AI foresight keynote: what changes in your company over the next three years

I forecast AI from what I build.

A foresight talk on artificial intelligence, built from what I see running in production in my own products. Your leadership team leaves with three dated horizons (6 months, 18 months, 3 years), the signals that will tell whether each scenario holds, and the decisions to make now.

Check my availability

What the talk covers

Five topics that affect every company, covered with real examples and live demos when the format allows.

From assistants to AI agents

AI that answers a question is giving way to AI that acts: it picks up the phone, books a meeting, writes up a call or chases a customer. What already works in production, what is still fragile, and what it means for your teams.

The jobs that change first

Repetitive tasks change first, such as data entry, quotes, meeting notes, follow-ups and tender responses. I show how to spot them in your own processes before buying any tool.

What an AI won't copy

When producing becomes almost free, value moves elsewhere. Using the IRIA index, we go through the five moats that protect a company: certifications, partnerships, service guarantees, proprietary data and sovereignty.

Trust in the age of deepfakes

Anyone can now clone a voice, doctor a video or automate CEO fraud. I cover the reflexes to build into the leadership team and the wider staff before the first incident.

Sovereignty, rules and dependencies

The AI Act applies in stages, and your AI tools often depend on a handful of foreign vendors. How to keep control of your data and stay free to switch providers later.

Three horizons, three signals to watch

Every scenario gets a date and a signal that will confirm it or not. That means you can judge the talk six months later.

In 6 months: AI inside everyday tools

Email, office suites and business software all ship an assistant. Teams already use it, with or without rules, often with data that should never leave the company.

Signal to watch: the share of your staff using a personal AI tool for work.

In 18 months: agents inside processes

Agents handle whole tasks end to end: taking a call, qualifying a request, preparing a quote. Companies that mapped their processes see real gains; the others automate their mess.

Signal to watch: how many tasks your vendors offer to handle with no human in the loop.

In 3 years: competition is about moats

Whatever is easy to produce no longer protects you. Margins go to companies with clean data, service commitments, certifications and control of their infrastructure.

Signal to watch: new competitors in your sector whose offer is built on AI first.

Foresight from a builder

My foresight starts from the field. I begin with what the products I run actually measure (calls handled by a voice AI, generated meeting notes, sovereign infrastructure) and from verifiable public signals, then derive dated scenarios. When I don't know, I say so, and I point to what to watch to settle the question.

Why this talk, with me

I'm a developer and a founder of AI SaaS products. Every week I see what an agent can do, where it fails and what it costs in production. That practice lets me separate what is really coming from what is only announced, and give a leadership team markers it can check for itself.

Frequently asked questions

What is an AI foresight keynote?

It's a talk that helps a leadership team anticipate how artificial intelligence will affect its company and its sector. It sets out dated scenarios, explains what makes them likely and gives the signals to watch, so you can decide today where to invest and what to protect.

How is it different from a futurist talk?

The scenarios start from what already runs in production in products I operate, and each one has a horizon and a verifiable signal. You can hold the talk up against reality six months later.

What time horizon does it cover?

Three horizons: 6 months, 18 months and 3 years. Beyond that, the uncertainty is too high to guide a business decision, and I'd rather say so.

Who is it for?

Executives, leadership teams, business clubs, conventions and strategy offsites. No technical background is needed; the demos are chosen so everyone can follow them.

Can it be tailored to our sector?

Yes. A short call beforehand lets me pick the examples, jobs and signals that fit your sector. Formats range from a 45-minute keynote to a 2-hour talk with Q&A.

Do you speak anywhere in France, and in English?

Yes. I'm based in La Roche-sur-Yon (Vendée), I mainly speak across Pays de la Loire and western France, and I travel anywhere in the country. The talk can be delivered in English.

Give your leadership team a three-year head start

A short call is enough to nail down the topic, format and date. Tell me your sector and what keeps your leadership team up at night, and I'll propose the right talk.

Check my availability

See all my AI talks · The sovereign AI, AI Act and GDPR keynote · The IRIA AI resistance index · How to choose an AI speaker