Why calling it a renewal could be your next software-buying mistake

Procurement teams that still talk about software “renewals” are, in effect, assuming the answer before the question has even been asked. That was the central argument made by Michael Shields in a recent OnlyCFO essay, which reflects a wider shift in finance and procurement as artificial intelligence compresses decision cycles and makes vendor landscapes more fluid.

The old playbook was built for a slower market. Contracts were reviewed months before expiry, stakeholders wer...

e consulted early and the incumbent was often retained after a measured renegotiation. That approach made sense when the main task was to avoid last-minute surprises. It looks much less adequate now, when product capabilities can change quickly, new entrants can emerge almost overnight and switching costs are falling in some categories even as the cost of a poor decision remains high.

Shields argues that the language matters because it shapes behaviour. A “renewal” implies continuity, whereas an expiring contract forces a more deliberate judgement about whether the existing supplier still deserves the business. In his view, that distinction is more important now because buyers are no longer simply managing time to expiry; they are managing changing business needs, changing pricing models and a changing competitive field.

That perspective fits with the broader procurement technology market, where platforms are increasingly being built around continuous monitoring rather than periodic review. Tropic, which positions itself as an intelligent procurement platform for AI and SaaS spending, says it tracks committed versus consumed spend, forecasts potential overruns and benchmarks pricing against peers. The company says it helped customers generate $85 million in savings in 2025, underlining the appeal of tools that try to combine automation with human judgement.

The shift is also visible in the way organisations are thinking about AI-specific forecasting. Runrate, a FinOps specialist, argues that traditional budgeting models are often too blunt for AI because usage can scale in non-linear ways as adoption spreads across teams and use cases. Its framework recommends a bottom-up approach, starting with current spend and then modelling growth by agent or workflow rather than relying on broad top-down assumptions.

That warning is echoed by AI spend management tools that are entering the market to deal with a common problem: too little visibility. AI SpendOps, for example, focuses on token-level attribution and policy controls across multiple model providers, while AIFinOps emphasises forecasting, pricing intelligence and executive reporting. The common thread is that AI costs are becoming more dynamic and more operationally complex, making static assumptions increasingly dangerous.

Shields’ case is that procurement must adapt to that reality by moving from calendar-led administration to signal-led decision-making. Instead of waiting for a contract to approach its end date, he argues that finance and procurement teams should watch for changes in usage, stakeholder satisfaction, market competition and pricing. Mid-contract reviews, pulse surveys and benchmark checks, he says, can be more useful than a once-a-year scramble.

That does not mean every incumbent should be challenged or every new vendor embraced. On the contrary, one of the more practical warnings in his piece is against both complacency and novelty bias. Some established suppliers have improved materially and may still be the best choice with better terms. Equally, new AI-native rivals are not automatically superior, even if they promise dramatic gains. The issue is not whether to favour the old or the new, but whether the business problem has changed enough to justify re-testing the decision.

The build-versus-buy question has also become more complicated. McKinsey has said that a substantial share of companies have already declined at least one software purchase because they believed AI coding tools could let them build the capability internally. That may be true in some cases, but it creates a different risk: teams can underestimate maintenance, governance and support costs, or build tools that look efficient on paper but perform poorly in practice. Retool, meanwhile, has reported that many builders shipped software outside formal IT oversight in the past year, highlighting how quickly shadow IT can reappear under a new label.

The result is a procurement environment in which neither long-term loyalty nor fast switching can be treated as a default. Shields points to a statistic he heard at Gartner’s procurement conference in September 2026 suggesting that a large majority of buyers regret a technology purchase within the first year. Whether that is because requirements were badly defined, benchmarks were weak or internal alignment was poor, the implication is the same: organisations are often getting the original decision wrong, not merely the renewal.

This is why the terminology change is more than semantics. If teams describe everything as a renewal, they risk treating the incumbent as the natural answer. If they describe the same event as an expiring contract, they are forced to ask a harder question: what problem are we trying to solve, what evidence do we have, and what is the best route now?

For many finance leaders, that may mean richer data, more frequent reassessment and a greater willingness to challenge assumptions on both sides of the buying decision. In a market where software pricing, product capability and user expectations can all shift within months, that discipline may matter more than contract length itself.

Shields closes on a pragmatic note. The aim is not perpetual churn, but better judgement. The task is to know, before the CFO asks, which contracts deserve real work and which simply need to be let go.

Source: Noah Wire Services