The Death of the Consulting Hour — When Value Becomes the Unit of Cost

For roughly a century, the price of professional expertise has been calculated the same way: time multiplied by rate. The hour was the unit. Whether you were buying legal counsel, a strategy engagement, a marketing retainer, or a development sprint, the meter ran on human time, and the invoice was a record of how much of it you consumed. You did not buy a result. You bought hours, and you hoped the result came with them.

That arrangement is ending — not slowly, and not at the margins. Among US accounting firms, hourly billing fell to just under 4% in 2025, down from nearly 8% a year earlier, while fixed-fee pricing climbed to 54%, according to Ignition's 2025 pricing benchmark. The most durable pricing convention in the history of services is being retired in real time. The cause is not a fad in how firms package their fees; it is that AI has broken the link between time and value. When a system compresses a five-hour task into one, the hour stops measuring anything worth paying for.

This is the most important thing the agentic shift does to you, the buyer. We have written separately about what it does to the firms selling these services — The Operating Model Shift — From Agency to Agentic Engineering — and about the market it is creating in The $200B Agentic Services Opportunity. This piece is the other side of the table. When the hour dies, the determinant of what you pay shifts from hours consumed to value delivered — and that single change rewrites how you scope, contract, allocate risk, and pay. For an ambitious CEO, this is not a threat to manage. It is the best procurement leverage you have been handed in a generation. Here is how to use it.

We are also writing this from inside the change, not above it. Facet is moving its own engagements off the billable hour — pricing the work we do for clients on the value it creates, not the time it consumes — because the way we now build makes that not just possible but more honest. So this is part argument and part disclosure: the model we are recommending you buy is the model we are choosing to sell.

Why the Hour Was Ever the Unit — and Why It's Breaking

It is worth being clear-eyed about why time-and-rate lasted as long as it did, because the reason it worked is the same reason it is now failing.

The hour was never a good measure of value. It was a good proxy for value — the only practical one available — for as long as the work genuinely required human time. If a contract review or a working feature took a skilled person forty hours, then forty hours was a reasonable stand-in for the worth of the thing produced. Buyer and seller could agree on a number neither could see directly by agreeing on a number both could count. The entire apparatus of professional services — timesheets, utilization targets, rate cards, retainers — was built to measure and bill that time.

The proxy breaks the moment time and output come unglued, and that is precisely what agentic delivery does. When the routine production of a service runs through autonomous agents, the marginal cost of one more unit of output collapses toward the cost of compute — effectively near zero — and the time it takes bears almost no relationship to the value it creates. A task that "should" cost forty hours now lands in an afternoon. If you are still paying by the hour, you are paying for a scarcity that no longer exists; if your vendor is still billing by the hour, they are quietly hoping you don't notice.

This is why the shift feels less like a pricing trend and more like a floor giving way. The hour was just how we agreed to count value. Remove the link between time and value, and the hour becomes what it secretly always was — an arbitrary number — and everyone can see it at once.

The New Determinant: From Hours Consumed to Value Delivered

So if not hours, then what? The honest answer is that the unit of cost is moving to the thing the hour was always standing in for: the outcome, and the value that outcome creates for your business.

This sounds abstract until you watch it happen in a real market. Take customer support. Zendesk, one of the largest players in the category, now prices its autonomous AI agents per resolution — on the order of $1.50 for each customer issue the agent fully resolves on its own, as reported by Futurum and IDC. You are not buying seats or the agent's time. You are buying closed tickets, and you pay only when one closes. The unit of cost is the outcome itself.

That is the template generalizing across services. The determinant of cost is no longer "how much effort did this consume" but "what result was delivered, and what is it worth to me." Because the marginal cost of agentic production is so low, the price can finally be tethered to the value of the result rather than the cost of producing it — which is exactly how value-based pricing has always been defined: not "how long will this take?" but "how much is this worth to the client?"

For a buyer, this is the whole game in one sentence: you are no longer purchasing a vendor's time; you are purchasing a result, and the price should reflect the result's worth to you. Everything else in this article is a consequence of that sentence. Once value, not time, is the unit, every habit you built around buying hours has to be re-examined — because most of them were designed to control a meter that no longer runs.

The Incentive Trap: Why Hourly and Fixed-Fee Both Pull Against You

Before we get to how you should buy, it is worth naming the thing the hour was hiding: every legacy pricing model puts the vendor's incentives in tension with yours. Value-based pricing is the first one that does not — and that, more than the headline savings, is why it matters.

