# An AI agent in production in two weeks. Audit €1,500 excl. VAT · Sprint €4,500 excl. VAT Audits from €1,500 excl. VAT, AI Sprints from €4,500 excl. VAT. One workflow wired into your tools, with the code in your repo and a log of every action the agent takes. ## Why AI projects get stuck at the demo stage. Your team already uses ChatGPT, but nothing is connected to your tools. The prototype worked in the meeting, then nobody had time to make it reliable. That's the part we take on: we pick one process worth automating, connect it to your real data, and it runs under monitoring. - A prototype that worked in a meeting, then nothing - ChatGPT open in a tab, nothing wired into your tools - Nobody to harden it or watch it ## Three steps, each paid separately. You can stop after any step. ### AI audit + roadmap From €1,500 excl. VAT · 1 to 2 weeks Your processes reviewed, automations ranked, a 90-day roadmap. - Interviews with your team - A map of your processes - The 3 most profitable automations, ranked by time saved and risk - The GDPR points to deal with - A 90-day roadmap, presented in a review session ### AI Sprint From €4,500 excl. VAT · 2 weeks One workflow in production, with the scope fixed before work starts. - One workflow in production: agent, assistant or automation - A human approval step when the agent touches sensitive data - A log of every action the agent takes - A test set to check the outputs - A runbook and an onboarding session for your team - The code in your repo ### Monthly run On quote · Monthly The workflow monitored and improved every month. - Monitoring of the workflow - Model and prompt updates - Tests re-run after every change - A monthly check-in on what's running All prices exclude VAT. ## Every agent action, in the log. The view you get with the Sprint: every step, its duration, and the errors it recovered from. Shown here on a fictional example. ### The audit is credited against the Sprint If you go on to an AI Sprint, the quote takes the audit you already paid into account, and the terms are written down before you sign. - 00:00 s · Ticket: parcel not received - 00:01 s · AI model · reading the ticket - 00:03 s · Answer looked up in Notion - 00:04 s · Carrier silent, retrying - 00:09 s · Parcel found: due Thursday - 00:11 s · Reply drafted - 00:11 s · waiting for review - 04:02 s · Approved by support - 04:03 s · Reply sent to the customer ## What runs under your agents. The building blocks we put together for your agents and automations. We pick the model for the task, the cost and where your data has to stay. ### Technologies we deploy - Agents: Hermes Agent, OpenClaw, NanoClaw, Claude Code - Models: Claude, OpenAI, Mistral AI - Connections: n8n, MCP Brands mentioned belong to their respective owners. Klyx is neither affiliated with nor a partner of these vendors. ## Already in production. Hoppy, the AI assistant for Hop Hop Immo, and CiteMe, co-founded by Maxime, our lead engineer. [See the work](https://klyx.fr/en/work) - [Hoppy, the AI assistant for Hop Hop Immo](https://klyx.fr/en/work/hop-hop-immo): Hop Hop Immo · 2026. An AI assistant that helps buyers find their way through new-build housing in south-west France. - [CiteMe, measuring visibility in AI search](https://klyx.fr/en/work/citeme): CiteMe (co-founded) · 2026. A GEO tool that tracks a brand's citations across ChatGPT, Gemini, Perplexity and other AI engines. [See all prices](https://klyx.fr/en/pricing) By sector: [Real estate](https://klyx.fr/en/use-cases/real-estate) · [Web agencies](https://klyx.fr/en/use-cases/agencies) ## Frequently asked questions. ### What's the difference between an AI agent and an automation? An automation follows fixed rules: when an email arrives, it adds a row to the CRM. An AI agent reads, decides and acts within limits you set, for example sorting a customer request and drafting a reply. Many projects use both. ### n8n, Make or custom code? It depends on the process and on who will maintain it. n8n or Make work well when your team wants to change the steps itself. Custom code makes sense when you need fine-grained control, tests or high volume. ### What if the audit finds nothing useful? You keep the process map and the roadmap, even if you go no further. ### Is this audit a Bpifrance Diag Data IA? No. The Diag Data IA is delivered by experts accredited by Bpifrance, and Klyx isn't one of them. Our audit is a private audit at a fixed, published price. ### Is our data safe? Your accounts and tools stay in your name, and we map the data before writing any code. Data the agent handles goes through the model provider's API, which by default doesn't use it to train its models. If you need EU hosting or a local model, we plan for it in the quote. ### Who maintains the agent after the sprint? Your team, using the runbook and onboarding included in the sprint, or us through the monthly run. ## Start with the process that costs you the most hours. Book a 20-minute call. We'll work out together whether it's worth automating. - [Book a call](https://klyx.fr/en/contact) - [Get a free quote](https://klyx.fr/en/contact)