For decades, software consulting ran on a simple transaction: clients bought developer hours. Whether structured as time-and-materials, retainers, or staff augmentation, the underlying unit of value was the same. You paid for human effort applied over time, and you hoped that enough hours would produce the outcome you needed.
This model survived because there was a reasonably predictable relationship between hours invested and results delivered. A competent team working for six months would produce roughly six months’ worth of software. The relationship between input (time) and output (working system) was legible enough that pricing by the hour was an acceptable approximation of pricing by the result.
AI has broken that relationship. When an engineer directing AI agents can deliver in two weeks what previously required three months of team effort, the input-output relationship that justified hourly billing collapses. The client is not buying hours anymore. They are buying the outcome. And they are starting to demand that the pricing reflect that reality.
This is not a minor adjustment to fee structures. It is a fundamental restructuring of how software gets bought, sold, and delivered. The industry is calling it outcome engineering, and it represents the most significant shift in software economics since the move from licensed software to SaaS.
The Hourly Billing Paradox
Hourly billing creates a structural misalignment between client and vendor incentives. The client wants the project done as quickly and cheaply as possible. The vendor, whose revenue is a direct function of hours billed, benefits from the project taking longer. Both parties are aware of this misalignment, and both expend energy managing it: clients through oversight and milestone tracking, vendors through estimates and progress reporting. The overhead of managing the misalignment is itself a cost.
This misalignment was tolerable when the variance in delivery speed was narrow. A fast team might be 30% more efficient than an average team. That variance was manageable. Clients could evaluate efficiency, compare rates, and make reasonable assessments of whether they were getting good value for their hourly spend.
AI has widened the variance to the point where hourly billing becomes absurd. A team using AI-native methods can be 500% to 1000% more efficient than a team working traditionally. When that happens, the hourly model produces outcomes that satisfy nobody. If the AI-native team bills at the same hourly rate and finishes in one-fifth the time, the client pays one-fifth the price but wonders whether the work is really done. If the team slows down to bill more hours, they are deliberately wasting the client’s money. If they raise their hourly rate fivefold to compensate, the optics are terrible even if the total project cost is the same.
Every option within the hourly model creates friction. The model itself is the problem.
What Outcome Engineering Actually Means
The term outcome engineering has been gaining traction since early 2026, catalyzed in part by Cory Ondrejka’s Outcome Engineering manifesto. Ondrejka, CTO of Onebrief and co-creator of Second Life, articulated a framework that has resonated widely across the industry: the realization that what we are engineering is not software. It is outcomes. Code is the mechanism, not the product.
This reframing has practical consequences for how work is structured, priced, and delivered.
In an outcome-engineered engagement, the deliverable is not a codebase or a set of features. It is a defined business result: reduced processing time, increased conversion rate, automated workflow, cost savings target. The engagement is structured around achieving that result, and compensation is tied to delivering it.
This is not a new idea in consulting generally. Management consultancies have experimented with outcome-based pricing for years. McKinsey reports that 25% of its engagements have shifted to outcome-based models. Deloitte operates engagement models that share risk and value creation with clients. But in software consulting specifically, outcome-based pricing has been rare because software outcomes are hard to define precisely and even harder to attribute cleanly to the consulting team’s work.
AI changes this equation in two ways. First, it compresses delivery timelines enough that the gap between project start and measurable outcome shrinks dramatically. When you can build, deploy, and measure in weeks rather than months, the outcome becomes visible fast enough to price against. Second, AI enables more precise specification of expected outcomes. When the specification is the primary artifact and the implementation is largely automated, the specification itself becomes a contract: here is exactly what will be built, here is how it will be validated, here is the business metric it is expected to move.
The shift from time-based to outcome-based pricing is not theoretical. The data shows it is already happening at scale.
73% of consulting clients now favor value-based or outcome-driven pricing over traditional hourly rates. This is not a preference expressed in surveys. It is reflected in actual buying behavior: clients are walking away from firms that insist on hourly billing and toward firms that will commit to results.
The SaaS industry, which provides a useful parallel, shows the same pattern. Credit-based pricing models have increased 126% year-over-year. Hybrid usage-based models surged from 27% to 41% of SaaS companies in a single year. The entire software industry is migrating away from models that price on inputs (seats, hours) toward models that price on outputs (usage, outcomes, value delivered).
In consulting specifically, the firms that have made the transition are outperforming. Hybrid consulting teams using AI deliver projects in 40% to 60% of traditional timelines while maintaining margins. The math works because the cost of delivery has dropped faster than the price, creating margin expansion even at lower total project costs.
