Claude Code training: put an agent to work for you

Claude Code is not just for developers. We train your teams to hand whole tasks to an agent: file analysis, reports, internal tools, automations.

Let's Discuss

Trusted by

PayFitDoctolibSumUpNewfundInfraViaNextStageLouve InvestLe PointRamifyEVS CanadaSantianeMoose KnucklesNominationHystra-SanofiStrangeBeeCathay CapitalGiquelloMedicofiHibooTageosComwattKparKAgapéeEuropCar MartiniqueEtandexGroupe BremaudGroupe SprintArcane IndustriesCobi EngineeringVEUPCampingCar Park

What is Claude Code?

An agent that acts, not a chatbot that answers

Claude Code is Anthropic's agent. You give it a working folder: it reads your files, runs commands and delivers a complete result.

The name is misleading. We use it as much to cross-reference three Excel exports as to write software. And it runs in a desktop app, not only in a terminal.

  • It reads, writes and organises your files on its own
  • It plugs into your tools through MCP connectors
  • It remembers your methods through skills and CLAUDE.md

Claude Code

~/sales-reporting

Cross-reference the three exports in the sales folder and prepare the monthly review

Reading sales_q3.xlsx, crm_export.csv and targets.pdf

4-step plan proposed, approved

monthly-review.html created, gaps flagged in red

Ready for the next task

The programme

Four modules, from the first prompt to the first agent

Every module starts from a real task of your teams. We install, we practise, and you leave with something that runs.

01

Getting started

Install the app, pick a working folder, understand permission modes and plan mode. Guardrails first, speed second.

02

Working method

Scope a request, review a plan, manage context and roll back. The CLAUDE.md file keeps the project's memory from one session to the next.

03

Skills & MCP

Encode your methods into reusable skills and connect Claude Code to your tools: CRM, Notion, Google Drive, databases.

04

Agents & automations

Schedule routines, chain sub-agents, deploy an internal tool. What used to be manual now runs on its own, with checks.

In practice

What a working session looks like

Screenshots from our own sessions. This is the interface your teams learn during the training.

Claude Code asking scoping questions before running a request
01

It scopes before acting

Before running anything, Claude Code asks its questions. One minute of scoping saves twenty minutes of fixing.

Claude Code interface showing plan mode, the conversation and the site preview
02

It plans, you approve

Plan mode details every step before any change. The preview shows the result live, in the same window.

Interactive HTML presentation generated by Claude Code from a brief
03

It delivers a finished file

An interactive presentation generated from a few lines of brief, ready to review and share.

Train your teams on Claude Code?

30 minutes to map your use cases and build the right track.

Let's Discuss

Frequently asked questions

Do I need to know how to code for the Claude Code training?

No. You need to be comfortable with your computer and your files, not to have a developer background. Claude Code writes the code: you describe the expected result and you review. We regularly train business profiles to build their first agents.

What is the difference between Claude Code and Claude Cowork?

Cowork runs tasks on your files from the Claude app, within a simple frame. Claude Code goes further: it works in a project folder, plugs into your tools through MCP, keeps a project memory and can run routines on its own. For a business team, Cowork is often the first step and Claude Code the next one.

Which Claude plan do we need for the training?

Claude Code is included in Claude's paid plans and in the enterprise offers. We help you pick the right plan before the first session, based on the intended use. Our Claude Code pricing comparison details the limits of each plan.

Is it safe for our data and files?

The training starts with guardrails: a dedicated working folder, permission modes that require approval before any sensitive action, and plan mode to review before running. We also define with you which data can go through the tool and which stays out.

Is the training on site or remote?

Both. The first workshop works best on site, to install the tool on every workstation and unblock the first cases together. Later sessions and project support work very well remotely.

What can a team do at the end of the training?

Every participant leaves with at least one concrete case running on their own files: an automated report, an internal tool or a scheduled routine. And with the method to build the next ones without us.

Concrete results, accelerated growth. Shall we talk?

Book a meeting