I help SMEs plan AI initiatives realistically, implement them on a solid technical foundation, and evolve existing solutions in a targeted way.
What matters is not the next tool, but a dependable foundation of clear data flows, roles, security rules, and an architecture that fits the company.
Thomas Stermole
AI architect & certified AI Manager (Austrian Standards)
It is not the next tool that decides, but whether value, data, roles, and technical guardrails fit together from the very beginning.
The problem: between the pressure to be efficient and regulatory reality, chaos often arises instead of clarity. Those who start too fast create new dependencies rather than real relief.
Employees adopt AI tools before responsibilities and boundaries are clear. With sensitive data, that quickly becomes risky.
Betting too early on the wrong stack or unclear data flows creates new dependencies. Correcting course later costs significantly more.
The problems usually lie elsewhere: unclear processes, poor data, missing roles. AI only becomes valuable once it fits reality.
I work directly, with a clear view of what makes sense professionally, is technically viable, and is organizationally realistic. Alongside client projects, I build my own AI products — which sharpens the eye for real requirements.
Working directly saves friction: you collaborate without detours with one person who connects strategy and technical reality.
Data sovereignty, hosting, governance, and data protection from the start — not only once the wrong decisions have been made.
Not the next framework or SaaS service. First it has to be clear what is sensible and feasible.
Directly with me — no detours, no lengthy coordination, no separation of strategy and technology.
It matters to me that AI is used ethically, morally responsibly, and fairly. I do not support AI initiatives for mass surveillance, weapons, manipulation, or the wholesale replacement of employees.
AI should relieve teams, prepare decisions, and keep responsibility visible.
Sensitive information needs clear data flows, secure architecture, and transparent rules.
No AI for mass surveillance, weapons, manipulation, or the wholesale replacement of staff.
Not every company needs an AI program right away. A clean workshop is often the better start. Once the direction is clearer, I provide support with architecture, roadmap, and implementation.
When clarity comes first: a workshop instead of jumping straight to tooling. Sort out potential, risks, conditions, and next steps together.
Suited to companies that want to start in a structured way and need a solid basis for decisions.
When the direction is clearer: guidance on architecture, stack selection, target picture, and implementation readiness — with a technical reality check.
Suited to companies that want to bring AI into processes and systems in a controlled way.
Each step builds logically on the previous one.
Clarify goals, priorities, and use cases.
Define data flows, roles, and technical guardrails.
Build and test a dependable prototype.
Trial the solution in daily use, gather learnings for rollout.
Not piloting something as fast as possible — but setting clear steps that hold.
Make sense of goals, the starting point, processes, data, and technical conditions together.
Identify use cases that deliver realistic value — instead of too much at once.
Define the target picture, hosting, data access, and integrations so the direction is viable.
Define the next steps so a workshop, pilot, or rollout can start without idle time.
Not every idea is a good use case. What matters: making internal knowledge more usable, easing routine work, keeping data under control.
Contracts, policies, documents, PDFs — made findable and usable in context faster via RAG, internal search systems, or secure knowledge assistants.
Research, preliminary analysis, summaries, documentation — noticeably accelerated once quality and control are clarified.
Where data, know-how, or internal information must stay protected: sovereign systems with clear roles and data flows.
Not more tools, but better decisions: which models, which infrastructure, which hosting variant make sense in the long run.
Close to the mid-market, with experience from larger organizations and regulated environments. What matters is not the size of the name, but the quality of the implementation.
„Thomas grasped the complex project quickly, communicated clearly, and delivered reliably. That blend of technical understanding and professional collaboration makes the difference.“
Martin Guntermann
Managing Director · experdoo GmbH
My AI expertise builds on years of experience in web development, platform architecture, UX, and systems thinking. That sharpens the view: AI is never isolated, but always part of data, processes, people, and business reality.
This combination of strategy, architecture, and operational implementation ensures that decisions not only sound good conceptually but also hold up in everyday use.
Thomas Stermole
Freelance AI architect · Certified AI Manager
Business understanding and architecture combined with real implementation experience — instead of just recommending tools.
Open source, local models, and controlled data flows are not an add-on but part of clean architecture.
Real projects, real constraints, maintainability, and sensible next steps — not generic AI promises.
My own AI products like FlowMind sharpen the eye for what systems really need to deliver in operation.
A short conversation is often enough to make sense of value, risks, and next steps. Clear, honest, and without sales pressure.