Blog: Investment
Your AI subscriptions are not an investment plan
Published by Forever Foundry
One team buys an AI assistant. Another adds a specialist tool. A third already has similar features inside software the business pays for.
Each purchase may make sense. Together, they still need a business case.
AI spending is becoming routine
On 12 October 2026, Capital on Tap published its July to September spending analysis. In its UK SME cardholder dataset, 14.5% of active businesses paid for AI tools, up from 8.0% a year earlier. Around one in five businesses paying for AI paid for both ChatGPT and Claude.
These are payment patterns, not proof of usage or value. The analysis excludes free plans, app-store subscriptions and invoiced contracts from its provider figures. It is not a census of UK businesses. Read the research and methodology.
The FOFO™ view
Several tools are not automatically wasteful. Several purchases without a shared understanding of the work deserve attention.
Imagine a professional services firm where sales uses AI to draft proposals, delivery uses another tool to prepare project plans and finance uses a third to explain overruns.
The subscriptions are visible. The underlying problem may not be: sales promises work that delivery has not scoped, and finance only sees the gap after the project starts. That is the kind of friction we described in finding AI opportunities between two teams.
Cancelling a licence might save a little. Understanding that recurring mismatch could reveal a more valuable investment.
Equally, different teams may genuinely need different capabilities. Standardising everything onto one tool before listening can remove something useful.
Review the work before the renewal
For each paid tool, establish:
- Which business problem is it addressing?
- Who uses it, and what has improved?
- Does another existing service meet the same need?
- What checking, rework or integration does it create?
- What would happen if the subscription stopped?
Compare total cost with the outcome, not just the licence price. For proposal work, that could mean preparation time, scope corrections and project-margin variance. Separate measured improvements from assumptions. Finance can help define those measures early: read why your CFO should help choose the AI problem.
AI Compass™ listens across the organisation to reveal the friction behind individual requests for technology. Slid3rs™ helps teams rank the opportunities by business value and delivery complexity. FlowRoom™ supports agreement on what deserves investment next.
The aim is not the fewest AI tools. It is a defensible reason for every investment, and a clear view of the problems still unsolved.
Find out what deserves investment next. Explore AI Compass™.
Related reading
Next