Don’t Start With the Tool
One of the most common mistakes organisations make with AI is also the simplest: They start with the technology.
You hear it in boardrooms and leadership meetings: “We want to do something with AI.”
It sounds progressive. It signals ambition. But it’s the wrong starting point.
AI is not a strategy. It’s a capability. And like any capability, it only creates value when applied to a clearly defined business problem.
Why “Let’s Do Something With AI” Is a Red Flag
When leaders begin with the tool, the next steps usually follow a predictable path:
A pilot is launched because the technology looks promising
A small team experiments
Something interesting is built
The business struggles to see why it matters
Eventually, the initiative stalls or is quietly deprioritised. Not because the technology failed, but because it was never anchored to a meaningful outcome.
In practice, this leads to:
Wasted effort
Fragmented experiments
Low adoption
Poor ROI
Growing internal scepticism
Over time, the narrative shifts from “AI will transform us” to “AI doesn’t really help here.” The reality is more uncomfortable.
AI didn’t fail. The framing did.
What Happens Inside Organisations
What we often see inside organisations is this:
A leadership team approves an AI pilot because competitors are talking about it
A vendor demonstrates a compelling product
There is pressure to move
But no one has clearly defined the operational friction the tool is meant to remove.
Six months later, the tool exists, but workflow hasn’t changed. Managers aren’t accountable for using it. The output isn’t tied to performance metrics. The experiment becomes optional rather than essential.
The friction point wasn’t solved, because it was never clearly identified.
The Problem-First Alternative
A more effective approach starts somewhere far less exciting:
Where are we underperforming?
Where are costs increasing without clear value?
Where are teams spending time on low-leverage work?
Where are decisions inconsistent or slow?
These are business questions, not technology questions.
Only once the problem is defined should AI enter the conversation, because:
Sometimes AI will be the right solution.
Sometimes automation will suffice.
Sometimes the issue is process, not technology at all.
The discipline of defining the problem first protects from novelty-driven investment.
A Simple Leadership Filter Before Any AI Investment
Before approving any AI initiative, leaders should ask three questions:
What specific business outcome are we trying to improve?
How will we measure whether it improved?
If this works, what changes in daily behaviour?
If those answers are vague, the initiative is premature.
Organisations that generate real value from AI are rarely the ones building the most impressive demos. They are the ones solving the most practical problems.
The most powerful AI projects often look unremarkable from the outside. But internally, they remove friction, reduce cost, improve quality, or speed up decisions.
That is what moves the business.
AI does not need to be “cool” to be valuable.
It needs to matter.
At Amity Insights, the focus is simple: AI guidance without hype, practical, strategic, and grounded in real organisational outcomes.