Strategic AI consulting for SMEs, complemented by professional web solutions and AI-powered content production.
Everything from a single source: from the first AI analysis to production-ready output. For companies in Upper Bavaria and across the entire DACH region.
About the founder, Tobias
Strategy · AI · Web Solutions · Content
Cross-industry experience
From mid-sized companies and startups to the public sector, with roles in e-commerce, administration, compliance and academic research.
Integrated skill profile
Understands technology, regulation and business logic, connects solutions and opens up new perspectives.
Certified AI specialisation
Master's degree in management with a focus on AI (2023) and further complementary certificates.
Sustainability as a guiding principle
Using AI responsibly, as a means to sustainable business success.
Vision
Why artificial intelligence?
Human intelligence means creativity, empathy and experience. Artificial intelligence means computing power, pattern recognition and knowledge transfer.
We help you bring the two together.
An upgrade for your processes.
More room for strategic value creation.
Our Services
Consulting Services
Tailored AI consulting, aligned with your specific needs.
Munich · Traunstein · Serving all of Germany, Austria and Switzerland
Insights
AI Know-How for Your Work
To the point and always up to date: we regularly publish relevant articles on AI, mid-sized businesses, workflows and techniques. Carefully researched and concisely presented.
AI Style or AI Slop: How do I tell quality AI from AI junk?
Generative AI makes design accessible to everyone while producing a flood of trivial content. How do you recognise quality AI use, and how do you avoid AI slop?
Learn more
AI in Mid-Sized Businesses: What Will Matter in 2026
Building AI skills to stay capable in times of talent shortages?
Learn more
EU AI Act: What to Expect in Q2 and Q3
The German government launched the national implementation of the European AI regulation on 11 February 2026. Where should SMEs pay particular attention now?
Learn more
Prompt Engineering: ChatGPT as a Vibecoding Assistant, Not a Generator
Enter prompts by hand or automate them? That is not a question of style but of success, because the method shapes the quality of an AI project from the very start.
Learn more
AI Briefing
Relevant AI updates for SMEs, the self-employed and freelancers
📅
Once a month we curate the most relevant developments in the world of artificial intelligence: compact, clear and with straightforward guidance for decision-makers. From regulatory updates to new tools and strategic impulses.
For more and more companies, AI is a building block of competitiveness. Those who acquire the right tools and skills today grow with technological change instead of being overtaken by it.
Not replacing people, but enabling them. Three examples.
👥HR+
Imagine this: your HR lead finally has time for what matters again, the people in your company. Because AI efficiently supports application screening, drafting references and recurring questions about leave policies.
📣Marketing+
With automated image editing and prompt-based content production, your marketing team can raise your company's public image to a new level without hitting its own capacity limits. Campaigns, social media posts and newsletters come together faster, more consistently and more precisely for the target audience.
📊Controlling+
With AI-supported data analysis, your head of controlling can produce extensive reporting in hours rather than days and derive forecasts with you even more precisely.
Consulting Services
Tailored AI support, aligned with your level of maturity and specific needs.
Generative AI Implementation
We assess where AI genuinely pays off in your company and implement practical solutions directly. No major IT project, but a clear focus on efficiency, added value and relieving your team.
Practical AI knowledge, tailored to your industry and specific use cases. Individual workshops and training sessions build confidence, orientation and productive competence for everyday work.
Not losing creativity, but easing the workload. Two examples.
🌐Web presence+
A familiar dilemma: either laboriously clicking a website together in a builder or running expensive feedback loops with an agency. Fortunately, that is over. Our specialised AI setup combines the best of both worlds: high quality and cost efficiency.
🎬Visual content+
Everyone knows the hurdles: equipment, external teams, cancellations, quality differences, lengthy planning. For smaller projects our approach is: let AI do the work, stay professional. As a supplement, short clips can look good with AI-generated elements too. But for longer portraits and personal content, human closeness remains irreplaceable. There, AI comes across not as innovative but as misplaced and overdone.
Production Services
Web and content built on the right AI setups. Highly professional, secure and tailored to your specific project.
Web Solutions
Highly professional websites, functional and cleanly coded in source code rather than assembled in a builder. Thanks to innovative AI methods, fast delivery at an unbeatable price-performance ratio compared with traditional agencies.
Highly professional AI-generated videos, images and dashboards. Visual content at agency level, faster and more cost-efficient through the targeted use of state-of-the-art AI tools.
Proven AI setups and services that solve familiar problems and unlock new potential.
FAQ
Frequently Asked Questions
We support all companies, the self-employed and freelancers who want to make the most of AI tools in their processes. The consulting is guided by your needs, your company size and your industry.
The non-binding strategy call (free of charge) lasts 30 to 60 minutes. The project duration varies with size and intensity. Developing an AI strategy for your company takes 2 to 4 weeks. A comprehensive implementation of AI applications in your processes (including quality assurance) takes 2 to 4 months. Here too: speed meets diligence.
We meet you where you are. Whether beginner, expert or somewhere in between, the focus is on developing your company, your team and your AI value chain.
We combine technical expertise with a focus on sustainability: individual, holistic and forward-looking for your company. Every solution is practical, regulatory sound and designed for long-term value creation.
Our consulting is available digitally across the entire German-speaking region (Germany, Austria and Switzerland). For on-site meetings, upskilling sessions and workshops, we arrange appointments individually with our clients. In south-east Bavaria and Munich we charge no travel costs for in-person appointments.
Of course. EcolutionAI doesn't just advise, it also produces. We create highly professional AI content, content tailored to you and websites with next-generation SEO, thanks to innovative AI applications at a price-quality ratio that stands out from traditional marketing agencies.
The hourly rate for projects in AI implementation, compliance and conception is between €105 and €125 plus VAT, depending on complexity.
Fixed prices are agreed for upskilling, more information under Upskilling.
Tiered price packages are also offered for web solutions and content, more information under Web Solutions and Content.
We assess where AI genuinely pays off in your company and implement practical solutions directly. No major IT project, but a clear focus on efficiency, added value and relieving your team.
The new tangible technology for SMEs
Generative AI is no longer a distant prospect, it is a tool. Now that Claude, Perplexity and Comet, Google AI, ChatGPT and comparable models are widely available, technologies that were reserved for large corporations with their own data science teams just a few years ago are open to small and mid-sized companies too.
For decision-makers in mid-sized businesses, this means the barrier to entry has never been lower. The question is no longer whether your company should use AI, but where the greatest leverage lies.
Future work: creativity and AI efficiency merge, taking your processes to a new technological level.
Use cases: which departments in your business benefit?
Generative AI does not deliver value in a single area only, it can work horizontally across the entire company. Here are the most common areas of application that we identify and implement with our clients:
Marketing & CommunicationWebsites, content creation, newsletter drafts and SEO optimisation, in a fraction of the time it used to take.
Sales & Customer ServicePersonalised offers, automated first responses, intelligent FAQ systems and lead qualification.
HR & RecruitingJob postings, onboarding material and internal knowledge bases.
Finance & ControllingAutomated reports, data preparation, anomaly detection and decision templates.
Management & StrategyMarket analyses, competitor research, scenario planning and management summaries at the push of a button.
Transformation
From process gridlock to the digital fast lane
Before
Analysis_Q1_2026_v3_FINAL2.xlsx
A
B
C
D
E
F
1
Revenue
Region
Q1
Q2
Δ%
Status
2
142.800
DACH
38.200
41.600
+8,9%
open
3
89.430
EMEA
21.100
24.800
+17,5%
review
4
67.210
APAC
15.900
18.200
+14,5%
open
5
53.100
NAM
12.400
13.900
+12,1%
final
6
31.990
LATAM
7.800
8.600
+10,3%
open
7
28.450
MEA
6.200
7.100
+14,5%
review
Inbox — Outlook
100+
M. Weber
Re: Q1 figures – corrections needed!!
09:14
S. Schmidt
FW: FW: FW: budget approval urgent
08:52
J. Berger
Where do I find the current reports?
08:31
L. Huber
Meeting tomorrow – please check the agenda
yesterday
T. Koch
Compliance document v4 attached
yesterday
Sales
Pipeline update – action items open
Mon
Compliance_Policy_2026.pdf
Excel
Outlook
PDF
Teams
Chrome
YouTube
Manual processes. Information overload.
After — with efficient AI implementation
AI assistant
EcolutionAI
Analyse the Q1 revenue data by region, identify the top 3 growth drivers and create a prioritised action list for the management meeting on Friday. Take the open compliance points into account.
Q1 summary: Total revenue: €412,980 (+12.3% YoY)
Top 3 action areas:
Secure EMEA growth (+17.5%)
Systematise the LATAM pipeline
Automate compliance approvals
Compare EMEA vs. APAC and create a forecast for Q2…
Revenue Q1
412.9K
↑ 12.3%
Efficiency
+38%
Time saved
Prioritised tasks
Consolidate the Q1 report
Complete the compliance check
Finalise the EMEA strategy
Prepare the board presentation
Automated analysis
Dynamic assessment
Opportunities
+EMEA growth scalable
+Automation saves 12h/week
+Compliance rate at 94%
+Data quality improved
Risks
−LATAM pipeline unstable
−Budget Q2 still open
−3 approvals pending
Summary
Structured workflows. Automated analysis.
