Pascal EK Loui · Independent AI consultant · AI Product Builder · Claude expert · 16 agents in production · 72-hour AI sprint from €3,500 excl. VAT · Paris and remote.
An AI consultant identifies the tasks where artificial intelligence delivers a measurable gain, scopes the use cases, selects the tools and leads deployment to production. I am an AI consultant in Paris with fifteen years of product work behind me, and I build what I recommend myself, with sixteen Claude agents in production.
The market is full of AI consultants who deliver a report and a roadmap. The report is often right. It just doesn't run. My rule: every engagement ends with something that works on your data. Otherwise it doesn't start.
Proof, not promises. Sixteen Claude agents in production. More than 100 reusable skills. Three platforms built solo: sceniq.studio, tondroit.fr, DevisGo. And AI products delivered for Accor (CallBOT ALL), Forvis Mazars (MAIA), Carrefour and the Ordre des experts-comptables.
| Profile | Part of the problem | Delivers | Paris price range 2026 |
|---|---|---|---|
| Traditional AI consultant | Diagnosis and strategy | A report, a roadmap | Day rate €600 to €1,200 |
| Data scientist | Models and data | A trained model | Day rate €550 to €800 |
| AI developer | Code and integration | One technical component | Day rate €550 to €1,000 |
| AI agency | Everything, with a team | A POC, sometimes production | POC €12,000 to €50,000 |
| AI consultant who builds | From scoping to production | A running system | Sprint from €3,500 excl. VAT, fixed price |
Market ranges: Hays France 2026, Malt, Koino's 2026 AI agency ranking. Mine are the rates I actually charge.
The question is no longer whether you need AI. It's how to get from pilot to production. That is exactly where an AI consultant who builds makes the difference over one who only recommends.
Fifteen years of framing problems, testing with real users and shipping under constraints, for demanding clients. That's the part of the job AI doesn't replace. Screen design is another matter: an agent can do it. I kept the judgment and changed the deliverable. A track record of failures and shipped products from before the models arrived is what separates an AI consultant from a ChatGPT user.
An AI consultant identifies the tasks and processes in a company where artificial intelligence delivers a measurable gain, scopes the use cases, selects the tools and models, then leads implementation all the way to production. Consultants differ on one point: do they deliver a running system or a report? I deliver the system, because I build it myself with Claude agents.
On the Paris market, a senior AI consultant charges between €600 and €1,200 per day (Hays, Malt, Koino 2026). I don't bill by the day. A one-week automation audit, quoted on request; a 72-hour AI sprint from €3,500 excl. VAT, with a working agent tested on your data; a 6-month to 1-year embedded engagement at a fixed price. The budget is agreed before we start.
An agency puts a team and a chain of handoffs between you and the result. An independent AI consultant works alone, which is an advantage if they can build and a limitation if all they can do is recommend. I combine both: the consultant's scoping and the agency's build, with sixteen agents in production to handle execution. See also my AI agency in Paris page.
A data scientist builds and trains models from data. An AI consultant starts from the business problem, decides whether it calls for a model, an agent, a RAG system or nothing at all, and leads the move to production. In 2026, most needs at SMEs and in product teams don't require training a model. They require orchestrating agents on top of existing models.
A 45-minute call to frame the need. Then either a one-week audit or straight into a 72-hour sprint, depending on how mature the topic is. At the end of the sprint, you have an agent running on your data and a costed go/no-go recommendation. You then decide on deployment, with me or with your own team.
The Anthropic ecosystem for agents: the Claude API, Claude Agent SDK, Model Context Protocol (MCP), reusable skills, sourced RAG, Claude Code. Next.js, Supabase and Vercel for production. Figma for design when it's needed. I stay model-agnostic when the use case calls for it.
Based in Paris, on site for the kick-off and workshops, remote for production. Clients in France, the United Kingdom, the United States and Asia, with an active company in Hong Kong. Reply within 48 hours.
45 minutes to frame the need, tell you what can and can't be automated, and put a price on the next steps. The first call is free. Reply within 48 hours.
Or by email: e.loui@uxdesignparis.fr · Phone: +33 6 35 31 50 35