FREN

Book a call

00 /Cambon AI · Applied R&D for small and mid-sized companies

Your hardest problems are the ones that pay the most.

A researcher to frame the problem, an entrepreneur to turn it into value. With a PhD in AI and more than ten years as an entrepreneur, I tackle what weighs on your margins: optimization, forecasting, targeting, decision support. From a well-framed problem to a solution running in production.

Let's talk about your problemSee the projects

01 /The shift

Providers are moving up the stack.

The AI market in three layers

Return on investment

Layer 3 · Hard problems

Payoff: margin, new markets, new products

Example problems

  • route optimization, shop-floor scheduling
  • demand forecasting, pricing
  • targeting: the right offer, to the right customer, at the right time
  • anticipating failures (predictive maintenance)
  • robotizing a workshop, automating a lab

Before

Out of reach for SMEs

Labs, specialist consultancies, corporate R&D

Now

Cambon AIProblem framing, prototype, solution
↑ IT services firms

starting to move up

Layer 2 · Production problems

Payoff: productivity on your core business

Example problems

  • setting up a data stack
  • building custom business software
  • connecting software systems
  • plugging AI into internal documentation: support, tender generation

Before

IT services firms

Custom development, slow and expensive

Now

Cambon AI designs and buildsData architecture, reliable document AI, and more
IT services firms

now building to order

Layer 1 · Support functions

Payoff: productivity on overhead costs

Example problems

  • chasing payments, invoicing
  • sorting email, writing meeting notes
  • accounting: entering invoices, bank reconciliation
  • HR: tracking leave, preparing payroll

Before

SaaS vendors

One tool per function

Now

AI agentsMainstream tools are replacing SaaS

Set up by small providers, or by teams themselves after training

Yesterday, optimizing delivery routes or forecasting demand precisely took an R&D team. Today, a small or mid-sized company can take it on.

The whole market is moving up a notch. Support functions are shifting from software to AI agents, which teams are learning to use themselves. Custom development is getting faster and cheaper. And problems once reserved for labs and large corporations are now within reach.

That's where I work: on the problems that move margins, markets and products.

02 /

What I do

Hard problems

Goal: margin, new markets, new products.

  • Optimizing routes, scheduling a workshop
  • Forecasting demand, setting prices
  • Targeting the right offer, to the right customer, at the right time
  • Anticipating failures
  • Robotizing a workshop, automating a biotech lab
  • Data science for biotech: choosing the next experiments based on previous ones

Production problems with a strong R&D component

Goal: productivity on your core business.

  • Setting up a data stack and making use of your data (data science)
  • Building custom business software
  • Connecting your software systems
  • Plugging AI into your internal documentation: support, tender responses
  • Deploying AI tools in regulated environments

What I don't do: train your teams on mainstream tools, or deploy agents for support functions. Others do that very well.

03 /

How I work

  1. Frame. A free 30-minute first conversation. Then I formalize the problem and estimate the expected return on investment before committing to delivery. We decide together whether to go ahead.

  2. Prototype. A proof on your data, in a few weeks. We measure the actual ROI before investing further.

  3. Deliver. I build the solution and put it into production, or hand it over to your teams.

  4. Deploy and go to market. I make sure your teams adopt the solution. And when it becomes a new product, I build the go-to-market with you: positioning, pricing, first customers. I was an entrepreneur for eleven years and sold my own solutions: I know what it takes for a product to find its customers.

I can also support you over time, a few days a month, as a fractional Chief AI Officer.

04 /

Why me

  • Researcher.

    I formalize the problem and prove the solution holds.

  • Entrepreneur.

    I pick the problems that pay, and I know how to sell what comes out of them.

  • Builder.

    I build and deliver, with no middleman.

05 /

Who I am

Stéphane Cambon

Stéphane Cambon, PhD in artificial intelligence and entrepreneur.

I founded and ran Ownpage for eleven years. We built, sold and deployed personalization software used by about 40% of major French media, and the company was sold in 2024. I then served as COO at Opper Marketing. I keep working on the economics of the press: I published an open-source global licence model for French publishers.

Before that, I was a researcher. My PhD focused on planning under multiple constraints for autonomous systems (Université Paul Sabatier, 2005). I then took AI into the field, across very different sectors: aerospace, with predictive maintenance on the Rafale and the A320; healthcare, with anesthesia software in hospitals; video games, with the AI of the strategy game R.U.S.E.

I never let go of scientific research. My latest work proves mathematically that a robot cannot reach certain objects.

06 /

Projects

All projects →

07 /

A problem that won't budge?

Tell me about it in 30 minutes. If I'm not the right person to solve it, I'll tell you.

Book a call