🏆 Best-practice examples
Classification
Automatically sorting and categorising
Comparative analyses
Comparing contracts and documents intelligently
Document extraction
PDFs to structured data in seconds
Browser workflow optimisation
Eliminating time and energy drains
How exactly does EcolutionAI help you?
We don't switch on the technology straight away, we start with your workflows. In a structured process, we identify together the points where generative AI has the greatest leverage. Then we implement, pragmatically, measurably and sustainably.
Our approach follows a clear three-step process. First we analyse your existing processes and prioritise quick wins with a high ROI. Next we configure and test the right AI tools in your real working environment. Where needed, we upskill your team so the new tools don't end up in a drawer but make a difference in everyday work.
The result: your company gains time, reduces repetitive work and creates room for strategic, creative and value-adding activities, exactly where your people make the difference.
Our range of services
Model consulting
Evaluating AI tools, carrying out feasibility analyses and deriving clear recommendations for your specific use case.
Prompt & PoC development
Building prototypes, defining quality metrics and creating practical guidelines for your team.
AI vision & structure
Aligning stakeholders and developing access concepts and role models, for clear responsibilities.
Reporting & monitoring
Building transparent, traceable structures for ongoing performance measurement.
💼
Day rate
€1,000 to €1,250plus VAT
Depending on complexity and the topic-specific preparation required.
Professional roles as an executive assistant at a tech start-up, as a research assistant in the German Bundestag and as an adviser at ministerial level.
Academic background
Master's in Sustainability and Quality Management (focus on AI, integrated management systems, impact analysis). Bachelor's in Political Science and Law (focus on corporate policy, statistical methods, civil law).
Certified AI adviser
Continuously trained by the Stifterverband für Künstliche Intelligenz e.V.: data and algorithm ethics, use cases and the EU AI Act.
Commitment to the sensible use of technology
Active in professional associations and initiatives for the responsible use of AI in business and society.
For ideas, discussions and professional exchange, let's connect on
LinkedIn or
Github.
Your benefits
Why EcolutionAI?
🏆Successful projectsProven expertise across various industries
📈ROI-PushMeasurable results within a few months
🔄Up to dateModel updates and legislative changes on the radar
🤝CollaborativeLocal experts & personal consulting, not an anonymous service provider
🎯TailoredIndividual solutions instead of standard packages
🛡️GDPR & EU AI ActRegulatory compliant and future-proof
Principles
Our values
🛡
Responsibility
Deploying AI where it creates value, regulatory compliant, ethically considered and thought through for the success of your company.
#DataSovereignty #Relief #Fairness
◎
Transparency
Traceable processes, honest communication and explainable AI without black-box thinking. These are the firm principles of our work.
#Klarheit #Teamplay #OpenMinds
◆
Innovation
New AI models and updates appear every week. We filter what's relevant, test strategically and integrate only what creates real added value.
#Comparison #PoC #UseCases
Nationwide impact
From southern Bavaria into the whole DACH region
Just as people come to our region from all directions, we also travel nationwide to people and companies to support them through a major transformation.
Practical AI knowledge, tailored to your industry and specific use cases. We give your team confidence, orientation and productive competence, from getting started to advanced use.
🎯
With AI literacy training, you learn how to use the right AI tools for your workflows legally and securely.
⚖️
With AI literacy training in line with EU AI Act Art. 4 you already meet the requirement to upskill your organisation and your team.
How are the content and techniques delivered?
A short look (2 min.): how participants learn interactively on the Miro workspace. Subtitles can be switched on via the CC icon.
An interactive learning format instead of a PowerPoint lecture: the Miro workspace engages participants ideally, letting them learn actively and take part. The format works for in-person, hybrid and online settings.
Online sessions usually use Zoom. Through shared access, screen sharing and integrated exercises, participants actively practise their AI skills in real time.
Design your first projects during the upskilling itself? With Figma's interface design tools, you can give your successful AI outputs a face: prototypes, mockups, templates and much more.
In the growing model jungle, you keep developing your AI ecosystem
New language models, diffusion models and agents, plus weekly updates: the requirements keep shifting. Add to that subscription models, token systems and multi-LLM platforms. After the upskilling, you know how to keep your AI setup up to date and how to choose the tools by their suitability.
Our upskilling formats
🎯 In-house workshops
Full-day or half-day workshops on site with you or remote. We work with real case studies from your company and develop concrete application scenarios for your team together.
📊 Management briefings
Compact sessions for leadership: strategic context for AI, opportunities and risks, regulatory requirements and concrete recommendations for action, in 90 to 120 minutes.
🧠 Awareness training
For all employees: a basic understanding of AI tools, responsible use, data protection awareness and practical tips for productive use in everyday work.
✍️ Prompt Engineering
Advanced techniques for communicating effectively with AI systems. Structured prompt strategies, chain-of-thought, few-shot approaches and domain-specific optimisation.
Why upskill with us?
Our upskilling sessions are not standard lectures, they are tailored individually to your industry, your tools and your level of knowledge. The focus is always on direct applicability in everyday work.
Every participant leaves the session with concrete techniques and workflows that can be used straight away. We work with real use cases from your company, not generic examples.
On request, we support your team after the upskilling too: with follow-up sessions, a prompt handbook and ongoing availability for questions.
Adaptable to each format, depending on the type of event you are planning:
🎯 Internal workshop🎤 Conference📈 Employee development📚 Content for continued learning
Contents of the AI literacy training
🤖
1. AI technology
Terms & fundamentals
✍️
2. Prompt Engineering
Tuning and refinement 2026
💡
3. Use Cases 2026
Related to your processes
⚡
4. Productivity
Implementing simple workflows with AI
🧭
5. Ethics & awareness
Responsible use of AI
⚖️
6. Legal framework
EU AI Act, GDPR, CADA
🔥
7. Deep Dive
Discussion, exchange, Q&A
Our upskilling packages
Transparent, individual and with no hidden costs.
Entry
Flash briefing
For teams that need a fast, solid entry into AI-relevant topics, without a long lead time.
The upskilling sessions are all tailored individually to you and built interactively, not a generic lecture. The prices already include extensive preparation and follow-up work (course materials, templates, certificates of attendance).
FAQ
Frequently Asked Questions
No. The EU AI Act requires neither a certificate nor a competence assessment. However, the upskilling should be documented, so you receive an upskilling record for your files.
The obligation under Art. 4 of the EU AI Act has applied since 2 February 2025. Since then, companies must ensure a sufficient level of AI literacy among their employees. Supervision and enforcement take effect from 3 August 2026.
To all companies that use AI, regardless of the risk level of the systems used. It applies to employees who use AI or decide on its use.
No, the choice of format is up to you. In terms of content, an online workshop has no disadvantages or limitations whatsoever. What cannot be replaced, of course, is the dynamic, directness and personal interaction of a good in-person course.
We draw all course content from current specialist literature, online articles from reputable editorial teams, scientific studies and recent reporting. Information or claims from AI influencers, LLMs, social media channels and YouTube are only used in exceptional cases and labelled as such. Sources are listed for all exercise units, so everything remains traceable and verifiable.
Yes, that is generally possible and even sensible, for example when concrete use cases from your area already exist and need to be thought through. The EcolutionAI learning material is then used only where you have not yet prepared anything. We agree the exact process together.
Plan upskilling for your team
Let's work out the right format and content for your requirements together.
An initial call is always free and non-binding, whether by video, phone or through the contact form. Costs only arise once we work together concretely, according to the terms stated on the relevant service page.
Whatever is on your mind: whether an initial assessment of AI in your company, a concrete project (web solutions, upskilling, compliance review) or an open idea you'd like to structure together. We listen, ask questions and give an initial expert assessment.
You decide whether and how things continue, no sales pressure, no follow-up calls without your consent. On request, you receive a written summary after the call with concrete next steps and, if relevant, a suitable offer.
SMEs in 2026: countering the skills gap with in-house AI competence
Tobias Kisslinger · 29 January 2026 · 6 min read
Unfilled tech roles, now what?
Many mid-sized companies face a structural problem: tech positions remain unfilled, digitalisation projects are delayed, innovation cycles lengthen. The question is no longer whether action must be taken, but how.
Generative AI is becoming a standard tool in this. According to an OECD study (November 2025), 39 % of SMEs that use generative AI and had previously identified skills gaps report having at least partially closed those gaps with the help of AI. That is remarkable, but not a given.
Because generative AI does not replace skilled staff. It increases the productivity of existing teams, but requires subject expertise to produce valid results and avoid risks. Since the skills shortage will hardly resolve itself in the short term, companies must work on their substance.
Upskilling instead of waiting: systematically qualifying existing teams.
