Enterprise clients are increasingly revisiting IT-services contracts before their scheduled renewal dates as AI reduces the time and number of people needed to deliver the same work. In some cases, clients are seeking price cuts of 25-30%, while others are asking for more work at the same cost or for a share of the productivity gains.
The shift is emerging as a gap opens between how quickly AI changes the delivery of IT services and how long enterprise contracts remain in place.
“AI moves in months; contracts move in years, and clients no longer accept absorbing that gap,” said Sridhar Mantha, CEO, Generative AI Business Services.
The trend is most visible in software development, testing and application maintenance, where AI-assisted coding and other tools can reduce the effort required to complete work. Clients that can demonstrate these productivity gains are increasingly questioning why they should wait for a contract renewal before seeking changes.
Mantha said clients with mature AI-enabled delivery teams are increasingly asking how productivity gains reach them. While a lower price is one option, clients are also seeking gain-sharing arrangements, additional scope at the same cost or service-level agreements based on faster delivery.
Gartner said AI-related mid-contract renegotiations were becoming more common, although they were not yet widespread across the enterprise IT-services market.
“AI-driven mid-contract renegotiations are becoming more common, but they are not yet widespread across the enterprise IT services market,” said Biswajit Maity, senior principal analyst at Gartner.
According to Maity, most of the activity is concentrated among large enterprises with strong sourcing maturity and the ability to measure AI-related productivity gains. Most clients are still assessing the benefits they have realised rather than reopening contracts as a matter of routine.
The issue is particularly relevant to contracts based on the number of people or hours used to deliver a service. If AI allows three engineers to deliver work that previously required five, clients can question whether the original pricing still reflects the work being done.
“Time and material contracts face the most direct pressure through reduction of team size to achieve same goals, and so does any pricing model built primarily around headcount,” Mantha said.
The same issue can arise in service desks, where AI agents can handle password resets, ticket routing, knowledge searches and basic troubleshooting with limited or no human intervention.
Maity said time-and-material, FTE-based and ticket-based models face the greatest pressure because they are linked to human effort, hours worked or transaction volumes.
Clients are also assessing whether their own AI capabilities have reduced the amount of work they need from vendors.
“As enterprises develop their own AI tools and platforms, or establish centers of excellence (COEs), the need for vendor-delivered work may implicitly diminish, including dependency on labor arbitrage,” Maity said.
For Indian IT-services companies, repeated mid-cycle resets could put pressure on contracts that depend on billing for people and hours. As AI reduces the effort required for existing work, providers may have to shift towards pricing based on outcomes rather than the resources deployed.