Walk the three models by the incentive they create:

• Hourly / time-and-materials rewards slowness. The vendor is paid for effort, so every inefficiency is revenue. A faster method, a reusable asset, a clever shortcut — each one costs the vendor money under an hourly model. You are paying a firm to be efficient while paying it more the less efficient it is. Nobody is corrupt here; the meter simply points the wrong way.
• Fixed-fee rewards corner-cutting. Flip to a fixed price and the incentive inverts but does not align. Now every hour the vendor doesn't spend is margin. The pressure is to do the minimum that clears acceptance, declare victory, and move on. Fixed-fee is better than hourly for budgeting, which is why it is winning the migration in the Ignition data — but it still has the buyer and seller pulling against each other across the table.

Value-based pricing rewards the outcome — for both sides. When the vendor is paid for the result and its worth to you, the vendor's interest and yours finally converge on the same object: a better outcome. The vendor wins by making the result more valuable, not by spending more hours or fewer. For the first time the contract is not a tug-of-war over effort; it is a shared bet on value.

This alignment is the real prize. Value-based costing means both the firm building the solution and the business receiving it are anchored to the same number — the value created — instead of negotiating against each other over a rate card or a fixed bid. When the unit of cost is value, "do more work" and "do less work" both stop being strategies. The only strategy left is "create more value," which happens to be the only thing you ever wanted to buy.

The catch, historically, is that value-based pricing was easy to admire and nearly impossible to run. To price on value you have to be able to deliver iteratively toward a defined outcome cheaply enough that converging on "good" doesn't bankrupt the margin. Under a human-hours cost structure, every round of revision was expensive, so vendors protected themselves with scope walls, change orders, and — underneath it all — the hour. The model that aligned incentives was the model nobody could afford. That is exactly what the agentic harness changes.

Why Value-Based Pricing Is Newly Possible: The Agentic Harness

Here is the mechanism. Value-based pricing is viable now — not in theory, in practice — because agentic delivery makes two things cheap that used to be ruinously expensive: production and, more importantly, iteration.

Production cost we have already covered: when agents do the routine building, the marginal cost of output trends toward compute. But the quieter revolution is iteration. The reason fixed-fee delivery is so adversarial is that feedback is expensive — every "actually, can we change this" costs the vendor hours it wants to protect. Value-based delivery requires the opposite: it needs many cheap rounds of "is this the value we wanted? not quite — adjust" until the outcome is genuinely worth what it costs.

That is precisely what an agentic harness provides. A harness is the loop-and-guardrail infrastructure around the agents: the defined roles, the tools each agent may touch, the order of operations, the checkpoints where a human steers, and the quality gates an output must clear before it advances. Inside that harness, a round of feedback is no longer a budget event — it is one more turn of the loop. The system drafts, a human (or a reviewing agent) judges it against the outcome you defined, the feedback goes back in, and it drafts again, converging on the value target instead of betting everything on a single fixed deliverable. We have written about the same engine on the buy-side of our own work — the bottleneck was never the draft, it was the iteration around it.

This is why we say value-based pricing is newly possible because of the harness, not merely alongside it. Stack the three properties:

1. Anchoring. Agentic delivery lets the work be scoped and priced from the value/outcome backward — you define "done" and "good," and the harness is built to drive toward exactly that, rather than toward a staffing plan.
2. Cheap convergence. Because iteration is cheap inside the loop, the vendor can afford to keep refining until the outcome actually clears the value bar — the thing value-based pricing always required and human-hours delivery could never sustain.
3. De-risked delivery. Multiple small, verified rounds replace one big fixed deliverable, so the risk of "we built the wrong thing for six weeks" collapses. Each loop is a checkpoint; bad direction is caught in a turn, not a quarter.

Put those together and an elegant method appears: deliver value surrounded by an agentic harness, iterate toward the defined outcome through cheap feedback rounds, and price the engagement on the value that results. The harness is not a delivery detail you can ignore as a buyer. It is the thing that makes the pricing model you want safe to sign.

A worked example: scoping, pricing, and iterating one engagement

Make it concrete. Suppose you want to cut the time-to-first-response on inbound sales leads, and you believe faster, better-qualified follow-up is worth real pipeline. Here is how the same engagement runs under the old model versus the agentic, value-based one.

The old way. You write an RFP for a "lead-ops automation project." Three firms propose teams — so many consultants at so many rates for so many weeks. You pick one, approve a fixed bid (which is really an hours estimate with a bow on it), and six weeks later you get a deliverable. If it is wrong, you are in change-order territory. The vendor's incentive the whole time was to spend as few hours as the acceptance criteria allowed.