The strategic answer is therefore: upskilling instead of waiting.
The financial losses from unfilled tech roles can be at least partially offset by systematically upskilling existing employees. The barriers to entry are lower than ever, AI tools are becoming more user-friendly, more multimodal and more context-aware.
At the same time, the market is growing more complex: a multitude of models, constant updates, new pricing models. Without clearly defined use cases per department, workarounds, inconsistent outputs and, ultimately, frustration arise. AI needs strategic direction, not just login credentials.
AI competence: from an IT question to an organisational capability
AI competence is becoming a basic professional skill, across all departments.
The OECD data show a clear shift in competence profiles. For employees, data analysis and interpretation skills as well as creativity and innovation competences are becoming particularly important.
Generative AI changes not only processes but also job profiles. It is less about pure execution and more about evaluating, contextualising and strategically using results. For SMEs this means: AI must not remain an isolated IT project. The relevant use cases lie equally in sales, administration, production and at management level.
Companies that enable their employees to critically examine AI tools, direct them purposefully and use them responsibly secure operational efficiency and strategic resilience.
Regulation becomes operationally relevant: the EU AI Act
From August 2026, comprehensive obligations apply to AI systems in the EU.
With the German government's implementation of the European AI Regulation (EU AI Act), comprehensive obligations apply from 2 August 2026, in particular for providers and operators of high-risk AI systems. Risk management, technical documentation, human oversight and transparency requirements become binding.
Important: the regulation affects not only high-risk applications such as lending, HR decisions or critical infrastructure. Even seemingly "harmless" systems such as chatbots are subject to transparency obligations.
For most of the economy, the Federal Network Agency (BNetzA) becomes the central AI supervisory authority.
Where AI is integrated into already regulated physical products, such as machinery, vehicles or medical devices, the relevant product safety authorities will in future also oversee the AI components.
The consequence for SMEs is clear: those who establish governance structures early, define responsibilities and document AI systematically create certainty for decisions and investment. Compliance becomes a competitive factor.
Conclusion
2026 will not be a year of experiments for the SME sector, but of professionalisation. Productivity gains from generative AI are real, but only where competence, governance and clear strategic objectives come together.
AI competence is no longer an additional qualification. It is becoming a basic professional skill, comparable to basic digital skills 10 or 20 years ago.
The competitive advantage comes not from merely using AI, but from using it systematically, responsibly and anchored across the organisation.
Would you like to anchor AI strategically in your company?
We support you with implementation, compliance and enabling your team.
Prompt engineering: ChatGPT as an interface in vibe coding?
Tobias Kisslinger · 9 February 2026 · 5 min read
The new leap in quality in vibe coding
The term "vibe coding" is still young and describes AI-supported software development, from code generation to helping shape the structure of entire applications.
At its core, vibe coding is not an entirely new phenomenon: since 2022, ChatGPT has generated code snippets and helped developers with debugging, refactoring or setting up projects. For consistent, holistic source code, such as complex dashboards, prototypes, apps or scalable web applications, the system long ran into context and structure limits.
The market has since diversified: OpenAI is no longer the only big player in the GenAI business, but part of an increasingly competitive ecosystem of specialised models and development environments. Over the past twelve months, such tools have made a clear leap in quality that goes well beyond experimental gimmickry. Anthropic in particular, with Claude, has set standards here in terms of context processing, code understanding and structured output.
The barrier to entry for non-experts has dropped massively
Vibe coding is moving from an experimental approach to a serious productivity instrument.
For a long time, the tech community held the view that AI-supported code generation was at best a playground for non-developers, suitable for private experiments or small proofs of concept, but hardly for scalable applications. No-code and template solutions for websites, design or content accordingly remained widespread: why deal with source code, API integrations and debugging when projects can be clicked together efficiently in a building-block system?
The limits of such systems are, however, obvious: they are convenient but structurally limited. Individual logic, proprietary workflows or applications that are demanding both visually and functionally used to quickly burst the frame of pre-built structures, and made full-stack developers, specialised agencies or professional web developers all but indispensable.
This very barrier to entry is beginning to dissolve. With the latest model generations at the latest, such as Claude Opus 4.6 (January 2026), Anthropic has shown that AI is now able to structure complex codebases consistently and to help think through architecture decisions. Vibe coding thus moves from an experimental approach to a serious productivity instrument, including for ambitious non-experts.
Does ChatGPT build a bridge between humans and specialised AI tools?
Whether ChatGPT, as a useful partner, improves project work with specialised models such as Claude depends on the complexity of your own project. For clearly defined tasks, such as "create a landing page for project X with the following parameters", a precise prompt is enough: the model structures the page logically, integrates the defined requirements and delivers immediately usable code or a visual preview. Iterative adjustments can then be refined through dialogue, quickly, efficiently and without in-depth architecture planning.
It is different for large projects that involve system logic, data flows, interfaces and design principles. An unstructured "mega-prompt" rarely leads to consistent results here, and can also become costly. Many AI models work with usage-based limits; inefficient iterations quickly drive resource consumption, and therefore spending, into the triple-digit range.
Prompt engineering is becoming a strategic discipline, far beyond the IT department.
For complex projects, prompt engineering becomes a strategic discipline: requirements, context and target architectures must be clearly defined. In this process, ChatGPT can serve as a conceptual interface, for structuring and for working out precise specifications. The decision as to which coding models are used can, given appropriate user expertise, either be specified by the user or left to the AI model.
A field of competence is emerging here that goes far beyond the classic IT department and will be relevant for almost all departments, from marketing and product development to controlling and HR.
Appropriate upskilling of employees becomes a prerequisite for being able to use the opportunities of these technologies as a productivity and innovation boost in the company.
Would you like to use prompt engineering purposefully in your company?
We support you with practical upskilling and strategic consulting.
We review and validate existing AI applications for legal conformity, with compliance as a stabilising force.
The legal situation at a glance
01
Mandatory since 2 February 2025
Art. 4 of the EU AI Act requires a sufficient level of AI literacy for everyone who uses AI or decides on its use, regardless of the risk level.
02
No certificate, but documentation
Neither a certificate nor a competence assessment is required. The upskilling should be documented, so an upskilling record stating name, date, duration and content is sufficient.
03
Supervision from 3 August 2026
From this date, supervision and enforcement take effect. Those who upskill and document now can demonstrably meet their duty of care.
We offer technical compliance solutions for companies that already use or develop AI: we assess your tech stack for regulatory risks and implement governance that stands up to audits.
Inventory of your AI systems and classification by risk class (prohibited practices and high-risk systems).
Technical review of existing models for transparency and documentation gaps.
Creation of a concrete "remediation roadmap" to achieve compliance readiness.
Setting up and operationally introducing an internal risk management system (RMS) in accordance with ISO/IEC 42001.
Governance: defining ethical AI principles for your company as well as guidelines for employees.
On request, we act as your external AI officer: we continuously review your company for AI compliance against new standards, regulations and technologies. You also receive regular briefings and updates, so you stay on top of things without having to plan for a new full-time position.
Scope of services in detail
From the first analysis to documented implementation, we guide you to AI compliance in a structured way, in line with GDPR and the EU AI Act.
1
Needs analysis & risk assessment
Identifying planned AI use cases
Assessing data protection relevance (GDPR, EU AI Act)
2
Legal basis review
Analysis of the legal bases (e.g. consent, legitimate interest)
Checking for a data protection impact assessment (DPIA) under Art. 35 GDPR / Sec. 35 KDG
3
Documentation & record of processing activities
Creating and updating the record of processing activities
Developing templates for AI-specific processing operations
4
Data protection impact assessment
Carrying out a DPIA based on the documentation
5
General recommendations for technical and organisational implementation
Guidelines for the safe use of AI systems
AI transparency report under the EU AI Act
📋
EU AI Act
Systematic preparation for the requirements of the EU AI Act, from risk classification to technical documentation.
🔒
GDPR & data protection
GDPR-compliant AI implementation with privacy-by-design principles and transparent data processing.
🏗
ISO 42001
Building an AI management system in line with ISO 42001, for demonstrable quality and continuous improvement.
📊
Audit & Reporting
Automated audit trails and compliance reports for internal and external reviews, verifiable at any time.
A compliance check for your company
Let's check together whether your AI systems meet the current regulatory requirements.
Highly professional websites, functional and cleanly coded in source code rather than assembled in a builder. Thanks to innovative AI methods, fast delivery at an unbeatable price-performance ratio compared with traditional agencies.
Innovative web solutions powered by AI
We don't build websites in a builder system, but with clean source code, supported by state-of-the-art AI tools. The result: faster development cycles, higher code quality and a price-performance ratio that clearly stands out from traditional agencies.
🔍 SEO 2.0
Next-generation search engine optimisation: we use AI-supported analysis for keyword strategies, semantic structuring and technical SEO, so your website not only looks good but also gets found.
📦 Your code, your property
The entire source code passes fully into your ownership. No dependencies, no hidden licence fees, you keep full control over your digital infrastructure.
