IT Services Tie AI Productivity Gains to Contract Pricing
Summary
Cognizant says fixed-price and transaction work reached nearly 58% of revenue, showing how contract structure determines who keeps AI productivity gains.
From October 23 to 29, 2026, Booz Allen Hamilton (BAH), Cognizant Technology Solutions (CTSH), and Sopra Steria (SPPSY) said on earnings calls that AI is changing how services contracts are priced. Cognizant's fixed-price and transaction-based work has reached nearly 58% of revenue.[1][2][3]
Contract structure decides who keeps lower delivery costs
IT services and government consulting have often been billed by labor hours and reimbursable costs. When AI reduces the labor needed for software development, operations, or analysis, a vendor billing by time may lose billable volume. Customers can also demand lower prices at renewal, so technical adoption alone does not guarantee higher margins.
Fixed-price, transaction-based, and outcome-based contracts allocate the savings differently. Customers buy a defined result, software unit, or transaction volume. If the vendor uses AI to lower delivery costs, it can still collect the agreed price and retain part of the difference. The vendor also assumes the risk of cost overruns and failed delivery.
Three companies already see the contract mix changing
Booz Allen said the fixed-price mix in its national security business increased three percentage points year to date. Management expects the shift to create more customer value and improve profitability over time.[1] It described outcome-based contracting as a bet on its delivery capabilities because lowering the cost curve can produce retained savings that would flow back to customers under cost-plus work.
Cognizant quantified the scale of the transition. Fixed-price and transaction-based work represented nearly 58% of third-quarter revenue, up 140 basis points year over year and more than 500 basis points from the third quarter of 2022.[2] Management tied the change to the convergence of human and digital labor and discussed outcome pricing, AI-infused rate cards, and platform monetization.
Sopra Steria shows that vendors do not retain every gain. The company said productivity improvements from AI and operating efficiency are already part of renewal negotiations, with some savings shared with customers. At the same time, about 60% of its business can take results-based commitments, providing a contractual route to retain efficiency gains.[3][4]
“Our mix of fixed price and transaction-based work is steadily increasing. In the third quarter, this mix was nearly 58% of our revenue, up 140 basis points year over year, and more than 500 basis points since the third quarter of 2022.”[2]
The control point moves from headcount to delivered results
As the contract mix shifts, services revenue can become less directly tied to employee count. Competition may depend more on which vendors can deliver the promised result consistently. Relevant follow-up indicators include the fixed-price and transaction share of revenue, unit delivery cost, renewal discounts, and project overruns, rather than the number of AI tools deployed.
The route has clear limits. SAIC said procurement offices delayed some large awards while implementing fixed-price directives. Infosys said competitive customers sometimes demand productivity gains beyond what AI can actually deliver.[5][6] Fixed-price work can therefore amplify the gains of strong execution while leaving estimation errors, delays, and pricing pressure with the vendor.
Companies exposed to the change
- CACI International (CACI): About 35% of its business is fixed-price, and it sells agile software development in fixed-price units. It is exposed to the same gain-retention mechanism, but the current materials do not quantify realized AI margin gains.[7]
- Amentum (AMTM): Its federal portfolio faces the same fixed-price conversion. Management expects implementation by mid-2027 to be positive for profitability, but that remains a forward-looking assessment.[8]
- Infosys (INFY): It shows the other side of the rule. Customers are seeking AI productivity through renewal pricing, indicating that labor-billed providers may struggle to retain all of the savings.[6]
Sources
[1] Drillr · Booz Allen Hamilton · October 23, 2026 · FY2027 Q2 earnings call
[2] Drillr · Cognizant Technology Solutions · October 29, 2026 · Q3 2026 earnings call
[3] Drillr · Sopra Steria · October 29, 2026 · Q3 2026 earnings call
[4] Drillr · Sopra Steria · July 29, 2026 · H1 2026 earnings call
[5] Drillr · SAIC · August 31, 2026 · FY2027 Q2 earnings call
[6] Drillr · Infosys · October 23, 2026 · FY2027 Q2 results press conference
[7] Drillr · CACI International · August 6, 2026 · FY2026 Q4 earnings call
[8] Drillr · Amentum · August 11, 2026 · FY2026 Q3 earnings call
This article identifies potentially overlooked industry changes and companies. It is not a stock recommendation.