This article explores the evolution of content management and the emergence of "agentic" web development, based on a conversation with Aurélien Georget, co-founder of the industry-standard headless CMS, Strapi, and founder of the AI-native website builder, FIMO.
How Strapi Was Born: From Student Project to Enterprise Standard
Aurélien Georget began his journey in 2013 as a student at a master’s degree program, where he met his co-founders. Initially working as freelancers building websites, they noticed a significant gap in the market: as mobile applications grew in popularity, the existing toolset—specifically WordPress—could no longer handle the complexity of managing content across both web and mobile platforms effectively.
They built their own solution, which eventually became Strapi. What started as a student project on GitHub is now a powerhouse used by major global corporations, including Apple, Adidas, and IKEA. After 10 years of scaling the company, raising $50 million, and managing a remote team of 70, Georget recently launched FIMO, an AI-native website builder designed to address the next phase of the web’s evolution.
Why Launch FIMO Now? The Case for Starting Over
When considering how to integrate AI, Georget faced a strategic dilemma: should he force AI into the decade-old "cargo ship" that is Strapi, or build a new, AI-native competitor?
Because Strapi supports millions of users and high-stakes corporate contracts, rapid architectural changes are difficult. Georget concluded that the market window for AI-native tools is open right now. Starting FIMO from scratch allowed his team to learn, iterate, and potentially apply those insights back to Strapi, while avoiding the "tech debt" of retrofitting an existing platform. In just six months, a team of five people managed to build a product that addresses what Georget identifies as the fundamental gap in current AI coding tools: long-term maintenance.
The Agentic Web: Building for Agents, Not Just Humans
Georget proposes a radical shift in how we perceive the internet: "The web was mostly built by humans for humans. The web, I think, is going to be built by agents, mostly for agents and humans."
In this new "agentic" era, AI coding agents—such as Cursor or Claude—will not just write the initial code; they will maintain it. They will proactively update content, add pages, and adjust strategies in response to analytics or competitor movements. FIMO is being built to provide the "first generation of autonomous, proactive, and intelligent websites."
When Content Management Breaks
Georget explains that the breaking point for content management happens at scale. When a website grows from 10 pages to 100,000, human-led maintenance becomes inefficient. It becomes nearly impossible to ensure consistent tone, branding, and accuracy across every page.
He uses an illustrative thought experiment: Imagine a future where an autonomous agent managing Apple.com decides to remove all products from the homepage except for AirPods during the Christmas season because data shows that is the highest-margin product for that specific window. While some might find this "sad" from a human creative perspective, for a business, it is an optimized, data-driven decision. Georget argues that companies will eventually delegate routine content management to agents while reserving human oversight for core brand identity and high-stakes creative decisions.
FIMO vs. Lovable: Verticalization as a Moat
While many AI tools like Lovable or Replit allow users to generate websites, Georget notes that they are primarily "app builders" rather than "content management platforms."
FIMO differentiates itself by providing a structured content layer:
- Structured Data: Unlike static code generators, FIMO stores content in a database with a robust interface, enabling rich text, media libraries, and multilingual support.
- Context Awareness: Because the content is structured and external to the code, agents can "reason" about it, making it easier to perform global updates or SEO adjustments.
- Verticalization: FIMO is optimized specifically for marketing and corporate websites, providing a specialized environment that general-purpose AI coding tools aren't designed to handle.
What Does AI-Native Really Mean?
For Georget, being AI-native is a mindset that changes how products are built:
- No Blank Pages: Every interaction starts with an AI-generated suggestion, eliminating the friction of starting from scratch.
- Vibe-Coded QA: Instead of filing tickets in tools like Linear, engineers and designers use AI to "vibe-code" fixes instantly. This reduces the cycle time for small UI adjustments from weeks to minutes.
- Lean Innovation: FIMO was built by a small, isolated team that operated without traditional constraints, such as separate support or documentation departments. Much of this work was automated by AI, allowing the team to move at unprecedented speeds.
Advice for Founders and PMs
Georget offers three key takeaways for those building in the AI space:
- Don't Over-Optimize Economics Early: Ignore margins in the beginning. Focus entirely on usage, traction, and finding the right hook.
- Protect Innovation: If you are building within a larger company, separate the AI team from the main organization. Give them their own systems, slack channels, and workspace to prevent distraction.
- Automate the Role of the PM: Product managers should stop writing user stories and managing backlogs. Instead, they should maintain a "suitcase full of agents" to automate administrative tasks, leaving them more time to develop "product sense," watch competitors, and truly understand the market.
Ultimately, Georget believes we are heading toward a web where content is generated by agents, posing new questions about trust and authenticity. However, he remains optimistic that this shift will simply redistribute human effort toward more creative, high-value tasks, allowing us to move away from the screen-bound drudgery of manual maintenance.