AI, understood and put to work

Training and consulting for companies that want to use LLMs without overestimating them — or missing out.

Write to us See the programmes

Two ways to get AI wrong

Most teams we meet sit at one end or the other. Neither is a strategy.

Overestimating it

  • Betting a roadmap on a demo that never held up in production
  • Shipping confident, wrong answers to customers
  • Pasting confidential data into tools nobody vetted

Underestimating it

  • Banning the tools while the team uses them on their phones
  • Leaving hours of routine work on the table every week
  • Watching competitors learn faster

What we do instead

We give your team a realistic map: what today's models actually do well, where they fail, what they cost, and what that means for your work. Then we practise on your own cases.

Who it is for

Three audiences, three levels of depth.

  • Managers & decision makers

    Understand the technology well enough to decide: which use cases, which risks, which budget, which questions to ask vendors.

  • Small teams

    Operations, marketing, support, finance: put LLMs to work on daily tasks, with methods for checking the output and protecting data.

  • Development teams

    Coding assistants, agents, RAG, evaluation: how the job changes, what to adopt now and what to leave alone.

Training programmes

Short, hands-on, on your own cases. Every session ends with a written recap and next steps.

Half day

AI foundations for managers

For Managers, executives, team leads

  • What an LLM is and is not, without the jargon
  • A method to sort use cases by value and risk
  • The questions to ask before signing with a vendor
One day

Working with LLMs day to day

For Any team, no technical background needed

  • Prompting that works: context, examples, verification
  • Your own tasks turned into repeatable workflows
  • Data hygiene: what you can and cannot paste
One to two days

Agents, RAG & automation

For Technical teams, product owners

  • How agents, tool use and retrieval actually work
  • A small RAG pipeline built on your own documents
  • Evaluation, guardrails and cost control
One to two days

The AI-augmented development team

For Developers, tech leads, engineering managers

  • Coding assistants and agentic workflows in a real codebase
  • Code review, testing and documentation with AI in the loop
  • How roles in a team change, and how to manage that
Half day

Privacy, security & compliance

For Managers, IT, DPO, developers

  • GDPR and the EU AI Act applied to your use cases
  • Cloud, on-premises or local models: how to choose
  • A policy your team will actually follow

Groups of up to twelve. Larger teams are split into several sessions so everyone gets hands-on time.

Consulting

When a session is not enough.

How a session is built

Nothing off the shelf.

  1. Listen

    A first exchange by email to understand your context, your tools and what your team already does with AI.

  2. Tailor

    We adapt the module to your sector and bring exercises based on your own documents and tasks.

  3. Hands-on

    Half the time is practice. Everyone leaves with something they built or changed.

  4. Follow-up

    A written recap, next steps, and a check-in by email a month later.

Why ForeTale Studio

We teach what we practise.

  • We ship with these tools every day

    Our apps, games and this site are built with agentic workflows. We know what breaks, and what it costs.

  • We have managed development teams

    The studio's background is leading engineering teams before working with LLMs: we speak both languages.

  • Privacy is our default

    Our products work offline and keep data on the device. We bring the same discipline to your use cases.

Practical

Where
On site around Lyon, or remote
Group size
Up to twelve participants
Languages
French or English
Quote
Within 48 hours, fixed price

Questions

What companies ask us before writing to us.

Do participants need a technical background?

No. Only the agents & RAG module and the development-team module assume some technical background. The foundations, daily-use and privacy modules are built for non-technical audiences.

Which tools do you teach?

We work with the major model providers and with open models you can run locally. We teach methods rather than a single vendor, and help you choose based on your constraints.

Is our data safe during the training?

Exercises use your documents only if you want them to, and everything can run on a private workspace or on local models. Data handling is a topic in every module.

How many people per session?

Up to twelve, so everyone gets hands-on time. For bigger teams we run several sessions, or combine a short talk for everyone with workshops for smaller groups.

Can the training be funded?

Financing options depend on your company's size and sector. Tell us about your situation and we will explain what is possible.

Can you also build the thing afterwards?

Yes. Diagnostics and pilots are the natural next step, and the studio builds applications and integrations.

Let's talk about your team

Write to us