“The next billion-dollar startup may not come from a Silicon Valley skyscraper, but from an apartment with three people and fifty AI agents.”
This statement, which would have sounded absurd five years ago, now describes a documented and measurable reality. We are going through a profound organizational shift—and product teams are on the front line.
Part 1 — Reality on the ground: tiny teams, massive outcomes
The numbers that change everything
The data is now public and undeniable. In 2025, several companies reached revenue levels once associated with organizations of hundreds of employees—with teams that fit into a single meeting room.
Cursor, an AI development tool, reached $200M in annual recurring revenue in 21 months with a team of about 20 people (Propulsion Studio, Dec. 2025).
Bolt.new generated $40M ARR in two months with 15 people.
Lovable, launched in 2024, reached $17M in two months with a similar team size (BBF Digital, Dec. 2025).
Midjourney remains the most cited case: $200M ARR in two years with just 11 people (BBF Digital, 2025).
These are not anomalies. They reflect a repeatable model. The emerging metric is no longer headcount, but revenue per employee—and this number is exploding in AI-native startups.
The “Tiny Teams” concept: from trend to movement
Shawn Wang (Swyx), a key figure in the AI developer ecosystem, formalized this phenomenon under the name Tiny Teams. His target metric is telling:
teams generating more millions in revenue than they have employees.
The core insight: in an era where cognitive work can be augmented and automated on demand, trust and human communication become the main bottlenecks—not technical execution (Daniel Bentes, Medium, July 2025).
What this concretely changes for a 5-person product team
A Core Innovation Capital study estimates that in 2025, an engineer using AI tools can produce as much as a team of five traditional developers.
Applied to a team of five, this represents the potential output of a 25-person organization (Bonsai Labs Dispatch, Nov. 2025).
Where AI removes size constraints:
Code: Tools like GitHub Copilot and Cursor allow a single developer to generate and maintain what once required a team.
Customer support: IBM reports show AI can handle up to 80% of routine support tasks.
Marketing & growth: Platforms like HubSpot AI automate campaigns, segmentation, and reporting.
Product research: AI analyzes user feedback, detects patterns, and generates insights in minutes.
Documented result: 77% of SMEs using AI report productivity gains, with nearly half seeing results in under three months (Massify Online, 2025).
🔎 A week in a 5-person tiny team — what AI-native work really looks like
AI transformations are often described through abstract promises: productivity gains, automation, faster cycles.
But what does this actually mean, day to day, for a product team?
Let’s take a 5-person team:
1 Product Owner / Business Analyst, 2 developers, 1 designer, 1 growth lead—working with an integrated AI stack (Cursor, ChatGPT, Notion AI, Intercom AI, Amplitude).
This is not a projection. It is an operational reality.
Monday — From raw data to decisions
Traditionally, analyzing user feedback is slow, fragmented, and manual.
With AI, this step is radically transformed.
User feedback is synthesized in minutes, trends emerge instantly, and the backlog can be prioritized continuously.
👉 The Product Owner’s role shifts:
less time structuring information, more time making decisions.
Tuesday — Amplified technical execution
Development is no longer limited by individual coding capacity.
AI tools generate functional foundations quickly, while human expertise focuses on quality, architecture, and integration.
👉 The shift: from production to validation.
Wednesday — AI detects, humans understand
AI can process large-scale user feedback: categorization, prioritization, pattern detection.
But qualitative validation remains essential.
User interviews are still critical—not to collect data (already abundant), but to interpret what the data doesn’t explicitly say.
Thursday — Compressed design cycles
Design, once sequential, becomes iterative and near-instant.
Generating multiple variations allows rapid testing of hypotheses, drastically reducing time from idea to validation.
👉 The designer becomes an orchestrator of rapid experiments.
Friday — Continuous growth loop
Growth activities now operate in a short, continuous cycle:
content generation → testing → measurement → adjustment.
AI accelerates execution, but value lies in interpreting results and deciding what comes next.
Part 2 — How AI is transforming product teams (and roles)
The “K-shape” of the product job market
The product job market is becoming increasingly polarized. Analysts describe a K-shaped split: on one side, strong demand for AI PM profiles with solid AI expertise; on the other, a gradual disappearance of traditional generalist roles (Agents Today, August 2025).
This shift is already measurable: in 2025, two-thirds of business leaders say they would not hire a candidate without AI skills, and 71% would prefer a less experienced profile with strong AI capabilities (Microsoft Work Trend Index, cited in Agents Today, 2025).
⚠️ Strong signal: mastering AI tools has moved from a competitive advantage to a baseline requirement for any product professional.
What is changing in the daily work of a BA/PO
What AI now does (very) well:
- Drafting initial versions of PRDs (Product Requirement Documents)
- Synthesizing user feedback at scale
- Analyzing data and identifying behavioral patterns
- Generating user stories from interview transcripts
- Producing technical documentation and writing specifications
What remains irreplaceably human:
- Reading between the lines during user interviews
- Negotiating priorities with stakeholders who have conflicting agendas
- Defining product vision and making strategic trade-offs
- Building trust with the development team
- Detecting unexpressed needs — “problem finding”
“Replacing PMs with LLMs is about as likely as replacing accountants with spreadsheets. Tools do part of the job, but only humans understand context and are accountable.” PM veteran, cited in Product School, 2025
The rise of the “AI PM” profile
A new hybrid role is emerging: the AI Product Manager. Neither purely technical nor purely business-oriented, this profile embodies the convergence between product strategy and mastery of AI models.