🏆 Top quality and long-term support
Every website goes through comprehensive quality checks, from performance and accessibility to responsiveness. And after launch too, we support you for the long term.
With AI technology, we reach a new level of individuality: visual elements, layouts and content can be generated exactly to your ideas. The usual limitation to the results of photo shoots and graphic designers falls away, your wishes are implemented precisely, not just approximated.
Our packages
Transparent, fair and with no hidden costs.
Smart
The fast online business card: clear, professional, live in 2 to 3 weeks.
Ideal for the self-employed, freelancers and emerging companies who want to present their offer online with confidence.
A top-tier, high-converting web presence: with AI integrations and agentic features, directly interlinked with your business processes.
Ideal for ambitious companies that want to expand their web presence into a value-creation platform, with measurable efficiency and competitive advantage.
Answers to the typical questions about your web project.
You pay a fixed package price, with no hidden costs. Any extensions are clearly discussed and stated in advance. Once the website is complete, ongoing hosting and domain fees apply. We are happy to cover this for you from €40 a month onwards, including technical maintenance and minor adjustments. Additional services such as regular new content can be added modularly and transparently. If you'd rather manage everything yourself, only the one-off price for the website applies.
The Smart package is usually live after 2 to 3 weeks; for the Signature package we plan 3 to 4 weeks, depending on the scope and how quickly you provide us with content such as logos, text or images.
As little as possible. We are happy to take over logos and existing image material. You can provide text as bullet points, we handle the wording in close coordination with you. We produce additional content with state-of-the-art AI technology, matched exactly to your brand.
You receive 3 months of free maintenance: technical updates, minor text corrections and security patches. After that you can either take over yourself (you own the complete source code) or optionally take out a maintenance contract.
Yes. On request, we integrate a simple content management system that lets you update text, images and news entries yourself, with no programming knowledge. For more complex changes, we are available at any time.
We host on modern, GDPR-compliant platforms such as Vercel, Netlify or, on request, German providers. Technical data protection (SSL, cookie banner, privacy policy) is included as standard and implemented in a legally sound way.
Of course. Every website is designed from the start for all devices, from smartphone and tablet to large desktop screens. Content, images and navigation adapt automatically to the screen size and stay clear and easy to use on every device.
Builders offer pre-made blocks, we develop custom code tailored exactly to your company. The result: faster load times, better placement in search engines, no monthly licence fees and no dependence on a platform. The source code belongs to you, so you stay flexible at all times.
Discuss your web project
Let's find out together how your new website should look, efficient, modern and on fair terms.
Highly professional AI-generated videos, images and dashboards. Visual content at agency level, faster and more cost-efficient through the targeted use of state-of-the-art AI tools.
There is a controversial public debate about how far AI-generated content should be used. There is no blanket answer: the purpose of using AI differs depending on the project, quality, target audience, ethical questions and budget.
Visual content that works
Whether a sample visualisation, a short brand clip, social media content or data-driven dashboards, we produce visual content that gets your message across. By using AI tools in a targeted way, we shorten production cycles considerably without compromising on quality.
🎬
AI video
Professional video production with AI-supported generation, editing and post-production, for results that impress.
🖼
AI images
Bespoke imagery for your brand, generated, edited and optimised with state-of-the-art AI tools.
📊
Dashboards
Interactive, visually appealing dashboards that present complex data clearly and in an action-oriented way.
📑
Presentations & pitch decks
Professional slides for investors, clients or internal stakeholders, created with AI support, visually convincing and to the point.
🎓
Explainer videos & tutorials
Animated or AI-generated short videos for onboarding, upskilling or customer communication, clear, compact and professional.
🧩
Interactive content
Configurators, interactive landing pages or micro-animations, digital experiences that activate users and boost engagement.
📈
Infographics & data visualisation
Complex relationships presented visually, for reports, blog posts, press material or internal decision templates.
⚡
Fast & cost-efficient
Thanks to AI-supported workflows, we deliver agency quality in a fraction of the usual time and cost.
What we advise against
Artificial intelligence is a powerful tool, but not the right one for every purpose. Where authenticity and real trust are decisive, we deliberately advise against visually AI-generated content:
🎬
Large brand films
Emotion and brand identity arise from real people, places and moments, and that cannot be credibly created entirely artificially.
👤
Personal portraits
Real portraits create closeness and trust. In this context, AI-generated faces quickly come across as fake.
🤝
Prices, terms & personal matters
Topics such as pricing, terms and anything that belongs in a personal conversation should not be presented as visually AI-generated content.
Discuss your content project
Tell us about your project, and we'll show you what's possible with AI-supported content production.
The free strategy call is aimed at managers, ambitious self-employed people and anyone who wants to find out how AI can be used concretely and with clear benefit in their company. Together we work out where AI makes sense for you, from strategic assessment and regulatory questions to upskilling and implementation. No buzzwords, no sales pressure.
What to expect
In no more than an hour, we build a clear picture of your starting position together and you get an initial, honest assessment of which steps are worthwhile for you. Concrete, non-binding and as equals.
The AI Act, CSR/ESG reporting obligations, the GDPR and other regulations make demands on every company that uses or develops AI. This calls for a holistic approach to making your own systems measurable, for internal quality management, transparent information flows and a credible public image.
EcolutionAI is most likely the first provider to offer a scientifically grounded and field-tested software-based compliance check. All KPIs on your AI systems, all risk classes, all information, brought together in one place.
Get in touch with us to be among the first to build this innovative approach into your own processes.
🛡️
🌱
🏆
Security
Sustainability
Quality
What we check
The check covers all compliance-relevant dimensions: risk classification under the EU AI Act, data protection requirements of the GDPR, transparency and documentation obligations, and technical quality standards.
We assess both in-house AI solutions and purchased software and SaaS services with AI components. Particular attention goes to the traceability of decisions, bias detection and the responsible handling of personal data.
The check works both as a one-off assessment and as a recurring audit instrument, and on request we integrate it into your existing compliance process.
Assessment matrix for advanced AI users
For companies that have already implemented AI extensively or develop their own AI systems, we additionally check the following dimensions:
⚖️ Liability & responsibility
Does your organisation have rules on liability aspects for potential damage caused by the AI systems you use? We check whether clear responsibilities are defined and liability risks are appropriately addressed.
🔬 Data quality & auditing
Against which quality criteria is the data you use for training and running your AI systems checked? Are industry standards such as ISO 25012 or ISO 8000 applied to audit the datasets? We assess your data processes and identify gaps.
🎯 Bias check
Is the training data systematically checked for bias? We analyse whether suitable procedures for detecting and mitigating bias in your datasets and models are established.
🌱 Energy consumption & resource efficiency
How high is the electricity consumption of your AI systems in kWh? Are metrics used to assess the resource efficiency of your own or external data centres, such as ISO 30134, Power Usage Effectiveness (PUE), Carbon Usage Effectiveness (CUE) or Water Usage Effectiveness (WUE)? We assess the ecological dimension of your AI use.
♿ Inclusive design
Which inclusive design principles guide you in developing your AI systems? Which principles are a prerequisite before you purchase an AI system? We check whether accessibility and inclusion are anchored as quality criteria in your procurement and development processes.
Request a compliance check
Let's check together where your AI applications stand in regulatory terms, and what to do next.
Your strategy briefing on the latest developments in the global AI landscape.
Tobias Kisslinger · 26 June 2026 · 8 min read
🛠 New tools & releases
Anthropic brings Claude into Slack as a team member (Claude Tag)
With Claude Tag, anyone in a Slack channel can @-mention Claude. Claude then pulls the context from the conversation, completes tasks and replies directly in the channel. It builds up context over time and, when ambient mode is enabled, takes the initiative itself. Administrators control token budgets and data sharing.
Reve 2.0 and Ideogram 4.0: two new image models
Reve 2.0 uses a code-based layout that lets you fine-tune every detail of a generated image, and ranks as the second-best model in the arena rankings, ahead of Nano Banana 2. Ideogram 4.0 is at the same time the currently top-ranked open-weight image model.
Microsoft Build 2026: Scout agent and seven new models
At Build 2026, Microsoft presented Scout, its first "autopilot" agent, which is meant to monitor the account and act proactively, for example preparing meetings or spotting to-dos. Scout runs on OpenClaw and is available through Microsoft's Frontier programme. In addition, Microsoft presented seven new in-house AI models.
OpenAI expands Codex into a workplace tool
OpenAI released six new role-specific plugins across 62 apps and 110 skills for knowledge workers, to turn Codex into a universal workplace tool. Also new is "Sites for Codex": it lets you generate ready-to-use websites or apps, complete with a shareable URL, from a single prompt.
⚖️ Regulation & policy
Bundestag passes AI act: Federal Network Agency becomes central supervisor
On 11 June 2026, the Bundestag passed the national act implementing the EU's AI Regulation and designated the Federal Network Agency as the central market surveillance authority for artificial intelligence in Germany. Business associations welcome the legal certainty gained. The opposition and civil society, by contrast, complain of shortcomings in fundamental-rights protection and warn of a federal tangle of competencies.
