A simple vision: AI tools that save several precious hours on repetitive tasks, managed by qualified people who are experts in applying these technologies, not simple gadgets installed to tick a box. This is the vision we share with symbolist, an AI agency based in Lille that thinks the same way we do on the subject. This synergy and the complementarity of our two methods is what led us to present their work in detail rather than vaguely mentioning them between two publications.
The observation that brings us together
Since late 2024, the use of generative AI has almost doubled among French small and medium-sized businesses. The Bpifrance Le Lab barometer puts the adoption rate at 55% at the end of 2025, compared to 31% a year earlier. The figure is real, but it above all tells the story of the speed at which executives got started, not the solidity of what they built with it. With symbolist, we feel that automation and optimization through AI agents can save far more hours than what these 55% allow today.
Integrating AI into a one to three position structure remains simple: a few Claude licenses and two or three specialized agents almost always suffice. Integrating AI into a 50-person SME is an entirely different matter. The company already carries a heavy software history, data scattered across several databases and a team whose members do not all master AI at the same level. Symbolist’s work then makes perfect sense: the agency helps these larger SMEs structure this shift rather than improvising it. We agree with them on this point. For ScaleMyCrew too, AI is only a means of reaching an objective, never an end in itself.
Who is symbolist
symbolist is an AI agency founded by Arnaud Rofidal, Perrine Dewally and Stéphanie Bonnet. It is a collective augmented by AI: professionals from several fields (digital marketing, software development…) who know how to get the best out of the tool by entrusting it with repetitive tasks, while retaining control over judgment and decisions. The agency refuses empty discourse and prefers to prove through figures rather than impress.
This shows in the way they talk about their own work. Few superlatives, little gratuitous jargon. When they mention a client, they cite a result rather than a promise. We recognize this register, because it is also the one we maintain ourselves on this blog.
“We do not sell promises. We deliver systems that perform. Symbolist was born from the conviction that innovation only has value if it transforms durably. Companies do not need magic, they need method, rhythm and measurable performance.” This is how they present themselves and this requirement is found in the work they deliver.
The resemblance between the two methods holds in one sentence: neither symbolist nor us start with the tool. Each tool is designed to respond to a real need. You never choose a tool to find a use for it afterwards. In other words, AI must not serve to follow a trend. It must solve a precise problem, identified in daily work before anyone talks about a tool, and support teams on a daily basis. It is this proximity of method that naturally brought us closer to symbolist. At ScaleMyCrew, a system of AI agents serves both our clients directly and the dedicated offshore team in Madagascar that uses it every day to execute the work entrusted.
Their offering, in three pillars
Symbolist’s offering is organized in three pillars. Strategy & AI Change handles data diagnostics and the support of teams that must change their working habits to better adopt AI. Growth & AI covers search engine optimization and B2B prospecting, both supported by AI, with a CRM then taking over to manage commercial activation. Operations & AI brings together HR transformation and recruitment. This is also where the construction of AI agents is housed.
This last pillar is the one that speaks most directly to us, because it touches on a subject we have been working on internally for more than a year.
Building an AI agent system: two methods, one same logic
At symbolist, building an AI agent system follows four stages: a workshop framing to identify where AI genuinely has its place in the client’s processes, an architecture designed from the outset so that a use case can be added or a tool changed without rebuilding everything, the activation that deploys and connects the retained use cases, then long-term monitoring that corrects and improves the system once it is running.
This is where our two methods meet. Symbolist as well as ScaleMyCrew handle both the strategic framing, which use cases to retain and which architecture to build, as well as the implementation. And when necessary, both provide qualified people to manage this daily use of AI.
These people matter. A well-designed AI agent absorbs almost all of the operational work, all the repetitive daily work. But it cannot replace the eye of an expert, nor their judgment. There is also the fine-tuning: these final adjustments that adapt the result to the client’s precise case. This is where a qualified person becomes necessary to support the expert, a person who knows how to communicate with the tool. This is an equally decisive part as the agent itself.
But beyond all of this, ScaleMyCrew also offers dedicated teams that support SMEs on entire functions, HR or client relations for example, where these same AI tools make daily work more efficient.
What it produces at their clients, in figures
Symbolist documents its results on its site. For Tribee, a solidarity kitty platform for gifts, the work focused on growth and search engine optimization. For an anonymous international consulting firm on their site, the agency automatically qualified 5,000 companies through AI and raised the response rate on outreach by 28%. The same firm also gains 35 hours per month thanks to automated workflows with the Make tool. For BEWE, a company supporting more than 350 companies on Microsoft 365, the agency set up a unified HR cockpit in two months, adopted by all on-site teams.
These are precise results, not promises of total transformation. This is precisely what makes them credible.
Our shared vision for companies that move
What brings us together holds in one simple conviction: AI in business is only worth something if it starts from a real need, never from a trend. An SME that buys a tool because a competitor adopted it rarely builds something that lasts. An SME that starts from a precise task, a report that takes too long, a prospect followed up too late, builds a system that resembles it and remains useful in two years.
We also agree on what we make available to companies. Symbolist delivers an AI agent system and, if needed, qualified people to support the client’s teams in its use. A tool is only worth what it genuinely produces once in place: on this, we think alike. At ScaleMyCrew, we offer the same thing. In addition, we allow an SME to entrust part of its activity to dedicated teams in Madagascar, trained in the use of these same AI tools to execute the work entrusted on a daily basis.
What we are asked most often
Two agencies, one shared conviction
Symbolist’s work truly resonates with ours. Both agencies start from the same starting point: an AI tool must respond to a real need and remain in service of the people who use it, a system must be designed to last. It is this shared conviction that makes us want to present their work in detail, not just a courtesy between partners.
If your SME needs an AI agent system framed from the outset, their site is symbolist.ai. If you are also looking for a team to run this kind of system on a daily basis, we are here to discuss it.
(1) [Bpifrance Le Lab](https://lelab.bpifrance.fr/Etudes/31-des-tpe-et-pme-utilisent-l-ia-generative) : generative AI adoption rate among French small and medium-sized businesses, 2024-2025.
(2) [symbolist, cas clients](https://www.symbolist.ai/cas-clients) : results published by the agency (Tribee, consulting firm, BEWE).
Publié le 01/09/2026