Key skills:
- Understanding the fundamentals of AI models to assess technical feasibility
- Designing product strategies that leverage AI capabilities
- Writing PRDs for model-based features
- Mastering AI model evaluation frameworks
- Acting as a bridge between data/ML teams and business stakeholders
Salary trends reflect the growing demand: in the United States, senior AI PMs earn between $150,000 and $200,000 per year, which is 30–40% higher than traditional PMs (Product School, Eleken, 2025–2026).
Part 3 — The future of product work: fewer people, more impact?
The real question is not “how many?” but “how?”
Organizations that succeed with small, AI-augmented teams are not simply “doing less with less” — they are doing more, differently.
The “10/100/3” framework proposed by Benhamou Global Ventures illustrates this trend: startups reaching $100 million in revenue with 10 people in 3 years. What was considered science fiction in 2020 has become a serious benchmark in 2025 (Bonsai Labs, Nov. 2025).
What characterizes these organizations:
- Faster decision-making: fewer hierarchical layers, less internal politics
- Accelerated iteration: AI compresses prototyping and testing cycles
- Extreme focus: team size constraints force radical prioritization
- Short feedback loops: founders stay close to users
The real limitations that must be acknowledged
The enthusiasm around tiny teams should not obscure real constraints:
- The risk of burnout is real. When a team of 5 must cover the responsibilities of a team of 20, the cognitive load still falls on human shoulders.
- The talent bottleneck. The pool of profiles capable of working at the intersection of product, data, and AI remains limited.
- Tool dependency. An ultra-lean team relying on a complex AI stack is vulnerable to outages and pricing changes.
- The erosion of junior learning. 54% of engineering leaders expect to reduce junior hiring due to AI, creating a systemic risk for the talent pipeline (Agents Today, 2025).
What this means for your career trajectory
If you are a BA, PO, or Product Manager in 2026, here is the clear signal the data is sending:
- AI will not replace PMs — it will replace PMs who don’t use AI.
- Your differentiating advantage will increasingly lie in the strategic and relational layer: vision, alignment, deep user understanding.
- AI expertise is no longer a specialization — it is a baseline skill within a 3–5 year horizon.
- Small teams will continue to gain market share in domains where agility outweighs economies of scale.
Conclusion — What I observe in the field
Working with product teams of very different sizes — from AI-native startups to large organizations in the healthcare and energy sectors — one observation stands out: the highest-performing teams are not the largest.
They are the ones that have rethought how they work around AI.
The tiny team is not a dogma. It is a strong signal: value is no longer created by accumulating resources, but by combining human judgment with AI’s execution power.
In regulated sectors such as healthcare, energy, and finance, this reality is even more pronounced.
AI accelerates, structures, and optimizes — but responsibility, decision-making, and vision remain deeply human.
So the question is no longer: “Can you build a product with 5 people?”
The answer is already known: yes.
The real question lies elsewhere:
are you developing the skills that will allow you to be one of those 5 people?
Sources and références
- Bentes, D. (2025, juillet). The Tiny Teams Revolution. Medium.
- Vlasiu, A. / Bonsai Labs Dispatch (2025, novembre). The Next 10-Person Startup Is Actually a 3-Person Team + 50 AI Agents. Medium.
- BBF Digital (2025, décembre). The Tiny Team Revolution Changing the Face of Startups.
- Propulsion Studio (2025, décembre). Tiny Teams, Big Impact.
- Massify Online (2025). Tiny Teams, Big Outcomes: How AI and Partnerships Enable Startup Growth in 2025.
- Your Coffee Break (2025, septembre). The Rise of AI-Powered Tiny Teams.
- World Economic Forum (2024, août). How AI Brings Corporate Might to Small Teams.
- Product School (2025). Will AI Replace Product Managers? & AI Product Manager: Real Role or Buzzword?
- Agents Today (2025, août). The Great Reshuffling: How AI is Polarizing Product Management Roles.
- Product Leadership Institute (2025). AI Transformation in 2025.
- McKinsey & Company (2025). The State of AI 2025.
- Microsoft (2025). Work Trend Index Report.
- Eleken UX Design Agency (2026). How to Become an AI Product Manager in 2026.
💬 What about you?
Is your team ready for this paradigm shift?
How much of your work do you currently delegate to AI — and more importantly, how much more could you delegate?
👉 Share your experience — this topic deserves a serious discussion.
On 32steps.com, I help product teams integrate AI into their workflows — from use case definition to measurable impact.
My approach is based on two pillars: continuous learning and close collaboration with teams.
The goal: AI that is useful, concrete, and aligned with your business challenges — far from hype-driven trends.
I would be delighted to connect with you, especially to hear your feedback on Tiny teams + AI.
You might also be interested in this article:
“Managing Change in Business: Practical Methods and Insights“
Feel free to share this article on your social networks.
If you have any questions, contact us by email — we will be happy to get back to you.
Follow our Pinterest page for visual inspiration and to discover new ideas.
