Feature Speed, Strategy Bottlenecks: The Future of Building Products With AI Pt. 3
At some point during development, something surprising happened:
I stopped wondering whether I could build Sunday Planning.
Instead, I found myself racing to keep the backlog ahead of the AI.
When Anthropic released Claude Code Web credits, I could run multiple AI agents in parallel against my GitHub repo—each writing code, testing it, updating branches, resolving issues, and proposing improvements. It felt like onboarding a small engineering team overnight (I was actually travelling in Paris when the web credits were coming up to expiry and would organize my day around being able to create enough tasks to consume all these credits… spoiler, I still wasn’t able to).
And then it hit me:
The limitation wasn’t engineering capacity.
It was how fast I could define clear, valuable features.
This wasn’t a development bottleneck—it was a strategic one.
The real constraint had shifted from “how long will it take to build this?” to “how quickly can we understand the problem, test solutions, and make good product decisions?”
When AI Outpaces Your Backlog
I had assumed the hardest part of building an MVP would be the engineering effort. For years, product managers have been conditioned to plan carefully because everything takes time. Work must be sequenced. Trade-offs analyzed. Roadmaps negotiated. Every feature has a cost.
But when AI can implement multiple features in parallel—sometimes overnight—those old constraints crumble.
Suddenly I had more engineering bandwidth than I had validated ideas.
The cost of exploration had collapsed.
Iteration cycles shrank from weeks to hours.
This forced a mindset shift:
The scarcity isn’t development anymore.
The scarcity is clarity.
If a feature was vague, AI struggled.
If a feature was well-defined, AI executed beautifully.
If the context was missing, AI made incorrect assumptions.
If the product thinking was strong, AI amplified it.
I wasn’t moving slowly.
I was thinking too slowly for the speed of the new system.
The New Shape of a Product Team
This experience made me reconsider the future of product teams—especially in SaaS.
There’s a narrative floating around that AI will shrink product and engineering orgs. But I now believe the opposite. In a world where implementation becomes faster and cheaper, product teams will grow, not shrink, because there is suddenly much more exploration possible.
AI is removing the friction around building prototypes, early features, A/B tests, and internal demos. So instead of evaluating one idea a quarter, companies can test ten—or a hundred. UI variations, onboarding flows, recommendation engines, pricing experiments, new features… all of it becomes easier to try.
That means product teams need to:
discover more opportunities
test more hypotheses
validate more ideas
run more user research
explore more UX directions
refine more workflows
define better standards and constraints
Engineering doesn’t go away—it becomes focused on integration, quality, robustness, scale, and long-term maintainability.
AI can generate code, but it won’t ensure security, data consistency, performance, or reliability under real-world conditions. That’s still where human engineers excel.
Meanwhile, PMs and designers get the freedom to explore, ideate, and shape experiences far earlier and more fully than before. AI lets them produce fully functioning models, not just mockups. That means engineers ultimately receive clearer, more validated work to implement—reducing ambiguity and rework.
This is the new loop:
Product teams explore widely.
AI builds quickly.
Users give feedback early.
Engineers scale what proves valuable.
It’s not that AI replaces teams.
It’s that AI changes what teams do.
The bottleneck becomes insight, not output.
Understanding, not execution.
Direction, not code.
The Real Future: Thinking Becomes the Competitive Advantage
Through this whole process, one truth became unavoidable:
AI doesn’t remove thinking.
It amplifies it.
If you’re clear, AI makes you clearer.
If your vision is strong, AI makes it tangible.
If your ideas are thoughtful, AI takes you further.
But if you’re vague, AI amplifies the vagueness.
If your ideas are contradictory, AI collapses under the contradictions.
If your product thinking is weak, AI exposes it faster than any human team ever could.
This, I believe, is the defining shift for SaaS companies:
As AI accelerates engineering, product management becomes the strategic center of the organization.
Not because PMs “write specs,”but because they shape the direction, define the user experience, and ensure the team is solving the right problems.
In the future, the companies that win won’t be the ones who build the most features.
They’ll be the ones who understand their users deeply, test ideas continuously, and iterate faster than anyone else.
And AI makes that possible.
Looking Ahead
For Sunday Planning, this means we can build a more delightful, intuitive travel tool—quickly and continuously informed by how people actually use it. A product that evolves at the speed of insight, not at the speed of sprint cycles.
AI won’t replace engineering teams.
It won’t replace product teams.
It won’t replace creativity, imagination, or vision.
But it will reshape how teams work together—accelerating the parts that used to slow us down and putting all the weight on the thing that matters most:
Clear thinking.
Customer understanding.
Product sense.
The future of building products won’t be defined by how fast we can code. It will be defined by how clearly we can think.
And that’s the future I’m building toward.
Part 1 and Part 2 of this series — available now.
Be part of our early community — sign up for updates and beta access to Sundays Online’s first release: Sunday Planning here.
