Outcomes-Based Commercial Models: How AI Rewrites the Software Services Contract

For thirty years, the enterprise software services contract has priced one thing: human effort. Rates per hour, per developer, per month. The unit varied, but the logic did not. You bought labor, and more work meant more labor, which meant more cost.

Agentic AI breaks that link. When an agent library and a handful of engineers can deliver what once required a large team, the connection between effort and output comes apart, and with it the assumption the entire commercial model was built on. If a partner can produce the same result with a fraction of the people, why is the buyer still paying by the person?

That question is quietly rewriting the software services contract. The shift is from paying for effort to paying for outcomes, and it changes the economics for both sides. This piece looks at why the old model is breaking, what outcomes-based commercial models actually involve, where the risk moves, and what a buyer should ask for before signing one.

The time-and-materials contract was honest about one thing: it aligned the buyer’s cost with the provider’s cost. Both scaled with headcount. Neither side had to agree on what the work was worth, only on what it cost to staff.

That alignment is exactly what AI dismantles. Once agents perform a meaningful share of the execution work, the provider’s cost stops tracking the number of people on the account. A provider billing by the hour now has a structural reason to prefer more hours, which is precisely the opposite of what the buyer wants from an AI-enabled engagement. The pricing model and the technology are now pulling in opposite directions.

This is the same fault line that runs through most enterprise AI programs. McKinsey’s State of AI survey, published in late 2025, found that nearly two-thirds of organizations have not yet begun scaling AI across the enterprise, and that fundamental workflow redesign correlates more strongly with financial impact than any other change, yet only about a fifth of firms have done it. The commercial model is one of those workflows. A buyer can adopt AI-native delivery and still bolt it onto a contract that rewards effort, which quietly caps the return before the work begins.

An outcomes-based commercial model prices the result rather than the labor used to reach it. Instead of buying a number of engineers for a number of months, the buyer contracts for a defined outcome at a defined cost, with payment tied to results the finance team can measure.

The word “outcome” has to be specific to work. A vague promise of transformation is not an outcome. A measurable one is: a release-cycle time reduced against a documented baseline, a migration completed to a defined scope, a cost of delivery lowered by an agreed percentage, a defect rate held under a threshold. The discipline of the model comes from naming the metric before the work starts, which forces both sides to agree on what success is while they can still shape it.

This is the practical face of what the industry calls Services-as-Software: delivery bought as a measurable, consumable result rather than as staffed capacity. It reframes the vendor relationship around what the business actually needs, which is the outcome, not the hours that produced it.

The model does not fit everywhere, and a credible partner will say so. Genuinely open-ended research, work whose scope cannot be defined at the outset, or engagements where the buyer controls most of the variables that determine success are poor candidates for a fixed outcome commitment. Outcomes-based pricing works where the result can be named and the provider can meaningfully control the path to it. Much of modern software engineering (modernization, migration, quality engineering, product delivery) fits that description well, which is why it is where the model is taking hold first.

It helps to see this shift in historical context, because the services industry has changed its basis for value before.

Ascendion’s founder and CEO, Karthik Krishnamurthy, framed the pattern in Forbes as a sequence of value levers in the $1.4 trillion IT services industry. The first was economies of scale: delivering standardized work more cheaply at volume. The second was wage arbitrage: sourcing talent wherever it was most cost-effective, which reshaped the industry for two decades and has largely played out as those wage gaps narrowed.

The third lever is AI Arbitrage: using AI agents and AI-augmented teams to deliver outcomes at lower cost and higher speed than a labor-driven model allows, and redirecting the freed effort toward higher-value work. Each lever changed what the buyer was really paying for. Wage arbitrage moved the question from how many people to where the people were. AI Arbitrage moves it again, to how much of the work needs a person at all. Outcomes-based pricing is simply the commercial expression of that third lever. It is what the contract looks like once value is no longer a function of headcount.

The model is not theoretical. Ascendion offers clients the option of what it calls AI Commercials, priced against the AI-driven productivity and savings delivered, in place of the heritage model of billing for time and materials. In practice, that choice has produced double-digit savings that free capital for reinvestment, which is the point of the exercise: the saving is not the end state, it is the fuel for the next initiative.

Underneath any such commitment sits the delivery discipline that makes it safe to offer. AAVA™ supplies the platform and the agent library, agentic workflows supply the governance, and trained engineers supply the judgment that keeps AI-generated work accountable in production. This is the Carbon + Silicon model: agents handle the repeatable execution, people make the calls that carry consequences. The production record behind it, including a delivery system independently appraised at CMMI Level 5 for both Development and Services, is what lets a provider stand behind an outcome rather than behind a timesheet.

None of this removes the buyer’s responsibility to structure the deal well. It changes what a good structure looks like.

A CIO or CFO evaluating an outcomes-based model should press on a few specific points before signing.

Insist on a measurable outcome and a documented baseline. If the provider cannot state the metric and the starting point, the contract is effort pricing wearing new language. The baseline is what makes the result provable later.

Ask where the delivery risk sits, in writing. A genuine outcomes-based model puts meaningful provider fee at stake against the result. If all the risk still rests with the buyer, the model is not what it claims to be.

Confirm the provider can actually control the outcome. The credibility of any commitment rests on the platform, the engineering depth, and the delivery discipline behind it. Ask what runs in production today, and at what scale, because that is the evidence that a committed outcome is achievable rather than aspirational.

Keep governance inside the contract. Outcome pricing does not reduce the need for human oversight, audit trails, and validation, particularly in regulated industries. The outcome has to be reached in a way the business can defend to a regulator, not only delivered on time.