Mel, thanks for posting. Good article. As you highlighted 26% have full visibility. On the other end of the spectrum, 72% have limited/some or no visibility. Organizations must have visibility and control of their costs to be successful.
The solution is to move away from token-based AI pricing, since it makes costs hard to predict or track. Instead, companies should use flat, outcome-based pricing, paying per completed case or result rather than per unit of usage. This gives a clear, predictable cost upfront and ties spending directly to real business value instead of raw consumption. It also makes it far easier to prove return on investment, since success is measured by outcomes delivered, not by how much AI was used. This turns AI spending from an unpredictable guessing game into a controlled, measurable expense.
As AI adoption grows, tracking business value becomes more important than tracking token consumption. Outcome-based pricing provides organizations with greater cost predictability, simpler budgeting, and allows teams to focus on delivering measurable business results rather than monitoring AI usage metrics.
CFOs have a new problem: They can’t track AI usage.
Translation: AI spend is rising… but control isn’t.
The real issue? We’re measuring tokens, not business outcomes.
That’s exactly where Pega flips the model:
No token-based pricing
Flat, outcome-based pricing (per case)
My take: