Custom AI agents, designed and coded by an engineer

Agents that read your emails, draft your proposals and answer from your documents. I design them and I code them, from idea to deployment.

10+

years of software development

24–48 h

to receive your proposal after scoping

100%

of what I deliver belongs to you

// 01 · business agents

Business agents: a repetitive task, an agent that does it

An overflowing inbox, a proposal to put together, a brief to write: too variable for a simple automation. The agent takes care of it, you approve before sending.

  • Triggered by an email, a schedule or a form
  • Connected to Gmail, Drive, your CRM, your APIs
  • Every action traced in the logs
project → Email sorting agent

// 02 · knowledge base

AI knowledge base (RAG): an AI that answers from your knowledge

Your documents, procedures and training materials are indexed. The AI searches for the right passages before answering, and cites its source.

  • Search by meaning and by exact words
  • You choose what goes into the base
  • Queryable through an interface, an API or from Claude
project → Knowledge Hub

// 03 · mcp servers

MCP servers: connecting Claude or ChatGPT to your tools

MCP is the open standard that lets an AI assistant call an application's tools. Your teams stay in Claude or ChatGPT, which finally sees your CRM, your projects, your data.

  • The actions you authorize: search, read, create
  • A revocable token per user
  • Read-only if needed
project → Internal cockpit driven from Claude

// projects

What I've built.

All projects →
Simplement · AI agent

Proposal writing agent

An application where sales reps prepare their proposals with an AI agent.

Korya · Web agency

AI web project preparation

From the client questionnaire to the project brief, with no re-entry.

Korya · AI agent

Email sorting agent

Every morning, sorts emails by importance and prepares the action to take for each important email.

// method

How I work

I run the meetings myself. The person who listens to you is the person who codes.

Amory Rouault
— Amory Rouault
  1. 01

    First conversation

    You describe the task to me. I tell you frankly whether an AI agent makes sense.

  2. 02

    Scoping and proposal

    Scope, data, access, what the agent does and does not do. Proposal within 24 to 48 h.

  3. 03

    Development

    I code and I test on your real cases, with a demo at each stage.

  4. 04

    Delivery

    Deployed, documented, taken over by your team. The code and the access are yours.

// choose

AI agent, automation or application?

The three are often combined. The question is: who makes the decision?

// faq

Frequently asked questions

A chatbot answers questions. An AI agent acts: it reads your emails, checks your CRM, drafts a document. You approve the sensitive actions.

Yes, through their APIs: Gmail, Google Drive, CRM, database, internal tools. With an MCP server, your teams can also access them from Claude or ChatGPT.

Claude and OpenAI, chosen according to the task. Sometimes through OpenRouter, to switch models without rewriting everything.

You do, 100%: code, access and documentation. You can take it over in-house or hand it to another provider.

// contact

Let's talk about your AI agent

Describe the task that takes up most of your time. I tell you frankly whether an AI agent is the right answer.