The vendor scorecard just got a shorter shelf life
Forrester examined more than 200 technology and services markets to determine where AI is displacing established software categories fastest, and the resulting picture should concern any CIO still running a standard three-to-five-year vendor renewal cycle without revisiting assumptions annually. Application development and software, broader technology services, and growth and transformation services all rank among the sectors most exposed to near-term disruption from general-purpose AI capability, according to the research. The breadth of the study, spanning well over 200 distinct markets, is itself notable: this is not a narrow warning about one product category but a systematic mapping exercise across nearly the entire enterprise software landscape.
Craig Le Clair, a Forrester vice president, summarized the underlying mechanism bluntly: these software platforms built substantial intellectual property solving specific customer problems, and when AI solves those same problems more easily and cheaply, that capability is inherently disruptive to the vendor's business model. A category that took a vendor a decade to build a defensible moat around can now be approximated by a general-purpose model in a fraction of that development time and cost.
The spending data backs up the anxiety
This is not an abstract, forward-looking prediction sitting in a slide deck somewhere. Forrester's research found three in five IT professionals reported increased AI overspend at their organizations already, while wasted SaaS spending grew 10 percentage points year over year, a sign that procurement teams are struggling to keep license counts and actual feature usage aligned as AI capabilities get bundled quietly into renewals nobody has fully audited line by line. Ten points of year-over-year growth in waste is a meaningful jump for any finance team already under pressure to justify technology spend against tighter overall budgets.
That waste tends to surface in board reviews well before anyone traces it back to a specific bundled AI feature nobody asked for or ever actually used after the first demo, at which point the fix usually costs more political capital than it would have earlier, when a quiet line-item cut was still an option nobody had to defend publicly. Compounding that pressure, Forrester projects AI coding costs will exceed average developer salaries by 2028, a figure that inverts the traditional build-versus-buy calculus CFOs have relied on for planning purposes for years. When AI-assisted development becomes expensive enough to rival direct headcount costs, the long-standing argument for buying commodity software instead of building custom internal tools weakens considerably rather than strengthening as most finance teams still assume.
The software development lifecycle is collapsing
Le Clair's assessment of developer-facing tools specifically is especially pointed: the whole software development lifecycle is being collapsed and reduced in stages, which is directly disrupting the companies that built profitable businesses around servicing individual stages of that lifecycle. Testing tools, code review platforms, and project management software all built around a multi-stage human workflow face genuine pressure when AI compresses several stages that previously required separate specialized tools purchased from separate vendors.
That trend connects directly to a pattern McKinsey flagged earlier this year, that a growing share of enterprises are choosing to build software internally now that AI has made custom development materially cheaper than it was even two years ago. Forrester's research adds the vendor-side view of that same underlying shift: the addressable market for buying narrow point solutions is shrinking measurably as the cost of building rough equivalents in-house continues to fall.
What Forrester actually recommends CIOs do
The report's practical guidance centers on a continuous audit discipline rather than a one-time annual review exercise. CIOs should regularly assess which vendor features are becoming commoditized by AI capabilities already available elsewhere in their existing stack, and avoid signing new multi-year commitments to tools whose core value proposition is at meaningful risk of being absorbed by a general-purpose model sometime within the life of that same contract term. That is a harder discipline to sustain than a standard annual vendor review, since AI capability keeps advancing continuously inside renewal cycles rather than neatly between them on a predictable schedule.
Forrester frames this as a portfolio management exercise: reducing technical debt and controlling cost exposure requires treating the entire vendor stack as a living system under constant reassessment, not a fixed architecture that gets revisited once a year during budget season. Few IT organizations currently have the staffing or tooling in place to run that kind of continuous review well, which is itself becoming a gap that vendor management functions will need to close over the next budget cycle, likely by hiring dedicated headcount rather than layering the task onto an already stretched procurement team that is busy enough handling standard renewals as it is. That staffing gap is quietly becoming its own line item in next year's IT budget request.
The ERP and core systems angle
For CIOs currently negotiating ERP or other core platform renewals, this research offers a useful counterweight to vendor sales pitches that bundle AI features in as automatic, unquestioned value adds justifying a price increase. If a vendor's proprietary AI layer is simply replicating capability already available through a general-purpose model integration elsewhere in the stack, that overlap becomes a legitimate negotiating point rather than a reason to accept higher renewal pricing without pushback. That is precisely the kind of due-diligence question PE-backed operators should expect their own boards to start asking before the next renewal cycle.
That question matters most where AI line items have grown quietly as a share of total software spend without a corresponding review of what those features actually replace or whether anyone downstream has confirmed they work as advertised in the sales pitch that justified the price increase to begin with. Boards that skip this step this year should expect to answer for it during the next one. The broader implication for enterprise software strategy is that vendor lock-in arguments built on proprietary AI features now carry a shorter expiration date than the multi-year contracts they are typically used to justify. That shift should move real negotiating leverage back toward buyers willing to do the commoditization homework Forrester is describing here, rather than accepting vendor claims about differentiated AI value at face value during renewal conversations.