But not everyone is convinced. There is a thoughtful counterargument that pure fixed-bid, fixed-scope engagements are inherently difficult because scope uncertainty is irreducible. Any project complex enough to need external help is complex enough that requirements will change during delivery. The counterargument is correct about scope uncertainty. It is wrong about the conclusion.
The answer is not to abandon outcome-based pricing in favor of hourly billing. It is to structure outcome-based engagements in ways that accommodate scope evolution. Phased delivery with outcome gates at each phase. Value-based pricing for defined outcomes with change-order mechanisms for scope additions. Shared risk models where the vendor invests in achieving the outcome and shares in the upside. These structures are more complex than simple hourly billing, but they align incentives correctly, and they scale in a world where AI makes delivery speed highly variable.
The Backlog Is Dead
Ondrejka’s manifesto makes a provocative claim that we think is essentially correct: the backlog is a relic of human limitation.
In traditional software development, the backlog was where ideas went to be prioritized against limited human capacity. You had more things you wanted to build than people to build them, so you maintained a prioritized list and worked through it sequentially. The backlog was a rationing mechanism for scarce human attention and effort.
When agentic systems remove time and human bandwidth as the primary constraints on implementation, the rationing mechanism becomes unnecessary. Ideas are not rejected for lack of developer time. They are evaluated on a simpler criterion: does the expected outcome justify the cost in compute? If the cost of building something with agents is less than the value of the outcome, build it. If not, do not. The backlog collapses into a continuous cost-benefit calculation.
This has profound implications for how consulting engagements are structured. Instead of scoping a twelve-month roadmap and pricing a team to execute it, the engagement becomes a continuous partnership where the client identifies desired outcomes, the consulting team assesses the cost to achieve each outcome, and work proceeds on everything where the math works. There is no backlog because there is no queue. There is a portfolio of outcomes being pursued in parallel, constrained only by budget and the computational cost of delivery.
This model, continuous outcome delivery, is what we believe replaces the traditional project-based consulting engagement. It is closer to a subscription than a project, but what the client subscribes to is not hours or headcount. It is access to a team with the capability and the AI infrastructure to deliver outcomes on demand.
The Death of the Estimate
Hourly billing required estimates. Clients needed to know how many hours a project would take so they could approve a budget. Vendors needed to estimate hours so they could staff the project. Estimation was a core competency of consulting firms, and entire methodologies (story points, planning poker, reference class forecasting) existed to make estimates less wrong.
Estimation in the AI era is simultaneously easier and less meaningful.
It is easier because AI can prototype solutions rapidly, giving concrete data about implementation complexity before the engagement begins. Instead of estimating that a feature will take 200 hours based on analogies to previous work, you can have an agent build a prototype in an afternoon and know with much higher confidence what the actual implementation will require.
It is less meaningful because the relationship between effort and cost has changed. In hourly billing, the estimate was a cost proxy: 200 hours at $250/hour equals $50,000. In outcome-based pricing, the cost is a function of the outcome’s value, not the effort to achieve it. If the outcome saves the client $500,000 per year, the price is a fraction of that value regardless of whether it takes 20 hours or 200 hours to deliver.
This makes the traditional estimation ritual unnecessary. What replaces it is outcome definition and value assessment: precisely specifying the desired outcome, validating that it is achievable, and pricing it relative to its business value. This is a different skill than estimation. It requires business acumen, domain expertise, and the ability to define measurable outcomes, not the ability to guess how many hours a task will take.
How Agencies Must Adapt
The transition from hourly billing to outcome-based pricing requires more than changing the numbers on proposals. It requires restructuring how consulting firms operate.
Capability over capacity. The hourly model sold capacity: bodies with skills available for a period of time. The outcome model sells capability: the ability to achieve a specific result. This means firms need to invest in proprietary tools, frameworks, and methodologies that amplify their delivery capability beyond what raw talent can provide. The firms that treat AI as a way to make their existing developers bill more hours are missing the point. The firms that use AI to build delivery infrastructure that allows them to commit to outcomes are building durable competitive advantages.
Vertical specialization. When you sell hours, being a generalist works. Every hour of React development is roughly equivalent. When you sell outcomes, domain expertise becomes critical. A healthcare outcome requires understanding healthcare workflows, regulatory requirements, and integration constraints. A financial services outcome requires understanding compliance frameworks, data governance, and risk management. Firms that specialize in verticals can command 30% to 40% fee premiums over generalists because they can define outcomes more precisely and deliver them more reliably.
Risk management as a core competency. Outcome-based pricing means the firm takes on delivery risk. If the outcome is not achieved, the firm does not get paid (or gets paid less, depending on the model). This requires sophisticated risk management: the ability to assess outcome achievability before committing, the discipline to walk away from engagements where the outcome is poorly defined or the risk is too high, and the financial reserves to absorb delivery failures when they occur.
Continuous engagement models. Project-based engagements with defined start and end dates are giving way to continuous partnerships. The client does not hire you for a project. They hire you as their outcome delivery partner, with an ongoing relationship where outcomes are continuously identified, evaluated, and delivered. This model produces more stable revenue for the firm and better results for the client, because the firm maintains deep context about the client’s business rather than rebuilding understanding with each new project.
Talent model redesign. Hourly billing rewarded utilization: how many hours of your people are billable. Outcome-based pricing rewards leverage: how much value your people can deliver per unit of effort. This means equipping top performers with AI tools that amplify their impact by multiples, investing in the specification and context engineering skills that drive agent performance, and compensating practitioners on revenue and margin metrics rather than billable hours.
The Competitive Landscape
The shift to outcome engineering is reshaping the consulting industry’s competitive dynamics.
Over 1,200 new software consulting firms launch monthly in the United States alone. Most of these are small teams or solo practitioners who can deliver outcomes cheaply using AI tools. The barrier to entry for software consulting has collapsed. Anyone with an AI coding subscription and domain expertise can compete for outcome-based engagements.
This creates intense price pressure on commodity outcomes. Building a standard internal dashboard, automating a routine workflow, implementing a common integration, these are outcomes that small, nimble competitors can deliver at a fraction of what large consulting firms charge. The large firms cannot compete on price for commodity work, and they should not try.
Where large firms retain structural advantage is in complex, multi-system outcomes that require deep domain expertise, cross-functional coordination, and enterprise governance. Building a compliant healthcare data pipeline that integrates with legacy EMR systems, implements role-based access control, passes a SOC 2 audit, and delivers measurable clinical outcomes, that is not a weekend project for a solo practitioner with an AI coding tool. That is a complex engagement that requires the kind of depth, breadth, and institutional credibility that established firms provide.
The consulting market is bifurcating. Commodity outcomes will be delivered by small, fast, cheap competitors. Complex outcomes will be delivered by specialized firms with deep domain expertise and enterprise delivery capability. The middle, generalist firms selling hours at mid-market rates, is being squeezed from both sides.
The $1.5 Trillion Convergence
Industry analysts are projecting a $1.5 trillion “Services-as-Software” market that represents the convergence of services and products. This is not services companies becoming software companies or software companies becoming services companies. It is both converging on the same model: continuous delivery of business outcomes, enabled by AI, priced on value.
Software companies are adding services layers: implementation, customization, ongoing optimization. Services companies are building software assets: proprietary frameworks, domain-specific tools, automated delivery pipelines. Both are converging on a model where the client buys an outcome and the provider uses whatever combination of software and services is most efficient to deliver it.
This convergence is the natural result of AI making the boundaries between “software product” and “consulting service” meaningless. When a consulting team uses AI agents to build a custom solution, is that a service or a product? When a software company uses AI to customize its product for each client, is that a product or a service? The distinction no longer maps to anything real.
What matters is the outcome. Did the client get the result they paid for? Was it delivered at a price that reflects its value? Can it be maintained and evolved as the client’s needs change? These are the questions that define the new market, and they do not care whether the answer involves packaged software, custom code, human services, or AI agents.
Our Position
We made the transition to outcome-based pricing two years ago. It was not easy. It required rethinking our engagement models, our compensation structures, our risk management processes, and our delivery methodology.
The results have validated the transition. Client satisfaction is higher because incentives are aligned: we succeed when they succeed. Revenue per engagement is higher because we capture a share of the value we create rather than billing for the hours we spend. Delivery speed is faster because we are motivated to deliver outcomes efficiently rather than to maximize billable hours.
Most importantly, the outcome model forces a level of discipline in engagement design that hourly billing never required. When your revenue depends on achieving a defined result, you invest heavily in defining that result precisely. You push back on vague requirements. You insist on measurable acceptance criteria. You scope ruthlessly. This discipline produces better outcomes for clients even when the pricing model is not the primary concern.
The software consulting industry is in the early stages of a structural transformation. The firms that cling to hourly billing will find their clients migrating to competitors who offer outcome-based models. The firms that make the transition early, and build the operational infrastructure to deliver on outcome commitments reliably, will capture disproportionate share of a growing market.
The era of billing by the hour is ending. The era of engineering outcomes has begun. The question for every consulting firm, and every enterprise that buys consulting services, is whether they will lead the transition or be disrupted by it.
