Blog: Investment
AI began delivering value when the conversation moved to business performance
Published by Forever Foundry
On 17 September, Microsoft published lessons from its own internal AI transformation.
Its initial approach resembled a conventional technology rollout: provide the tools, train employees and encourage adoption. Microsoft describes licensing AI to more than 200,000 people, while finding that access alone did not change how work happened. In its early sales experience, usage plateaued and the expected impact did not materialise.
Microsoft changed its approach by starting with business outcomes and examining how people actually worked. It reports that adoption of selected use cases tripled in one sales group, alongside revenue per account manager 9.4% higher and close rates 20% higher.
These are internally reported findings. The sales figures compare groups with different levels of Copilot usage during January–June 2024; they do not establish that AI alone caused the difference or forecast results for other organisations. Read Microsoft’s account and methodology.
The FOFO™ view
For us, the strongest lesson is that deployment began producing value when the conversation moved from technology adoption to business performance.
A clear business objective gives people a reason to use AI, a way to decide where it belongs and a measure of whether it is helping.
That is why AI Compass™ starts by asking what should improve before asking how to use AI.
Imagine a manufacturer introducing AI to its quality teams.
A high adoption rate might show that people are generating reports or searching documents. It would not reveal whether quality decisions are faster, recurring faults are identified earlier or fewer orders are delayed waiting for approval.
Those outcomes depend on understanding the complete workflow.
Operators may know where information first goes missing. Quality teams may understand which exceptions require judgement. Operations may see the cost of waiting. Customer service may experience the consequences when decisions arrive late.
Listening across those functions changes the investment question from “How do we get more people using AI?” to “Which business problem is worth solving, and what role, if any, should AI play?”
Give every investment a measure that matters
Before funding an AI initiative, define the operational outcome it should change. That could be:
Shorter quality-release time
Fewer repeat failures
Reduced rework
More accurate planning
Faster customer response
Lower cost per successfully completed case
Record the baseline, identify who owns the outcome and agree how improvement will be measured. Include implementation and ongoing costs, then compare the opportunity with every other claim on the same budget.
The FOFO Process™ helps leadership teams make that decision. AI Compass™ listens across the organisation to uncover the business challenges and opportunities people can see. Slid3rs™ keeps prioritisation human, helping teams rank those opportunities by business value and delivery complexity. FlowRoom™ brings leadership together to agree what should happen first.
AI adoption can be a useful leading indicator. The measure of success is what becomes better for the business and its customers because people use it.
The same logic applies to prediction tools. Read why a better prediction is only valuable if you can act on it.
Next