“We need AI in the product.” Now what?
Over the past year I’ve heard this same sentence from B2B organisations:
“We need AI in the product.”
Sometimes it comes from the CEO. Sometimes the board. Occasionally it appears halfway through a roadmap discussion. And to be fair, the pressure is understandable. When competitors are announcing AI features and the industry conversation is dominated by AI breakthroughs, it’s hard not to worry about being left behind.
But there’s often an interesting gap in these conversations. The request to introduce AI frequently arrives before anyone has answered a much more important question:
Where does AI actually create value for the user?
Right now AI is everywhere in product marketing. Landing pages, feature lists and product announcements are full of AI assistants, copilots and automation claims. But when every product says it has AI, the technology itself stops being differentiation.
This leaves product leaders in a difficult position. They recognise the opportunity AI represents, but they also know that adding AI without a clear purpose can easily create complexity, unreliable experiences and confused users. The challenge isn’t simply introducing AI into the product. The challenge is deciding where it genuinely belongs.
Of course, you already know this, but how do you tackle this very real problem many product leaders are facing right now?
Why companies feel pressure to add AI to their product
Let’s first identify where the pressure is coming from. Several forces are pushing organisations to introduce AI quickly.
Competitive signalling. Product announcements increasingly include AI-powered features, even when the underlying value is unclear. Companies want to demonstrate they are evolving with the market.
The pace of the AI narrative. New models, tools and productivity claims appear almost weekly. This creates a psychological sense that decisions need to happen quickly.
Fear of strategic irrelevance. Many leadership teams are quietly asking a deeper question: could AI eventually mimic what our product does, making our product irrelevant?
That concern is legitimate. However, it is very different from simply adding AI features to the roadmap. Responding strategically requires stepping back and thinking about where AI actually improves outcomes.
So what do you do?
The pressure is real. You’ve explained to the board the potential downsides of simply trying to find a place to “include AI” in the product. But despite your efforts, the response is still the same: “We need AI in the product”.
So here are some avenues you can explore
1. Find value for the user
Start by asking a different question:
If automation were part of this product, where would it meaningfully improve the outcome for the user?
Working backwards from the outcome helps teams focus on value rather than technology. It encourages product teams to consider questions such as:
What task becomes significantly easier?
What analysis or decision becomes faster?
What friction disappears from the workflow?
What work no longer needs to be done manually?
An interesting opportunity appears when AI improves the core workflow of the user.
We’re working on a product where AI isn’t being positioned as a visible feature or a marketing headline. There’s no chatbot and no flashy assistant. Instead, AI is being used to quietly remove a task from the user’s process. Previously, users had to manually analyse and interpret certain information. Now the system performs that step automatically in the background. The user doesn’t need to interact with AI directly - they simply experience a product that feels more capable and efficient. This kind of implementation often delivers far more value than highly visible AI features. The product becomes smarter without forcing users to change how they work.
When AI is introduced at this level, it tends to feel like a natural extension of the product’s capability rather than a feature added purely for visibility.
2. Find value for the team
If AI doesn’t meaningfully improve the user experience, that doesn’t necessarily mean it has no value. It may still improve how the organisation builds and evolves the product.
Across many product teams, AI is already helping to:
accelerate design exploration and prototyping
support engineering workflows
analyse product data and customer feedback
surface patterns in large datasets.
We’ve been using AI to accelerate parts of the process, for example during the prototyping stage to accelerate the journey from idea to working concept. Early interaction patterns and design directions can be explored much faster, allowing teams to generate and iterate prototypes quickly. Ideas become tangible sooner, user testing happens earlier, and product decisions can be made with more confidence.
We’re also exploring how AI can speed up parts of the build process itself to allow developers to move from the Figma prototype to production faster, which shortens the feedback loop between design, engineering and product teams.
Users may never see this AI directly, but they benefit from it through better products being developed and improved more quickly.
The trust challenge in AI-powered products
From a design perspective, one of the biggest considerations with AI is trust.
When we observe people using complex digital products, one pattern appears repeatedly: users care far more about reliability than novelty. AI can deliver powerful capabilities, but it can also introduce uncertainty if outputs are inconsistent or difficult to verify.
For professional software in particular, losing user trust can be more damaging than not introducing an AI feature at all. This is why many successful implementations focus on augmenting the user rather than replacing them. AI surfaces insights, suggestions or automation while still allowing the user to remain in control.
5 step approach to respond to the pressure
In some organisations the direction is already set: AI needs to be part of the product narrative.
At that point the role of the product leader shifts. Instead of debating whether AI should exist in the product, the focus becomes ensuring it is implemented in a way that creates genuine value.
A pragmatic approach can help:
Prioritise removing effort from the user. The most valuable AI implementations remove work from the user’s workflow. If a task that previously required manual analysis or repetitive effort can be handled by the system, the product becomes more capable without forcing users to change how they work. But remember, particularly for complex B2B tools, users care far more about reliability than novelty. AI should support human decision-making and produce outputs users can understand and verify.
Avoid building AI purely for visibility. Highly visible features such as chat interfaces or assistants can signal innovation, but they are not always where the real value lies. The most effective uses of AI are often the quiet ones improving the product without demanding attention.
Find value for the product team. AI doesn’t always need to appear in the product to deliver value. Faster prototyping, improved analysis of product data and more efficient development workflows can dramatically shorten the learning cycle for product teams.
Start with contained experiments. Test a focused use case before committing AI across the roadmap. Rather than introducing AI as a large platform capability or headline feature, start with a small, well-defined problem where automation could meaningfully improve the outcome. More on this in our other article.
Final thought
Many product leaders are currently navigating two competing pressures. On one side, they want to build thoughtful products that genuinely improve how people work. On the other, organisations are responding to market signals, investor expectations and competitive announcements about AI.
That tension is real.
The leaders who handle it well are unlikely to be the ones shipping the most AI features. Instead, they will focus on identifying where AI removes friction, improves workflows and strengthens the core value of the product.
In other words, the real challenge isn’t simply adding AI to the product.
It’s deciding where AI actually belongs.
FAQs
-
A: Look for workflows that involve repetitive analysis, interpretation of large datasets or time-consuming manual steps. AI tends to create the most value when it removes cognitive effort or surfaces insights that would otherwise take users significant time to find.
-
A: Many AI features are introduced because of market pressure rather than user need. When teams start with the technology instead of the outcome, the result is often a feature that feels impressive in demonstrations but adds little value to everyday workflows.
-
A: AI can significantly accelerate design exploration, prototyping, engineering workflows and product analytics. By shortening the feedback loop between design, engineering and product teams, AI can help organisations learn faster and improve products more quickly.
-
A: Start by reframing the conversation around outcomes rather than technology. Instead of asking where AI can be added, identify where automation could meaningfully improve a user workflow or remove effort from a task. If the organisation is committed to introducing AI, focus on small experiments and real improvements rather than highly visible features.
Need help navigating the pressure to introduce AI into your product?
We help B2B product teams explore where AI genuinely improves workflows, protects user trust and strengthens the core value of the product - without adding unnecessary complexity.
👉 Book a call with our team to talk about how we can help you introduce AI in a way that actually improves your product.
