.svg)
.svg)
AI Isn’t a Chatbot Problem
Many AI products start with a chat interface. In reality, most users don’t want another place to have conversations. They want help completing tasks, making decisions, and reducing effort within the tools they already use.
.svg)
.svg)
AI Isn’t a Chatbot Problem
When generative AI exploded into the mainstream, almost every product team reached the same conclusion:
“We need a chatbot.”
New chat panels appeared everywhere.
A floating assistant in the corner became the default AI strategy.
But after working on AI-powered experiences in operational software, I’ve become increasingly convinced that most organisations are solving the wrong problem.
The challenge isn’t introducing another conversation.
The challenge is helping users get work done.
For many users, chat is simply a mechanism. It isn’t the destination.
Once the foundation is set, the focus shifts to design, iteration, and execution. Prototypes are tested, feedback is analyzed, and features are refined until the product becomes something intuitive, functional, and meaningful.
Design AI Around Outcomes
When designing AI experiences, I try to avoid asking:
“What can the model do?”
Instead, I ask:
“What outcome is the user trying to achieve?”
Sometimes the answer is generating a response.
Sometimes it’s summarising a long history.
Sometimes it’s identifying risk.
Sometimes it’s recommending the next action.
In many cases, users don’t need a chatbot at all.
They need a faster way to complete a task.
The most effective AI experiences I’ve worked on don’t feel like separate products.
They feel like a natural extension of the workflow itself.
The future of AI isn’t a chat window sitting beside the product.
It’s intelligence embedded directly into the moments where users make decisions.
Users Don’t Want Another Tool
Think about a Facilities Manager handling hundreds of maintenance tasks.
Or a Customer Service Agent responding to residents.
Or an Operations Manager reviewing service performance.
Their day is already fragmented across systems, notifications, emails and meetings.
Adding another destination where they need to ask questions often creates more work rather than less.
The real opportunity for AI is understanding the context users are already working within.
Instead of asking users to explain a problem to an assistant, the system should already understand:
- The task
- The asset
- The history
- The stakeholders
- The current status
The less context users need to provide, the more valuable the AI becomes.
Similar blogs
.svg){kind=small}
.svg)
LET’S MAKE CLARITY A HABIT.
Have a product, platform, or operational challenge in mind?
I help organisations simplify complexity, create scalable foundations, and build digital products that deliver lasting value.