EU Council approves Digital Omnibus: concrete deadlines set
The EU Council has approved the Digital Omnibus Regulation, setting out a clear roadmap. The labelling obligations from the EU AI Act already take effect from August 2026. In December 2026, a ban on non-consensual sexual AI content comes into force. The regulation of high-risk AI systems follows in December 2027, with binding rules for AI in products from August 2028.
US locks out foreign users: Anthropic blocks Fable 5 and Mythos 5
The US government has prohibited all foreign nationals from using the Fable 5 and Mythos 5 models. Because of the broad interpretation of this order, Anthropic has fully suspended public access to both models in order to comply. The trigger was a jailbreak discovered by Amazon researchers that bypassed Fable 5's safeguards.
💼 Labour market & studies
Study: AI pioneers are not cutting jobs, quite the opposite
The fear of massive job cuts among office workers may be exaggerated. A study by the tech companies Ramp and Revelio Labs analysed staffing data from 22,000 US companies and compared it with their spending on enterprise AI solutions between January 2021 and February 2026. The result: the largest AI adopters tended to grow their workforce after adoption and hired new staff faster than competitors, including entry-level workers. A caveat: these firms were often larger, more often venture-funded and were growing faster anyway; much of the growth was in the information sector. Even so, the findings call into question the assumption that AI is the main cause of mass layoffs. "If you read headlines in which CEOs attribute layoffs to AI, you should be sceptical," said the study's authors.
🔬 Research
Meta Brain2Qwerty v2: turning thoughts into sentences
Meta's updated research tool Brain2Qwerty v2 analyses brain signals captured via a non-invasive scanner and decodes whole sentences from them in real time, with 61% accuracy. It is Meta's most powerful brain-to-text system to date, trained on around 22,000 sentences from nine volunteers with 10 hours of recording each. The aim is to help people who can no longer speak after a stroke, accident or brain disease.
🧭 Strategic assessment
SemiAnalysis: how heavily AI labs subsidise their subscriptions
The analysts at SemiAnalysis maxed out the weekly usage limits of all Anthropic and OpenAI plans and compared the pure compute costs with the subscription prices. The result: a Claude Max subscription at $200/month can cost Anthropic up to $8,000 in monthly compute, and ChatGPT Pro can cost OpenAI up to $14,000. The gap shows how aggressively the labs accept losses to retain power users. For companies this means: today's subscription prices are not a reliable benchmark for the long-term cost of productive AI use.
🎲 Fun fact
For Meta's "thought-to-text" system Brain2Qwerty v2, nine volunteers each sat for around ten hours thinking sentences instead of speaking them: together about 90 hours of focused silence for 22,000 sentences of training data, ending with a 61% hit rate. Sometimes the greatest progress in AI research is simply to sit still and think.
Do these topics affect your company?
Let's discuss how you can use the latest AI developments strategically for your company.
Your strategy briefing on the latest developments in the global AI landscape.
Tobias Kisslinger · 29 May 2026 · 8 min read
🛠 New tools & releases
Anthropic releases Claude Opus 4.8, valuation reaches $965bn
Opus 4.8 leads in coding, knowledge work and, notably, honesty: Anthropic specifically trained the model not to make confident statements on a thin factual basis, a common weakness in frontier models. Also new are a cost-efficient Fast Mode and Effort Control for managing compute. Available at the same price as Opus 4.7. In addition, in May Anthropic secured the entire Colossus 1 supercomputer in Memphis (over 220,000 Nvidia GPUs) and at the same time doubled the 5-hour limits of Claude Code, the compute race keeps accelerating.
Claude generally available in Microsoft 365
Claude is now generally integrated into Microsoft 365: in Word, PowerPoint and Excel, with full context continuity across apps and documents. Repeatedly explaining the working context is no longer necessary. Available in the paid plans, with Outlook in public beta. For companies already using Microsoft 365, a clear productivity lever without a new tool landscape.
ChatGPT creates and edits PowerPoint decks directly
OpenAI now lets users create, update and revise PowerPoint slides directly in ChatGPT, without copy-and-paste. The generated slides remain fully editable in PowerPoint. Currently in beta. With this, Claude for 365 and ChatGPT compete directly for the workflow layer in everyday office work.
Microsoft 365 Copilot: faster and calmer in design
The new Microsoft 365 Copilot interface swaps the previous, rather colourful look for a cleaner, calmer design, consistent across Excel, PowerPoint and Word. Microsoft states that after the redesign the app loads twice as fast and response time on complex prompts improved by 10%. A sign that providers are increasingly using UX and performance as differentiators too.
ElevenLabs Dubbing v2: emotion across language barriers
ElevenLabs has unveiled its Dubbing v2 model: translations of videos into more than 90 languages, preserving the tone, acting performance and speech rhythm of the original. The aim is more authentic and at the same time more cost-efficient dubbing, relevant for content production in international markets.
Google Health app, Health Coach and Fitbit Air
Google is bundling health tracking into a new Health app that captures fitness, sleep and vital signs centrally. The Gemini-powered Health Coach analyses the data and gives personalised recommendations 24/7. In parallel, Google presented the Fitbit Air, a screenless wearable with a week of battery life. The app rolls out from 19 May, Fitbit Air goes on sale from 26 May. Some features require a premium subscription.
🤖 Robotics & automation
Japan Airlines tests humanoid robots at Haneda Airport
With tourism at record levels and a rapidly shrinking workforce, Japan Airlines is turning to humanoid robots to keep operations running at one of the world's busiest airports. A two-year trial is testing their use in baggage handling and cabin cleaning. Background: Japan's working-age population is projected to fall by around a third by 2060. Analysts expect the government to position humanoids increasingly as a demographic answer.
The US wants to push China out of robotics, easier said than done
A new bipartisan bill aims to ban the use of Chinese ground robots by US federal agencies, part of a broader push to sever sensitive technology ties with Beijing. The problem: US manufacturers themselves rely heavily on Chinese components. A ban further down the supply chain could hit precisely those companies it is meant to protect. Analysts criticise that the US still lacks a coherent strategy for the larger tech competition with China.
🇩🇪 Germany: Germany Stack, a sovereign AI cloud for public authorities
SAP, Telekom & Schwarz Digits build a €250m public-administration AI
With state funding, SAP and Deutsche Telekom are planning a joint, sovereign AI infrastructure for federal, state and local government, the Germany Stack. The stack is intended to handle sensitive administrative tasks: document processing, translations and automated summaries. One of the first concrete applications is KIPITZ, a specialised AI assistant for public-sector employees.
The federal government has provided around 250 million euros for the build-out. The contract went to a consortium of T-Systems, SAP, SVA and Schwarz Digits. But the procurement process stalled: on 28 May, a competing alliance of Adesso and Google filed a complaint with the procurement tribunal. Critics such as the Free Software Foundation Europe warn against too strong a dependence on individual providers and demand that software financed with taxpayers' money should generally be released as open source.
⚖️ Regulation
FTC enforces a 48-hour deadline for deepfake removals
The US authority FTC is forcing large tech corporations to implement the Take It Down Act: after a removal request, AI deepfakes must be removed within 48 hours and their re-distribution blocked. Violations can incur fines of up to 53,000 US dollars per individual case. Warning letters went to Alphabet, Amazon, Apple, Meta, Microsoft, Discord, TikTok, X as well as other platforms such as Reddit, Snapchat, Bumble and Pinterest. The letters do not constitute a finding of a violation, but serve as a preventive notice of gaps in existing control systems.
🧭 Strategic assessment
Three trends shape May. First, AI is moving into the office workflow: Claude becomes generally available in Microsoft 365, ChatGPT builds PowerPoint decks, Copilot gets faster and calmer. The AI sandbox in the browser is becoming a thing of the past; in future the model works directly in Word, Excel and PPT. Second, AI sovereignty is taking on industrial form: the Germany Stack creates, for the first time, a concrete, state-financed public-administration AI cloud without primary dependence on US hyperscalers, even if the award is still legally contested. Third, AI and robotics are pushing into physical skills gaps: Haneda robots and Fitbit Air show that demographic bottlenecks and health tracking are increasingly being answered technologically.
Our recommendation: if you use Microsoft 365 productively, evaluate concretely which workflows would benefit from Claude or Copilot. As a public-sector client, it is worth watching the Germany Stack procurement process, even if the complaint delays the start. And for platform operators: review take-down processes for deepfakes now, before regulatory pressure forces haste.
🎲 Fun fact
In May, Anthropic secured the entire Colossus 1 supercomputer in Memphis, once announced by Elon Musk's xAI as the "world's largest training cluster". More than 220,000 Nvidia GPUs changed hands, and on the same day Anthropic promptly doubled the 5-hour limits of Claude Code. The AI compute market now resembles 1970s oil trading more than the classic tech industry: hardware is swapped faster today than business cards at conferences.
