Blog: Diagnosis
Your next AI opportunity might be hiding between two teams
Published by Forever Foundry
An issue gets logged. Someone copies it into another system. Another person chases an update. A manager asks what is happening. Nobody calls this an AI opportunity. They call it Tuesday.
The work around the work
On 10 September 2026, Microsoft published an account of its use of AI agents in network operations. Alongside troubleshooting, its teams targeted cross-system handoffs and outage communications. Microsoft reports that users of its Falcon Agent saved around 300 hours during its first few months, an early, self-reported result, not a forecast for other businesses.
Our reading: the useful question is not simply what AI can do. It is how much work people perform just to keep other work moving.
Look between departments
Consider a manufacturer chasing a quality decision before dispatch. Quality holds the evidence. Operations needs an answer. Customer service needs a delivery date.
The opportunity might be an AI-assisted update, an integration or a straightforward low-code approval workflow. The diagnosis should determine the solution, not the other way round.
Ask the people on both sides of the handoff: what are you waiting for, what do you re-enter, and what happens when the answer arrives late?
Measure the outcome, not the activity
Before building anything, record the handling time, waiting time and frequency of the problem. Agree an owner and the approvals any automated action must respect.
Then test whether the change reduces effort or delay without increasing errors. Time released is useful capacity; it only becomes a financial saving when the business can explain how that value will be realised.
At Forever Foundry, AI Compass™ uses conversational AI to gather those experiences across the organisation. Slid3rs™ helps your people rank the resulting opportunities by value and complexity.
The next investment need not start with another platform demonstration. It can start with the handoff everyone has learned to tolerate.
Next