Why Voltaire exists
I could not answer a simple question about my own company: which growth path was the right one. Nobody could sell me the answer. That gap is the company.
I had a company called Cleen. Good product, real users, a market that existed. And a go-to-market I could not crack.
The problem was not effort. It was that I could not tell which of five plausible paths was the right one, and neither could anyone I asked. Every option had a story attached to it. None had a number. So the only way to find out was to pick one, spend three months and a chunk of runway finding out, and then pick again. Slow, expensive trial and error, dressed up as strategy.
I have run marketing at scale before. I was an entrepreneur in residence at Logitech. I have raised money and sold companies. I knew what good looked like, and I still could not buy it.
What I actually tried to buy
Consultants. The good ones are genuinely good. They are also episodic and priced for someone else. A study lands, the invoice clears, and the thinking walks out the door with the partner. You are paying for hours, so the invoice arrives whether the advice worked or not, and nobody comes back in six months to check.
Agencies. They execute. They do not decide. Ask an agency which of your five paths to take and you get the one they are set up to run.
Fractional executives. One person's judgment, at a day rate. Better than nothing and better than most agencies. Still one person's priors, and still no record of whether those priors have been right before.
Tools. Dashboards tell you what happened. I did not need a chart of a decision I had already made badly. I needed to know which decision to make next.
Every option was missing the same thing. Nobody would put a number on a recommendation and then stand behind it.
The thing that was missing
Big companies do not have this problem in the same way. They buy certainty. They hire a firm, run the tests, and decide against evidence, because they can afford six figures and two quarters to do it. Everyone else guesses and calls it instinct.
That gap is not about intelligence. It is about the cost of finding out. And the cost of finding out is exactly what collapsed.
A system can now read a whole business rather than a sample of it. Revenue, product, ads, CRM, market signal, all of it, continuously. It can hold that as a model instead of a slide, generate the candidate moves, score them against the evidence, run the cheapest test that settles the biggest bet, and record what actually happened against what it predicted. Then get better at the next one.
That is not a faster consultant. It is a different object. A consultant's memory leaves when they do. A system's memory compounds.
What we built
Voltaire reads your business, ranks the moves worth making, ships the ones you approve, and records the outcome against the prediction that produced it.
Three deliberate choices sit behind that.
Decisions come from a knowledge graph and a formal-logic engine, not from a language model. Ask a model the same question twice and you can get two answers, with no way to audit either. Models do the reading and drafting, which they are excellent at, under senior review. The decisions come out of the graph, where a conclusion traces to a rule and the records behind it, and the same question gets the same answer.
Part of the fee rides on the result. Every move ships with a prediction and an outcome window, and when the window closes the result is recorded against the prediction. That is only possible because the system measures what it recommends. It also means we are wrong in public, in your log, which is the only version of accountability that means anything.
Senior operators own the calls that matter. Agents produce the volume. A person signs anything substantive, and you approve what ships. Nobody should accept "the model said so" as a reason to move money, including us.
The honest part
Voltaire is early. We have no client logos to show you and no results table to point at, and I would rather say that here than let you find out on the call. What we have is a system, a method you can inspect, and a record that starts the day you connect.
But I know exactly what I was trying to buy three years ago and could not. This is that.
If you are sitting where I was sitting, holding four plausible paths and no way to price them, that is the conversation to have.
Fabrice Le Parc Founder, Voltaire