Felix Grivot on AI Adoption Mistakes and Leadership
Felix Grivot analyzes the pitfalls of vibe coding and guides leaders on setting clear guardrails.
When examining AI adoption in business, Felix Grivot warns against teams spending excessive time building micro-solutions instead of using existing tools. Speaking on Marketing - The Podcast, he advises managers to establish clear guardrails and stay customer-focused.
Felix Grivot, Head of B2B Marketing at FCamara, discussed the primary mistakes companies make when implementing artificial intelligence during a conversation with Bruno Gimenes on Marketing - The Podcast. The full episode also explores other aspects of the intersection between marketing, strategy, and technology in the corporate landscape.
In the video above, Grivot addresses the recent fascination with vibe coding tools such as Lovable and Claude, warning about the time efficiency of teams building their own proprietary solutions.
The risk of building solutions instead of using them
For Grivot, the ease of generating code through AI platforms has led teams to spend hours developing features that already exist off the shelf. He notes that the unchecked creation of micro-solutions can distract from the core business objective.
"There's a saying my grandmother used to have: sometimes the sauce ends up costing more than the pasta. So sometimes a company, marketing, or managers encouraging their teams to build things with AI end up spending more time creating the solution than actually using it," Grivot explains.
Grivot emphasizes that before embarking on the internal development of any script or microservice, leadership must assess whether the project is genuinely scalable and whether accessible market solutions already exist. As examples, he cites automations for email verification or automatic meeting logging in CRMs—tasks that often already have dedicated tools.
Getting the basics right and focusing on customer needs
Another challenge highlighted by Grivot is the excess of micro-analysis driven by the volume of data AI makes available. He observes that over-indexing on technical details, such as complex attribution models, can pull professionals away from direct engagement with their audience.
"Good marketing basics come down to listening to the customer and looking at the data you already have available," explains Grivot. In his view, the accumulation of AI-generated information can trap teams in building ad-hoc tools rather than analyzing actual research and understanding buying behavior.
Grivot highlights the importance of pacing oneself amid continuous industry developments: "It's hard to separate the hype from the ROI. But we need to take a step back and balance that information so we truly invest our time wisely, pursuing solutions that are genuinely functional and can scale."
Guardrails and hands-on leadership in AI management
When advising managers looking to guide their teams on AI usage, Grivot advocates for establishing guardrails—clear security boundaries and guidelines that direct teams on what is and isn't permitted.
"As a manager, and as advice for other managers, I think the most important thing is creating guardrails so your team can navigate this path, because your team is going to go down it anyway," points out Grivot. He explains that these rules prevent severe issues, such as improperly uploading proprietary databases to free versions of platforms like ChatGPT.
Beyond governance, Grivot advises leaders to actively take part in technical learning and cultivate curiosity alongside their teams, rather than simply delegating tech exploration. He stresses that AI operates solely on existing data and knowledge, relying on the human element to innovate: "AI isn't creative; it works with what has already been done. Only humans have that potential."