AI Spend Needs a Business Case, Not Just a Budget
19m
Connect Spending to the Work That Matters
AI spend tells only part of the story behind faster software development. Teams also need to know what that spending produced and why the work mattered. More code, more tokens and more activity do not automatically demonstrate business value.
Tempo Software Chief Product Officer Kevin Nanney explores that distinction with Mike Vizard. Nanney argues that building software has become easier, while deciding what to build has become harder. Leaders need better visibility into how human effort and AI consumption support company priorities.
The discussion moves from team-level costs to individual work items and strategic objectives. Nanney describes linking usage to projects, epics and stories rather than stopping at an overall budget. That context also helps distinguish useful discovery work from spending that has drifted off course.
Match Models and Resources to Business Needs
Nanney sees model routing becoming more closely connected to business context. Different tasks may justify different combinations of capability and cost. Choosing the most expensive model for every request is not the same as choosing the best fit.
He also cautions against making staffing decisions based only on rising consumption costs. Organizations need evidence about the work being performed before replacing people with agents or moving resources. Automation can change roles, but assumptions about savings still need testing.
The interview considers the spread of software creation beyond traditional engineering teams. Product managers and technically capable business users can now build more themselves. That flexibility creates opportunities, while raising questions about governance, accountability and enterprise readiness.
Catch Planning Drift Before It Becomes Expensive
Faster execution can pull teams away from their original objectives. Nanney explains why monthly or quarterly planning reviews may struggle to keep pace. By the time leaders spot a mismatch, the money and effort may already be spent.
Managing AI spend therefore requires more than a usage dashboard. Teams need timely insights into priorities, progress and the value of ongoing work. Nanney describes the role of workforce intelligence in bringing those signals together.
The conversation closes on the human challenge of supervising expanding agent activity. Producing information is easier than absorbing it. The priority is helping people decide what to continue, what to stop and where their attention belongs.