Frequently asked questions

We are a tech startup first. We start from the reality of your company, we support you across every block, from the understanding phase to implementation to continuous transformation, and the software we put in your hands lets us move much faster and be far more relevant: it holds the map of all your processes, what is implemented, the roadmap and the ongoing transformation.

Above all, we are an engineering team, working at the state of the art of AI. Our job is to connect that technical state of the art to the business context of each company. We spent months building AI-native companies: thriving businesses that now run fully autonomously. We now help companies drive that same transformation for themselves. We are builders and operators, not consultants.

Three deliverables, one per block. Understand: the map of your company as it works today, and the scope to start with. Company Brain: your company memory connected to your tools, queryable, hosted on your side. Business agents: the agents in production in your infrastructure, with their traces, and the custom tools where a generic one is not enough.

As an option, the Disruption Playbook: a projection exercise that shows how far your business can go with AI, and the plan to get there.

The Disruption Playbook →

It's the best moment. Laying the right foundations before processes set in means nothing to undo, and our work begins with understanding your company before talking technology. No prerequisites, whether in your teams or your infrastructure.

We work with SMBs, mid-caps and large groups, across very different industries: the method builds on what your teams actually do, not on an industry playbook. If your organization runs on processes and the teams who carry them, our approach applies.

No. The understanding phase maps how the company works - processes, tools, frictions - not how people perform. Our team or our agents interview employees to understand how work actually gets done, never to score or rank them.

Our stance doesn't change: humans are amplified, not replaced. What the map reveals is time given back to teams for what matters, and roles that grow rather than disappear.

The map is yours: what you decide to do with it is up to you.

With a 30-minute call, to understand your situation and explain our method. By the end, we usually have what we need to send you a proposal on a first scope; on a wider scope, a scoping call with a lead from the department concerned is added. Then we deploy, we measure, and we extend to the next departments.

Weeks, not months. The duration of the understanding phase mostly depends on your pace: the time it takes for all your employees to complete their interviews. Then what gets prioritized is in production in the following weeks: the Company Brain connected to your tools, then the first business agents on your existing processes.

A sponsor at leadership level, and often one point of contact per complex business line, whom we turn to when a question of domain expertise or context comes up. No need for a data team or a full-time project manager: we carry the load.

AI evolves every month, and so does your company. We train your teams to use the agents deployed, we provide long-term maintenance of what has been implemented, along with continuous transformation support: applied watch on your use cases, new opportunities detected, agents kept at the state of the art. Your teams gain autonomy with every iteration.

Continuous transformation →

Your data stays your data, and it is strictly confidential: we sign an NDA at the start of the engagement that forbids any use outside the project we run with you. It is encrypted and hosted on highly secured servers. And it is never used to train models, ours or the vendors': we use configurations where your content is excluded from training.

No. OpenAI, Anthropic, Google, Mistral, open-source models: we evaluate every option, open source or proprietary, open weight, on premise or not, and we switch when the state of the art evolves. Our only bias is what works best for you.

Yes. Regulatory and security constraints are part of the mapping from the start. Depending on the case and your preferences: European hosting, deployment inside your infrastructure, open-source models running locally, ...

That is a good starting point, and few companies your size are there. The question is: what do your agents know about your company today? A licence knows nothing about you, a use case is a process. What we build is the level above: a memory that holds your entire context, and business agents that work on top of it.

If your data is already unified, the project is only simpler: let us check together whether your agents have the same level of context as a human, whether the relevant external data is integrated, whether your meetings and exchanges reach the right task at the right time.

The Company Brain →

They execute in your infrastructure. Your teams use them from the model they already use (Claude, ChatGPT or another), connected through an MCP to your company memory, or from the interface we build for you. Several models possible, no vendor imposed.

Fixed fee, defined before we start: one scope, one price, no surprises along the way. The amount depends on the size of the organization and the ambition; that is precisely what we define together on the first call.

One department: its tools connected into the Company Brain, and its first business agents. We prove the value on that scope, then add the next blocks, department after department. You have nothing to choose on your own: we propose the path.

Before starting, we define together what success looks like: measurable criteria. We are fully transparent about what is worth doing: we never push for marginal gains, only meaningful ones. And if a use case doesn't deliver, we say so, and we redirect or adjust it toward what actually creates value.