Speed, Trust, and Feedback: The New Rules of Product at Strapi

In fast-growing product teams, there comes a moment when the playbook stops working. That moment hit Strapi hard.

Strapi—the open-source headless CMS used by developers and brands worldwide—was moving fast but feeling slow. The ceremonies, frameworks, and best practices that once brought structure were now bottlenecks. Feedback felt siloed. Teams were operating, but not always learning.

So they changed everything.

We sat down with Aurélien, co-founder and CPO of Strapi, to understand how they rebuilt their product org from the ground up: from ditching rigid processes to scaling a culture of feedback with AI.

Escaping the process trap

"We used to follow the book. Sprints, retros, estimations, ceremonies… all of it, but over time, it became clear: we were spending too much time talking about work, and not enough actually doing it."

To regain speed and autonomy, Strapi overhauling how their product teams work. Not all squads have fully adopted the new model yet, but the shift is underway. They eliminated time-consuming rituals like effort estimations, dropped fixed squad structures, and stepped away from traditional Scrum altogether. In their place, they introduced lean, flexible micro-teams inspired by Stripe’s model. Each feature is now owned end-to-end by an Engineering Lead, while PMs and designers support multiple features across squads. This new setup reduces handoffs, boosts ownership, and dramatically accelerates delivery.

But this shift only works with experience.

“You can only break the rules once you understand them. That’s why we only run this model with senior profiles.”

Building product sense at scale

To help new PMs ramp up, Strapi built a comprehensive Product Management Framework—not as a rigid playbook, but as a flexible guide. The doc lays out every step of their product process across discovery and delivery phases, defines responsibilities across roles, and outlines key rituals, from scoping problems to validating solutions to communicating releases.

Inspired by BlaBlaCar's Discovery Discipline
Highlights when to involve stakeholders, what documents are expected, and where validation is needed
Provides adaptable tracks for small, medium, and large features

“We want new PMs to have structure when they need it… and freedom once they don’t.”

The framework is meant to be outgrown. Once a PM develops deep product sense, they’re encouraged to deviate and adapt. Until then, it’s a shared language to ensure quality, clarity, and collaboration at scale.

Here’s a glimpse into how it works:

The process is split into Discovery (Scope, Immerse, Inspire, Shape, Design, Expose) and Delivery (Spec, Develop, Release, Follow-up).
Each step includes clear goals, expected deliverables (like initiative cards, prototypes, user stories), and responsible roles.
The framework adapts to feature size—small features skip some steps, large ones follow the full 10-step process.
Everything lives in a Notion card that acts as the single source of truth.

It’s both structured and pragmatic—a strong onboarding ramp for newcomers and a common language to align cross-functional teams.

Turning feedback into a reflex across the org

Before Cycle, feedback lived in silos. The product team had to hunt it down manually—interviews, surveys, tickets. They even tried other tools on the market, but none were adopted. It was slow, biased, and incomplete.

Now, feedback flows from everywhere.

A single Slack channel accepts feedback in any format (text, screenshots, videos)
Cycle automatically ingests, tags, and routes it
PMs triage and follow up

"I don’t even go into the tool—I just read the digest. It saves me hours each week."

Today, support, partnerships, and even Discord messages contribute to product thinking. The org sees more signal, earlier, and from more diverse sources.

How Strapi uses AI to move faster

Strapi didn’t just embrace AI—they embedded it deeply into their workflows to maximize velocity and focus.

Cycle automatically ingests, summarizes, tags, and routes feedback from Slack and other sources, making feedback instantly actionable.
GPT is Aurélien’s go-to assistant to break the blank page—from drafting strategy memos to preparing workshops.
V0 and Lovable help prototype interface ideas quickly, turning product vision into interactive mockups without waiting on design cycles.
Dust powers custom AI agents to moderate plugin submissions: checking naming, dependencies, and even flagging vulgar descriptions.
Super gives the team daily Slack recaps to stay on top of conversations without drowning in threads.
Actionable ties together NPS data, CRM info, and product usage metrics to surface at-risk customers and identify friction points.
Gong is used to monitor sales conversations and extract product-relevant insights from customer calls.

“What took 5 days now takes 10 minutes for documentation by using a lot of customer Claude Scripts.”

What can you steal from Strapi’s product org?

Strapi’s shift to a leaner, feedback-first org wasn’t just about speed. It was about trust. Trusting senior PMs to own their domain. Trusting teams to learn by doing, not over-processing. Trusting that the right culture would bring the right signal to the surface.

Here’s what stood out:

Less process = more product. Kill what slows you down.
Train with structure, scale with autonomy. Build Product Sense and let people break the rules.
Make feedback a reflex. Plug it into team rituals, not just the product function.
Let AI do the grunt work. Summarization, prioritization, detection—leave that to machines.
Empowered PMs ship better products.

“Cycle helps PMs do their real job. Talk to customers. Understand the market. Spot patterns. Prioritize smartly.”

And that’s how you turn feedback into a growth engine.