The Engineering, Product, and Design (EPD) model was supposed to be a better way to structure teams. And it was, for a moment.

But AI is already rewiring what each of those roles actually means. The convergence isn’t a distant prediction. It’s happening at different speeds depending on one thing: company size.

Here’s what the next two years look like across the spectrum and the concrete moves to make now if you want to stay ahead of it.

Timeline: Already here

At small companies, the EPD generalist is not a future role. It’s a survival requirement.

A two-person team using Cursor, Claude, and v0 is already doing what a five-person cross-functional team did in 2022. The PM layer at this scale is functionally gone, either absorbed into the founding engineer or handled by AI-assisted research loops.

What’s changing for Engineers: You’re expected to own the full product loop - discovery, build, ship, learn. Writing code is the smallest part of the job. If you’re waiting for a PM to tell you what to build, you’re already behind.

What’s changing for EMs: The EM title at a small company is rapidly becoming “GM of a small business unit”. You own outcomes, customer metrics, and team health, not sprint velocity. If you’re still measuring success by delivery cadence, you’re measuring the wrong thing.

Timeline: 2026-2027 - accelerating now

This is where the most meaningful structural change is happening. Mid-size orgs are big enough to have established EPD roles, small enough to feel the economic pressure to consolidate them.

The pattern showing up right now: EPD micro-pods. Small, high-leverage teams of 2-3 people replacing what used to be a full cross-functional team of 8-10. Each person carries more surface area. AI handles the coordination and execution overhead that used to justify headcount.

The PM role is compressing first. The parts of PM work that created real value - synthesizing user research, writing specs, stakeholder alignment - are increasingly AI-assisted or absorbed by engineers with good product sense. The dedicated PM who adds value at this layer is the one operating strategically, not administratively.

Design is shifting upward. Execution-level design (producing UI assets, iterating on variations) is increasingly AI-generated. The designers who thrive own design systems, interaction principles, and the deep question of how humans actually behave - not production work.

What’s changing for Engineers: Full-stack is the floor, not a differentiator. The new bar is understanding enough product and system design to make good architectural decisions without a room full of stakeholders. The engineers who will own their careers in this window are the ones who can go from customer problem to shipped solution with minimal handoffs.

What’s changing for EMs: Your headcount assumptions are stale. An AI-augmented team of six can output what a twelve-person team did twelve months ago. That changes how you plan, how you staff, and how you measure productivity. If you’re not recalibrating capacity models regularly, you’re making resourcing decisions on bad data.

Timeline: 2027-2029 and beyond, but the baseline is rising now

Large orgs will be the last to restructure. Specialization survives longer at scale because the cost of getting distributed systems, ML infrastructure, or security wrong is too high to leave to generalists.

But the baseline expectation for every role is rising fast.

Engineers are expected to have stronger product instincts. PMs are expected to understand technical feasibility at a depth that wasn’t required before. Designers are expected to understand AI behavior and its implications for UX - not just how to design for it, but why humans respond to it the way they do.

The new specializations emerging above that baseline: AI system design, human-AI interaction, and trust and safety. These aren’t niche research areas anymore. They’re production engineering problems showing up in every product team.

What’s changing for Engineers: Deep technical specialization still has a home at large scale but isolation from product context doesn’t. The specialist who understands their system and can reason about what the customer actually experiences is significantly more valuable than the one who only goes deep.

What’s changing for EMs: Managing AI-augmented teams requires rethinking what healthy performance looks like. Output metrics are less reliable when AI amplifies individual productivity unevenly. The EMs who lead well in this window are shifting focus to outcomes, decision quality, and team learning - not tickets closed.

For Engineers:

For EMs:

The EPD convergence isn’t a clean wave that hits all companies at once. It’s a gradient, already here at a small scale, arriving at mid-size in the next 18 months, and reshaping the baseline expectations at large orgs through the back half of the decade.

The roles aren’t disappearing. The scope is expanding and the handoffs are collapsing.

The engineers and EMs who thrive in this window are the ones who stop waiting for the job description to catch up and start operating at the level the role is becoming.

This wee, have a 15-minute conversation with a real customer - not a PM summary, not a support ticket. Try these three questions:

“What frustrates you most about the product right now?” Don’t defend or explain. Just listen.

“Are there any tools you use alongside ours to get the job done?” This surfaces the workarounds your team doesn’t know exist.

“What would make you recommend this to a colleague without hesitation?” This is the gap between where you are and where you need to be. It’s also the clearest signal of what your team should be building next.

Want something covered? Hit reply and tell me. I love hearing what you’re dealing with.

📬 Reply back to this email to book a 30 min session (free for subscribers!)

The roles are shifting. The expectations are expanding. But the engineers and leaders who stay curious, keep asking the “why”, and build with both speed and heart ❤️, those are the ones who will shape what comes next.

Remember, your code might run, but clarity scales!

Until next time,

Sabeen

P.S.

If you’re new here, welcome, grab your virtual beanbag and settle in. And whether you’re new or a returning reader, feel free to share this with a fellow builder who loves coffee-fueled “how to run Engineering” arguments (or discussions 🤪). The EPD debate alone should keep you busy for a while.☕💡