Why the AI Roadmap is Backward (and What Leaders Can Do About It)
Product teams are under pressure to prove they're "building with AI." Executives want to see AI
features on the roadmap, investors want to hear about AI capabilities in pitch decks, and
competitors are announcing AI integrations, so teams respond by adding AI features.
While this is a natural response, it’s also misguided. The best AI products don’t add features,
they eliminate workflows. When AI works properly, you don’t even need to brand it as “AI”
because it integrates so seamlessly into the product that it is no longer a distraction.
88% of organizations are now using AI, and nearly two-thirds remain stuck in the
experimentation or piloting phase.
The problem isn’t the technology—it's the approach.
We're treating AI like every other feature: something to add, ship, and hope users adopt. The
fundamental assumption is that users will adopt the feature simply because it offers AI.
But rather than treat AI as just an add-on, I believe teams should think of AI as a way to
streamline and expedite workflows. Perhaps the objective isn’t to add another step, but remove
a few.
Want more from The CPO Club?
Sign up for a free membership to complete reading this article:
Have an account? Log In
AI Integration: Effective AI should streamline workflows, not just add features that complicate user experience.
Workflow Elimination: Successful AI products focus on eliminating steps in processes rather than simply adding new functionalities.
User Adoption: Adding AI features rarely ensures adoption; removing friction is a more effective strategy.
Product Roadmap: Product teams must prioritize workflow elimination over merely adding AI components to enhance usability.
Competitive Edge: Focusing on seamless integration of AI can create a significant advantage over competitors emphasizing visible features.
The Race to Add AI
Leadership asks for AI features, so product teams audit their workflows and ask, where can we
add AI? They land on AI-powered search, smart suggestions, automated summaries, and
intelligent recommendations. These features get built, shipped, and added to marketing pages.
Then adoption stalls. The AI features sit unused, tucked into menus and settings panels. Teams
respond by adding more features, creating even more complexity and more things for users to
remember.
What Elimination Looks Like
Consider the steps involved in setting up a meeting. Teams send calendar invites back and forth
over email, join via a calendar link, take notes during the conversation, write a recap email
afterward, create action items in a project management tool, and schedule a follow-up meeting.
Most AI meeting products add a bot that joins meetings to take notes, a feature that generates
summaries, another that creates transcripts, and smart scheduling suggestions. Each feature
solves one piece of the problem, and each requires users to remember it exists and incorporate
it into their workflow.
An elimination approach works differently. The system already knows everyone's availability, it
infers the agenda, it captures summaries and transcripts automatically, it distributes notes,
action items, and schedules what's next. Attendees show up for the meeting and have
the conversation; the rest happens automatically. There are no AI features to point to because
the workflow disappeared along with the friction.
Building for Elimination
If teams are serious about building AI products that users will actually adopt, they need to
change how they approach the roadmap. Stop auditing the product for places to add AI and
start auditing where to remove friction.
When eliminating workflows, teams need to remove enough friction at once so that the new
experience is genuinely better than the old one. Half-eliminated workflows create confusion
about which parts are automated and which parts still require manual work.
This also requires more courage. Shipping AI features is safe because if users don't adopt
them, teams can say, it’s there if people want it. Eliminating workflows is risky because if the AI
doesn't work well enough, you've removed something users relied on. This is exactly why
elimination-focused products win; they force teams to make the AI genuinely reliable.
The Real Question
In your next team meeting, I would ask your colleagues: Are we building to prove we have AI, or
building to actually improve the product?
If the objective is to prove you have AI, teams will keep adding features. If the objective is to
improve the product, teams should ask what AI allows them to remove. This eliminates
interfaces, removes decision points, and deletes workflows, which in turn gives users less
Complexity.
While competitors are racing to ship roadmaps full of visible AI features, the real opportunity is
building products where users can't point to the AI at all. What can AI allow you to remove?
Stay in Touch!
Don't forget to subscribe to our newsletter for more product management resources and guides, plus the latest podcasts, interviews, and other insights from industry leaders and experts.