Do these topics affect your company?
Let's discuss how you can use the latest AI developments strategically for your company.
Your strategy briefing on the latest developments in the global AI landscape.
Tobias Kisslinger · 27 April 2026 · 7 min read
🛠 New tools & releases
Anthropic leak: Claude Code source code publicly accessible
A security researcher discovered the full source code of the viral coding tool Claude Code, around 1,900 files and 512,000 lines of code. The leak not only offers insight into how it works, but also reveals more than 20 as-yet-unreleased features on the roadmap. It is already the second Anthropic leak within a week; previously the secret model "Mythos" had come to light. For users this is a double-edged signal: transparency builds trust, while unintended disclosure raises questions about internal security.
Softr: no-code platform for production-ready business apps
Softr's new AI-native platform lets non-developers, too, build sophisticated business applications such as customer portals, CRM systems or inventory management, including database, permissions and user interface. The finished apps run on Softr's infrastructure, and adjustments are made visually without programming. An interesting building block for companies that want to implement quick in-house developments without a major IT project.
Claude Managed Agents: build your own agents in the cloud
With the new Managed Agents (public beta), Anthropic provides the complete infrastructure to run your own cloud-based agents. Users define tasks, tools and guardrails, and Claude handles the productive operation. Production-grade agents, long-running sessions and multi-agent coordination are supported. This significantly lowers the barrier for companies to put their own AI agents into productive use.
Gemini: direct file generation for Workspace and Microsoft 365
Google Gemini can now generate Word, Excel, PowerPoint, Docs, Sheets and Slides files directly in the chat. PDF, CSV, LaTeX, TXT, RTF and Markdown are also supported. The manual copy-paste step is gone, the finished file can be downloaded directly. A noticeable time saving for office-driven workflows.
Perplexity announces a 1-million-dollar competition
With the "Billion Dollar Build", Perplexity is launching an 8-week challenge in which teams are to build a realistic billion-dollar company using Perplexity Computer. The reward: 1 million dollars in seed investment plus 1 million dollars in computer credits, split across up to three winners. Registration until 14 April, ten finalists present live. An interesting signal of how aggressively Perplexity is courting developer mindshare.
📡 Updates & developments
OpenAI reportedly building its own smartphone
According to reports, OpenAI has teamed up with MediaTek, Qualcomm and Luxshare to develop its own smartphone, probably with its own operating system in which AI agents could replace classic apps. The device joins a growing hardware pipeline that includes a ChatGPT speaker, smart glasses and a smart lamp. The strategic thrust is clear: OpenAI does not want to remain just a software layer, but to be present directly on the device.
Copilot in Outlook goes agentic
Microsoft Outlook now supports an Agent Mode that manages mailbox and calendar more efficiently. Users describe the goal and Copilot takes over, with intermediate check-backs so that adjustments are possible at any time. Currently available in the Frontier programme. For companies with high email volumes, a potentially significant productivity lever.
Apple iOS 27: AI more deeply embedded in the Photos app
Apple reportedly plans three new AI features with Apple Intelligence: "Extend" for expanding image backgrounds with generative AI, "Enhance" for automatic lighting and quality corrections, and "Reframe" for changing perspective on spatial photos. Presentation expected at WWDC on 8 June. With this, Apple continues to catch up in consumer-oriented AI.
🇩🇪 Germany: Aleph Alpha & Cohere merge
German-Canadian AI alliance for more sovereignty
Heidelberg-based Aleph Alpha and the Canadian company Cohere have announced their merger in Berlin, in the presence of Federal Digital Minister Karsten Wildberger and his Canadian counterpart Evan Solomon. The goal: a "global AI champion" as an independent alternative to US providers, especially for public administration, finance, defence, energy, telecommunications and healthcare.
The Schwarz Group (Lidl, Kaufland) is participating with around 500 million euros as "structured financing". Cohere CEO Aidan Gomez emphasised customers' data sovereignty, while Wildberger spoke of a strong signal for Germany as an AI location. In parallel, Schwarz Digits is investing eleven billion euros in a new data centre in Lübbenau (Spreewald). With this, the topic of AI sovereignty in Europe is taking on industrial form for the first time.
🧭 Strategic assessment
Three trends shape April. First, agent platforms are going mainstream: Claude Managed Agents, Copilot Agent Mode and Perplexity's competition show that autonomous workflows are no longer research but product. Second, platform competition on the device is picking up speed: OpenAI's phone plans and Apple's AI photo tools mark the shift from "AI in the browser" to "AI on the device". Third, Aleph Alpha + Cohere create, for the first time, a serious European-Canadian sovereignty axis, flanked by Schwarz Group capital and data-centre construction.
Our recommendation: check whether your company must keep data in the EU, the new alliance creates concrete options here. Assess which recurring workflows are suitable for Managed Agents before competitors extend their lead. And watch the providers' hardware strategy: whoever bets on software lock-in today will find device lock-in tomorrow.
🎲 Fun fact
The AI location of the German-Canadian joint venture from Aleph Alpha and Cohere is in Heidelberg, the same city where, back in the 19th century, Friedrich Hölderlin wrote his famous verses on the relationship between reason and creation. Perhaps that fits quite well: anyone who reads Hölderlin knows that technical progress needs more than just computing power.
Do these topics affect your company?
Let's discuss how you can use the latest AI developments strategically for your company.
Your strategy briefing on the latest developments in the global AI landscape.
Tobias Kisslinger · 20 March 2026 · 7 min read
🛠 New tools & releases
Anthropic doubles Claude usage limits & 1M context window now generally available
Over two weeks, Anthropic is doubling the usage limits for all plans (Free, Pro, Max and Team) during off-peak hours. The bonus usage is applied automatically. In parallel, the full 1-million-token context window for Claude Opus 4.6 and Sonnet 4.6 has been made generally available, with no price increase. This allows significantly larger media files to be uploaded in chats.
OpenAI launches GPT-5.4 with computer use and a financial services suite
OpenAI has unveiled GPT-5.4, its most capable model to date, available in two modes: Thinking for everyday work and Pro for complex tasks. The highlight: native computer-use capabilities in Codex and the API. This makes GPT-5.4 the first OpenAI model that can autonomously click, type and operate software to carry out multi-step workflows. In addition, ChatGPT for Excel is launching in beta.
Microsoft launches Copilot Cowork built on Anthropic's technology
Working closely with Anthropic, Microsoft is bringing Copilot Cowork into the Microsoft 365 ecosystem. The solution integrates Claude Cowork's multi-step task execution with enterprise security, governance and compliance. Users hand over tasks, Cowork creates a plan and executes it across Outlook, Teams, Excel, PowerPoint and Word, backed by Microsoft's Work IQ intelligence layer. Currently in a limited research preview, with broader access to follow through the Frontier programme.
Google launches Nano Banana 2, faster AI image generation
Nano Banana 2 is the latest version of the popular image generation model and builds on Gemini 3.1 Flash. It produces images from 512px up to 4K resolution in various aspect ratios, maintains consistency across up to five characters and 14 objects in a single workflow, and can incorporate real-time information from web search for more precise results.
Luma AI launches creative agents for complete ad campaigns
Luma AI has unveiled Luma Agents: AI agents that handle end-to-end creative work across text, image, video and audio, powered by the new Uni-1 model. The agents coordinate with external models such as Google's Veo 3, ByteDance's Seedream and ElevenLabs Voice, and evaluate and refine their own outputs through iterative self-critique. Available via API.
📡 Updates & developments
Meta considers mass layoffs of up to 20% of its workforce
According to reports, Meta is considering extensive layoffs that could affect 20% or more of its almost 79,000 employees. The move comes in the context of massive AI investments: 600 billion dollars are planned for data centres by 2028, along with recent acquisitions such as Manus. The layoffs are not yet confirmed, a Meta spokesperson called the report speculative.
Elon Musk admits: xAI is being rebuilt from the ground up
Elon Musk has confirmed that his AI start-up xAI is being rebuilt from the foundations up, as it struggles to keep pace with the competition. In recent weeks several co-founders have left the company, including Tony Wu, Igor Babuschkin, Kyle Kosic and Christian Szegedy. The exodus falls in a delicate phase, as SpaceX, now the owner of xAI, prepares an IPO this year.
🏛 Germany: Bundestag debates the AI Implementation Act
Plenary speech by Digital Minister Dr Karsten Wildberger (20/03/2026)
On 20 March 2026, the first reading of the act implementing the EU AI Regulation took place in the Bundestag. In his speech, Digital Minister Wildberger emphasised that Germany needs AI regulation that enables innovation rather than stifling it. The draft law relies on existing authority structures instead of a new AI super-agency. The Federal Network Agency will receive a coordination and competence centre, an AI service desk for companies and an AI regulatory sandbox for legally sound experimentation. The central message: companies should quickly gain legal clarity and clear points of contact. Wildberger also argued that Germany must develop AI technologies itself instead of merely being a user. At the same time, he took societal concerns seriously: productivity gains must reach people, and AI in education must not lead to creativity and moral judgement being outsourced to chatbots.
