Stop Pricing Effort: Value-Based Contracts for the AI-Native Era

On Thursday, in We've Seen This Movie Before, I traced six technology waves, from RISC to agentic AI, in which enterprises bolted new technology onto old architecture instead of rebuilding for it. The Monday before, I showed what the economics look like when operations are built native. This piece takes up the question that follows: how do you price that advantage, and who benefits?

The market has just made that question urgent.

A reallocation, not a recession

Over the past few weeks, the largest IT services stocks have fallen sharply. As of 21 September, TCS, Infosys, Wipro and Accenture are each down roughly 28–38% year to date. At first glance that looks like a technology spending downturn. It isn't.

ISG's Q2 2026 Index shows where the money is going. Combined global contract value grew 43%, the fastest rate ISG has ever recorded. Cloud services grew 65%, and infrastructure-as-a-service 78%, fed by AI demand. Managed services grew just 2.7%. In the Americas, the picture is starker: managed services fell 12%, and IT outsourcing within it fell 23%, led by application development and maintenance as AI takes over work people used to do.

Enterprises aren't spending less on technology. They're spending less on human effort delivered as a service. AI is increasing demand for technology while reducing what customers will pay for the labour wrapped around it. ISG itself points to pricing deflation and more provider-funded, AI-powered transformation built into contracts. Providers are increasingly expected to invest in the AI that shrinks their own revenue.

That's the service provider's problem. The CIO has one too.

The CIO's squeeze

Headline IT budgets have kept growing, but the share CIOs control has not. Run-the-business costs absorb most of the budget, and every transformation initiative competes for what's left.

The reason is the pattern from Thursday's piece. In each of the six waves, transformation was layered on top of existing operations rather than replacing them. Each wave brought new tools, platforms and teams, but the old run costs stayed. Bolt-on transformation adds capability and adds cost. It rarely takes cost out.

So the CIO faces a board that expects AI to deliver efficiency, a CFO who sees technology spend rising, and a services contract that rewards the provider for effort rather than results.

Why today's outcome pricing doesn't fix it

The industry's answer has been outcome-based pricing, and providers report customers moving toward it. But most outcome-based contracts are defined in IT terms: uptime, SLA attainment, tickets resolved, mean time to repair.

That creates a trap. When AI cuts the effort needed to hit those targets, the natural move is to reprice the same scope at a lower cost. The client captures the productivity gain, the provider protects its margin for a while, and total contract value shrinks. Outcome pricing on IT metrics doesn't escape AI deflation. It formalises it.

This also explains the apparent contradiction in the market data: providers can truthfully say outcome pricing is gaining traction while traditional managed services stay flat. The flat line is partly what that shift looks like in contract value.

The deeper issue is one I've written about before: legacy operating models can't price efficiency. Managed-services contracts are anchored to cost reduction targets: year-on-year productivity commitments, resource-unit rates and continuous-improvement clauses. None of these gives the provider a reason to eliminate work entirely, because eliminated work is eliminated revenue.

What native makes possible

When I took over building an AI-native operations platform, we onboarded more customers year after year while keeping people and technology costs flat. Every new customer lowered our unit cost instead of adding to it.

In AI-Native Operations: When Growth No Longer Requires Proportional Headcount, I modelled what that means for six customers. With dedicated teams, 180 FTE deliver the work at $18M, leaving $12M of contribution on $30M of revenue. On a shared AI-native platform, about 60 FTE plus $1.5M of platform cost deliver the same work at $7.5M, and contribution rises to $22.5M. The remaining people aren't just handling exceptions; they build, govern and assure the platform. The figures are illustrative, but the pattern reflects what I've seen in practice.

That difference comes from intent architecture: designing operations with the explicit intent to eliminate work that shouldn't exist, not just automate it. Work that can be prevented is prevented, and people focus where expertise matters. Once operations are designed that way, growth stops requiring proportional spend.

The discount trap

The same model shows why the industry's instinct is wrong. A provider can cut its price 20% and still earn $16.5M, well above the traditional $12M. Most will do exactly that, and most outcome-based contracts are that discount in disguise.

But discounting hands away the headroom. The better move is to put part of it at risk against business KPIs: take a lower base fee, then earn the rest back, and more, as the client's business measurably improves. The provider can afford that risk because its costs don't rise with each unit of value delivered. The CIO gets a price that falls only if the business doesn't improve.

Price the business outcome

That means moving the outcome up a level, from IT metrics to business KPIs.

• Insurance: claim cycle time or policy onboarding time.
• Wealth management: asset onboarding time.
• Manufacturing: OEE and unplanned downtime.

These are numbers the CFO and the business already track. When they improve, the value is visible in business terms, not buried in an IT scorecard.

A value-based contract can be structured in three parts:

• A base fee covering the platform and the build of the intelligence layer, so the provider isn't carrying all the risk before results show.
• Value tiers defined by the size of the KPI improvement, for example claim cycle time reduced 15–25%, 25–40%, or more than 40% from an agreed baseline.
• A cap, so the client's finance team can budget with confidence.

The tiers must be defined by improvement achieved, not by features or service levels included. Otherwise it's a service catalogue with a new name.

Making it contractable

Value-based pricing isn't a new idea. It has struggled because no one could agree on how to measure it: where the baseline sits, how much of the improvement the provider caused, and whose data is the source of truth.

AI-native operations change that. The same enterprise intelligence that finds and closes the gaps in a process can also instrument the KPI. The system that reduces claim cycle time is the system that establishes the baseline, tracks the improvement, and separates its contribution from seasonality or other process changes. Baselining and attribution, the issues that usually stall value-based deals, become part of the platform rather than a negotiation.

Who wins

CIOs get a budget case in business terms the CFO accepts. Instead of defending technology spend, they show that the business grew, or a core process got faster, without IT costs growing with it. Savings from eliminated work fund the next outcome, not just a lower line item.

Service providers get revenue and margin that grow with value delivered rather than shrink with effort. That's the only way out of the deflation curve that doesn't depend on selling more hours.

The business gets what it has been asking for all along: technology that moves the numbers it runs on.

The market's verdict this quarter isn't that IT services is finished. It's that pricing effort is. The providers and CIOs who agree to price value, and build native operations that can deliver it, will be on the right side of this reallocation.

Enterprise AI operating models, for a higher performing tomorrow.