Skip to content
AI Journey Advisory

Spend and vendors

Before you renew an AI tool, answer four questions

Renewal is the moment of greatest leverage over AI spend. Four questions separate the tools worth scaling from those worth fixing or stopping.

3 min read

1. What are we paying for, and who uses it?

Start with a ledger that reconciles invoices, seats and usage to named owners. Unused seats and overlapping subscriptions are common, and they are the easiest savings to defend.

Where usage data is incomplete, record the gap rather than estimating around it. A known gap is a negotiating point; a hidden one is a risk.

2. What benefit can we evidence?

Benefits claimed for AI tools often mix three different things: cash savings, cost avoidance and capacity freed for other work. Each is real, but only the first reduces spend, so they should be shown separately.

Set a baseline for cost, adoption, quality and benefit now, even an imperfect one, so the next renewal can be judged on movement rather than opinion.

Exhibit 1Stop, fix or scale: matching the renewal decision to the evidence
DecisionWhen the evidence showsNext step
StopLow use and no evidenced benefitExit at renewal and reclaim the budget
FixBenefit is evident where used, but adoption, fit or contract terms hold it backClose the gap with an owner and a date before renewing
ScaleStrong use and evidenced benefit, with costs understoodExpand deliberately, with a baseline to measure against

Source: AI Journey Advisory, AI Spend and Value Review service scope.

3. Should we stop, fix or scale?

With costs and evidence side by side, each tool falls into one of three decisions. Stop where value is absent, fix where adoption or fit is the problem, and scale where the evidence is strong. Record the reason, and the evidence still needed before expanding.

4. What would the alternative cost over three years?

If a replacement is on the table, compare up to three vendors on features, cost, risk and fit, and model licensing, implementation and ongoing costs over three years. First-year pricing rarely tells the full story.

End with acceptance criteria the preferred vendor must meet, so the decision can be validated before contracts are signed.

How we can help

Services that put this into practice

More perspectives

Strategy and value

From fifteen ideas to three business cases

Most organisations do not lack AI ideas. They lack a fair way to choose between them. A simple, evidence-led funnel turns a long list into decisions leadership can fund.

3 min read

Governance and risk

Governance that lets AI move faster

Clear rules, named owners and a simple approval path make it easier to say yes to AI, because everyone knows what yes requires.

3 min read