🧭 Strategic assessment
The pace of AI development picked up markedly again in March. Four trends shape the quarter. First, computer use is becoming the new standard: GPT-5.4 and Claude can now operate software autonomously. Second, agentic AI is establishing itself as an enterprise feature, no longer just a research topic: Copilot Cowork and Luma Agents show that multi-step autonomous workflows are arriving in production environments. Third, AI regulation in Germany is becoming operational: the Implementation Act gives companies concrete points of contact and compliance paths for the first time. Fourth, the staffing situation at Meta and xAI shows that even big-tech players are under massive consolidation pressure.
Our recommendation: check whether your AI tools offer computer-use capabilities and how these can be integrated into existing workflows. Use the legal clarity of the Implementation Act to build your compliance structure now, before the deadlines take effect.
🎲 Fun fact
Google's new image generation model really is called "Nano Banana", named after an internal joke by the development team. The name came about because the model showed a striking preference for yellow tones in early tests. Google kept the codename as the official product name, a rarity in the tech industry's rather sober naming conventions.
Do these topics affect your company?
Let's discuss how you can use the latest AI developments strategically for your company.
The EU AI Act is the world's first comprehensive set of rules for artificial intelligence. Its aim: AI systems should be used safely, transparently and in line with fundamental rights. For companies in the DACH region this means concrete obligations to act, regardless of whether they develop AI themselves or merely use it.
Most of the EU AI Act's obligations become enforceable from 2 August 2026. AI is now deeply embedded in business processes, from customer service and fraud detection to recruiting and decision automation. As these systems increasingly influence financial outcomes and access to services, regulators classify them as a source of operational and societal risk.
Who does the EU AI Act affect?
The scope is not limited to EU companies. Under Article 2, the regulation applies to any organisation that develops, deploys or makes available AI systems within the EU market. Companies outside the EU are not exempt, what matters is where the AI is used, not where the company is based.
Article 3 distinguishes between providers (developers of AI systems) and deployers (users of AI systems in day-to-day operations). Both groups carry specific obligations that vary depending on the system's risk classification.
Risk-based classification
AI systems are regulated according to a risk-based model. High-risk applications are subject to the strictest requirements for governance, documentation and monitoring. Compliance requires companies to inventory their AI systems, assign risk classes, maintain technical documentation, implement governance processes and monitor AI performance across the entire lifecycle.
Core obligations at a glance
Inventory of all AI systems, risk classification according to EU requirements, creation and maintenance of technical documentation, implementation of structured governance frameworks, and continuous monitoring of AI performance across the entire system lifecycle.
Current developments: Q2/Q3 2026
AI Omnibus: simplification before the deadline
On 14 March 2026, EU member states granted the mandate for inter-institutional negotiations with the European Parliament on the AI Omnibus. As part of the Digital Simplification Package, it is intended to simplify the AI Act's high-risk requirements before the August deadline. At the current pace, a final text can be expected by May 2026.
European Commission opens consultation on AI supervisory procedures
On 12 March, the European Commission published a draft for implementing acts on evaluation procedures for general-purpose AI models by the EU AI Office, including the involvement of independent experts. The consultation runs until 9 April.
Cloud and AI Development Act (CADA) delayed
The CADA is now expected to be presented on 27 May 2026. The initiative aims to triple data-centre capacity in the EU, establish EU-wide requirements for cloud providers and create a unified cloud policy for public administrations. Together with the revision of the EU procurement rules, the package is being touted as the "Tech Sovereignty Package".
Digital Fairness Act (DFA) gains momentum
The DFA, a new legislative initiative to protect consumers in digital environments (dark patterns, addictive design, unfair personalisation, subscription traps), is picking up pace. The legislative proposal is expected for Q4 2026. More than 200 civil-society groups called on the Commission in an open letter for an ambitious proposal. Consumer Protection Commissioner Michael McGrath stressed that the measures would be "targeted and proportionate".
EU Inc.: the 28th legal framework
On 18 March, the European Commission published its legislative proposal for "EU Inc.", an optional, harmonised EU-wide legal framework for companies. It addresses the fragmentation that forces companies expanding across borders to navigate 27 national legal systems. Key features: company registration within 48 hours, fully digital, for under 100 euros and with no minimum capital. Plus a "once only" principle with a single EU interface to national commercial registers.
What does this mean for your company?
Security and governance leaders take on new duties. CISOs and GRC teams must ensure that AI systems are identified, assessed and managed through structured oversight processes. This includes: understanding how AI is used in the organisation, documenting compliance evidence and continuous risk monitoring across the entire system lifecycle.
Our recommendation
Start now with inventorying your AI systems and risk classification. Use the remaining time until August 2026 to build governance structures, before the deadlines take effect. EcolutionAI supports you with a structured compliance check and practical consulting.
Is your company prepared?
Let's examine together where your AI applications stand in regulatory terms, and which steps are necessary by August 2026.
AI style or AI slop: how do I tell quality AI from AI junk?
Tobias Kisslinger · 12 June 2026 · 6 min read
Quality AI or AI junk?
Generative AI makes design and content production accessible to everyone. But the same low barrier that opens up new opportunities also produces a flood of trivial content, often called "AI slop". The key question is: where is the line between valuable use of AI and AI junk?
Causes of AI slop
The low barrier to entry can be an advantage, but it also exposes weaknesses. Without an eye for aesthetics, the user misses the mark. Public perception has become more sensitive. Bad design is noticed even by non-experts. AI does not protect against this, on the contrary: used carelessly, the result can be even worse than a listless designer working entirely without AI.
AI slop is produced en masse, because AI content can be generated extremely cheaply and quickly without any human quality control. Typical drivers of its spread are monetisation through viral posts with ad revenue, scaling by flooding search engines and social media feeds, manipulation through fake images or politically distorted content, and prototyping on a large scale without staffing effort. Content farms and bots automate this process entirely.
GenAI is a young technology
We are still at the beginning of generative creativity. Some stylistic breaks are still forgiven, because they occur very frequently and can look modern in the eye of the beholder, while others simply come across as tasteless. Generic content, used to excess, disfigures the product you represent. The fascination with GenAI must not obscure that.
Overloaded, generic motifs with no clear focus are a typical hallmark of AI slop.
The growing volume of low-quality AI content also strengthens AI critics' case for far-reaching regulation, which correspondingly slows the adoption of high-quality and productive AI applications. Alongside the loss of quality, AI slop increasingly leads to a significant loss of trust among users and brands, as it becomes ever harder to tell what is true and what is false.
What can companies do about AI slop?
Transparency, quality control and human editing are essential means of countering AI slop. AI-generated content must always be checked and based on facts so that it delivers real added value. AI content should be openly labelled, and hallucinations prevented through structured knowledge databases (such as a knowledge base). Ultimately, this also increases user trust.
Modern language models (LLMs) enable fast, scalable content production and preparation of content, but only to good effect if they are precisely directed and professionally checked. Used wrongly, they amplify the information overload and complicate knowledge management. What matters is the balance between generic and human-made content: the unreflected spread of generic content should be avoided.
A) Avoid overloading
A quick, thoughtless prompt ends up overloading designs. A currently striking example that everyone should recognise: the neighbouring sports club has published the poster for its next summer party. The impression: a firework of exaggeratedly laughing faces, staged-looking motifs, little white space and garish effects. The overall impression is cluttered yet still soulless. The viewer does not know where to look.
B) Narrow down the subject
Whether image, video or graphic: generated content only seems credible if its purpose is clearly defined. This includes context, mood, objects, light and shadow, and the message. If, say, only one of these parameters is described in the prompt, the LLM adds the rest itself. And in most cases it fills in with arbitrary content that the well-meaning AI user judges to be fitting, but which to outsiders seems completely out of place and interchangeable.
C) Maintain consistency
In image series, marketing campaigns or graphic content, AI-generated parts often do not sit coherently together. Format, form, soundtrack or colours seem arbitrary and convey no clear brand message. The solution: customise the model, cache files in the LLM, create project folders, train skills in Claude for example, use proven prompt templates for recurring tasks and keep returning to them, instead of starting "from scratch" with every prompt.
The difference between AI style and AI slop lies not in the tool but in the standards behind it: a clearly defined purpose, consistent design and human review determine whether AI enhances a product or drowns in the noise.
Would you like to steer your AI content towards quality rather than slop?
We support you with upskilling and AI-assisted content production with human quality control.
Information pursuant to § 5 DDG (German Digital Services Act)
Tobias Kißlinger
Ecolution AI
Adalbertstraße 56
80799 München
Location & service area
In-person availability and on-site appointments: greater Munich area and south-east Bavaria (Traunstein, Berchtesgadener Land, Rosenheim, Mühldorf, Altötting, Rottal-Inn).
Appointments throughout the DACH region on request.
