
‘The question is no longer how much AI can produce, but how much of that output is genuinely usable’: How we use and pay for AI is undergoing a major shift
THE SO WHAT
If enterprises are moving spend toward trust, accuracy, and measurable outcomes, usage-based pricing tied to raw tokens or calls will come under pressure. Vendors and internal teams need to instrument quality and business impact now — or risk being de-scoped when CFOs start cutting "unusable" AI spend.
READ THE SOURCE
MORE FROM THE WIRE
Applied AIMidjourney pushes to expose studios own AI practices in copyright fight
Discovery into Disney, Universal, and Warner Bros.’ own AI use turns this lawsuit into a two-way audit of training data and internal tooling. If you’re using third-party datasets or models at scale, assume your own AI practices could be dragged into the open in future IP disputes.
Applied AIMistral CEO warns closed AI models give providers ‘immense leverage’ over your business
The leverage argument is shifting from pricing to data retention and control—closed providers holding logs and fine-tunes are now a structural dependency, not just a vendor. This is a board-level question: where do you need portability and self-hosting, and where are you comfortable with lock-in for speed.
Applied AINokia's new retro-styled feature phones have AI buttons for some reason — and users are dismissing the odd feature as 'useless' and 'dumb'
An AI button on a feature phone with no clear job to be done is a reminder that ‘AI-washing’ has a user backlash ceiling. If you’re adding AI entry points, tie them to one or two high-frequency tasks—otherwise you’re just burning trust and UI real estate.
Applied AINHS App will use AI to triage patients as part of £10bn tech overhaul
Putting AI triage into the NHS app for 200,000 patients is a live test of routing algorithms at national scale—getting symptom capture and escalation thresholds wrong will show up in A&E load, not just dashboards. Any healthcare or insurance operator deploying similar tools should be instrumenting outcomes and appeal paths from day one, not treating this as a UX upgrade.