Responsible for content pursuant to § 18(2) MStV (German Interstate Media Treaty)
Tobias Kißlinger
Adalbertstraße 56
80799 München
Dispute resolution
We are neither willing nor obliged to participate in dispute resolution proceedings before a consumer arbitration board.
Privacy Policy
1. Controller
Tobias Kißlinger
Ecolution AI
Adalbertstraße 56, 80799 München
Email: hello@ecolution.ai
2. Hosting
This website is hosted by Vercel Inc., 340 S Lemon Ave #4133, Walnut, CA 91789, USA. When this website is accessed, the server automatically records technical access data (IP address, time, page accessed, browser, operating system). The legal basis is Art. 6(1)(f) GDPR (legitimate interest in the secure provision of the website). The data is not merged with other data sources. Further information: Vercel Privacy Policy.
3. Google Fonts
This website uses fonts from Google Fonts (Google Ireland Limited, Gordon House, Barrow Street, Dublin 4, Ireland). When the page is accessed, your browser establishes a connection to Google servers, during which your IP address is transmitted to Google. The legal basis is Art. 6(1)(f) GDPR. Further information: Google Privacy Policy.
3. Cookies
Our websites use so-called "cookies". Cookies are small data packets and do not cause any damage to your device. They are stored on your device either temporarily for the duration of a session (session cookies) or permanently (persistent cookies). Session cookies are automatically deleted at the end of your visit. Persistent cookies remain stored on your device until you delete them yourself or until they are automatically deleted by your web browser.
Cookies can originate from us (first-party cookies) or from third-party companies (so-called third-party cookies). Third-party cookies enable the integration of certain services from third-party companies within websites (e.g. cookies for processing payment services).
Cookies have various functions. Many cookies are technically necessary, as certain website functions would not work without them. Other cookies can be used to analyse user behaviour or for advertising purposes.
Cookies that are necessary to carry out the electronic communication process, to provide certain functions you have requested or to optimise the website (necessary cookies) are stored on the basis of Art. 6(1)(f) GDPR, unless another legal basis is specified. The website operator has a legitimate interest in storing necessary cookies for the technically error-free and optimised provision of its services. Where consent to the storage of cookies and comparable recognition technologies has been requested, processing takes place exclusively on the basis of this consent (Art. 6(1)(a) GDPR and § 25(1) TDDDG); the consent can be revoked at any time.
You can set your browser so that you are informed about the setting of cookies and allow cookies only in individual cases, exclude the acceptance of cookies for certain cases or in general, and activate the automatic deletion of cookies when the browser is closed. If cookies are deactivated, the functionality of this website may be limited.
You can find out which cookies and services are used on this website in this privacy policy.
3b. Server log files
The provider of the pages automatically collects and stores information in so-called server log files, which your browser automatically transmits to us. These are: browser type and browser version, operating system used, referrer URL, host name of the accessing computer, time of the server request and IP address.
This data is not merged with other data sources.
This data is collected on the basis of Art. 6(1)(f) GDPR. The website operator has a legitimate interest in the technically error-free presentation and optimisation of its website, for which the server log files must be collected.
4. Contact form
If you contact us via the contact form, the data you enter (name, email, message and, where applicable, company and telephone number) is processed to handle your enquiry. The legal basis is Art. 6(1)(b) GDPR (pre-contractual measures) or Art. 6(1)(f) GDPR (legitimate interest). Your data is deleted after your enquiry has been fully handled, unless statutory retention obligations exist.
4b. HubSpot (appointment booking)
On our website you have the option of arranging appointments with us. For appointment booking we use the "HubSpot" platform. The provider is HubSpot, Inc., 25 First Street, 2nd Floor, Cambridge, MA 02141, USA.
For the purpose of booking an appointment, you enter the requested data and your preferred date in the form provided. The data entered is used for planning, carrying out and, where applicable, following up the appointment. The appointment data is stored for us on HubSpot's servers, whose privacy policy you can view here: https://legal.hubspot.com/privacy-policy
The data you enter remains with us until you ask us to delete it, revoke your consent to storage, or the purpose for storing the data no longer applies. Mandatory statutory provisions, in particular retention periods, remain unaffected.
The legal basis for the data processing is Art. 6(1)(f) GDPR. The website operator has a legitimate interest in arranging appointments with prospects and customers as simply as possible. Where corresponding consent has been requested, processing takes place exclusively on the basis of Art. 6(1)(a) GDPR and § 25(1) TDDDG, insofar as the consent covers the storage of cookies or access to information on the user's device (e.g. device fingerprinting) within the meaning of the TDDDG. The consent can be revoked at any time.
The data transfer to the USA is based on the European Commission's standard contractual clauses.
5. Your rights
You have the right to information (Art. 15 GDPR), rectification (Art. 16 GDPR), erasure (Art. 17 GDPR), restriction of processing (Art. 18 GDPR), data portability (Art. 20 GDPR) and to object to processing (Art. 21 GDPR). To exercise your rights, please contact: hello@ecolution.ai.
6. Right to lodge a complaint
You have the right to lodge a complaint with a data protection supervisory authority. The competent authority is the Bavarian State Office for Data Protection Supervision (Bayerisches Landesamt für Datenschutzaufsicht, BayLDA), Promenade 18, 91522 Ansbach, www.lda.bayern.de.
7. SSL encryption
For security reasons, this website uses SSL encryption. You can recognise an encrypted connection by the padlock symbol in your browser's address bar.
Accessibility statement
We (Ecolution AI), as the website operator, endeavour to design the website in line with the relevant accessibility regulations. The following legal provisions apply to us:
Bavarian Act on the Equality, Integration and Participation of People with Disabilities (Bayerisches Behindertengleichstellungsgesetz, BayBGG) and the Bavarian Accessible Information Technology Ordinance (BayBITV).
Website
This accessibility statement applies to the website ecolution.ai.
Evaluation
The website was evaluated with regard to contrast values and technical implementation using the "Google Lighthouse" tool. The result of 98% indicates excellent accessibility.
Feedback and contact details
Have you noticed shortcomings in the accessible access to content on our website, or do you have comments or questions about accessible access? Please get in touch with us at: hello@ecolution.ai
Status of compliance with the requirements
This website is not fully compliant with the accessibility regulations applicable to us. In detail:
Non-compliance with the accessibility legislation
The content listed below is not compliant with the relevant statutory accessibility provisions:
Easy language and sign language Compliance with easy language should be reviewed and optimised on an ongoing basis. In principle, the content is written to be comprehensible. Sign language is not offered.
Keyboard focus and keyboard operability Keyboard focus on individual elements of the page is generally provided. In principle, the website can be operated via the keyboard.
Alternative texts for images and graphics A graphic that has no informative function does not require an alternative text. Graphics without an informative function are, for example, spacers, colour areas, patterns or purely decorative photos. Such graphics should be marked with an empty alt attribute (alt=""). These are not set consistently throughout.
High-contrast colours and typographic clarity Text passages highlighted in colour increase attention for non-impaired people and indicate that they are a link or a similar function. For blind or colour-blind users, however, this is not self-evident and not readily perceptible. For information (text and icons) conveyed via colour, the contrast values are sufficiently high. Brightness contrasts meet the specific standards.
Disproportionate burden
The following content is not provided in an accessible form because this would place a disproportionate burden on us: where the effort is too high, corresponding trade-offs must be made (see Article 5 "Disproportionate burden" of Directive (EU) 2016/2102).
The following areas implemented on the website involve special technical configurations, and converting them to accessible use would require considerable effort. This covers all functions integrated on the website by third-party providers. For example, both a map via OpenStreetMap and the calendar/booking terminal via the external provider HubSpot Germany are embedded.
The current costs and effort would be disproportionate in relation to the currently expected benefit and needs for people with and without impairments. The financial and situational arguments, which have so far not provided for developing the functions mentioned in-house, prevail.
Accessible alternatives
For our non-accessible content we provide the following accessible alternatives:
Magnification: texts and content can be enlarged by more than 200 % to improve readability without loss of functionality.
The read-aloud function can be used thanks to correct syntax.
Keyboard commands help with orientation on the website.
Alternative access to information is offered via the navigation on the desktop as well as the mobile version of the website. The navigation elements are consistent and support understanding of where you are and which connections can be explored.
Enforcement procedure
If you believe that you are disadvantaged by insufficient accessible design of our website, you can contact the responsible enforcement body. You can reach them at:
Landesamt für Digitalisierung, Breitband und Vermessung IT-Dienstleistungszentrum des Freistaats Bayern Durchsetzungs- und Überwachungsstelle für barrierefreie Informationstechnik St.-Martin-Straße 47 81541 München Email: bitv@bayern.de
Preparation of this accessibility statement
This statement was prepared on 15/02/2026. The statement was last reviewed on 01/03/2026.
🍪 Cookie notice
This website uses only technically necessary cookies to ensure the functionality and navigation of the site. No tracking or analysis of your user behaviour takes place. Further information in our privacy policy.