The agentic, value-based way. You define the outcome and its worth: qualified leads contacted within five minutes, at a defined quality bar, and what that improvement is worth to you per month. Pricing is tied to that — say a hybrid of a modest base plus a per-qualified-lead fee, or a gainshare on the pipeline lift. Then delivery runs inside a harness: an agent loop drafts the qualification-and-routing flow, a human reviews the first cohort of real leads against the quality bar, the feedback goes back into the loop, and it tightens — escalation rules, edge cases, tone — over a handful of cheap iterations until it clears the bar in production. You pay as the outcome lands and holds. The vendor made money by making the result better and getting it to "good" faster, because under this structure that is the only way it makes money at all.

Notice what changed and what didn't. The work is still hard; judgment still matters; humans are still in the loop at every checkpoint. What changed is the unit of cost and, with it, whose interests the contract serves. The harness is what let the vendor afford to iterate to "actually good" instead of stopping at "technically delivered."

How Buyers Must Change the Vendor Relationship

The death of the hour is not something that happens to your vendors while you watch. It changes what you are responsible for in the transaction. Four things move.

1. Scope by outcome, not by staffing

The old RFP asked vendors to propose a team: how many people, at what seniority, for how many weeks. You were scoping the inputs and trusting that the right output fell out the end. In a value-based world, that is backwards. You scope the outcome — the measurable result you want — and leave staffing entirely to the vendor. How many humans and how many agents they use is their problem, not a line item you approve.

This is harder than it sounds. Scoping by outcome forces you to define, up front and precisely, what "done" and "good" mean — the result, the quality bar it must clear, and how you will measure both. That definition work used to be deferrable; you could buy hours and figure out the destination along the way. You cannot buy an outcome you cannot describe. The upside is that this is better procurement regardless of AI: a vendor handed a crisp outcome and a clear quality bar has nowhere to hide.

2. Price on the result, with a model that fits the work

Once you are buying outcomes, the pricing conversation opens up well beyond "fixed fee versus hourly." A few value-based structures are now in play, and the right one depends on how measurable the outcome is and how much risk each side will carry:

Model	How you pay	Fits when
Outcome / per-unit fee	A set price per delivered result (resolved ticket, processed invoice, qualified lead, shipped feature)	The outcome is discrete, countable, and attributable to the vendor
Gainshare	A share of the measurable value created — a percentage of cost saved or revenue gained	The value is real and trackable but not cleanly per-unit; both sides accept shared upside
Subscription-to-outcome	A recurring fee tied to a standing service level of results, not to access or seats	You want a continuous capability (an ongoing volume of outcomes) rather than a one-time deliverable
Hybrid	A modest base fee plus an outcome or gainshare component on top	Measurement is imperfect, or you want to de-risk a first engagement while still tying real money to results

Notice what every row has in common: none of them references time. The vendor's hours are now an internal cost they manage, not a number you reimburse. Notice, too, that the hybrid row is where most real first engagements land — pure outcome pricing requires measurement maturity that few buyers or vendors have on day one, and a base-plus-outcome structure is the honest, common-sense bridge.

3. Reallocate the risk — deliberately

Here is the part that should make an opportunity-minded CEO sit up. Under time-and-materials, you carried almost all the delivery risk. If the work took twice as long as expected, you paid twice as much. If the approach was wrong, you funded the dead ends. The vendor was paid for effort whether or not the effort produced anything — an extraordinary risk position to have accepted for a hundred years, accepted only because there was no alternative.

Value-based pricing moves that risk to the party that controls it. When a vendor agrees to be paid per result — or to put a share of its fee on the line against a gainshare target — it is absorbing the risk that its own approach works. That is not a concession you extract through hard negotiation; it is the natural shape of the deal once the unit is value, and it is leverage that did not exist when the only thing on offer was hours. The vendors confident in their agentic delivery take that risk willingly. The ones who flinch are telling you something.

4. Re-tool the paperwork: RFPs, SOWs, SLAs, and payment triggers

The contractual machinery of professional services was written, top to bottom, to buy time and effort. It does not survive contact with outcome buying, and four documents need to change.

• The RFP stops asking for proposed teams and rate cards and starts asking for proposed outcomes, the price per outcome, and the governance around how those outcomes are produced and verified.

The SOW becomes performance-based: it defines the result, the quality standard it must meet, and what happens on a miss — and is deliberately silent on how the vendor staffs the work. (Statement-of-work procurement is already the fastest-growing slice of services spend precisely because it is outcome-shaped rather than hours-shaped.)

The SLA stops measuring availability and response time — the right metrics when you bought infrastructure — and starts measuring the things that matter when you buy decisions: outcome accuracy, the rate at which the system escalates to a human, and the error threshold beyond which remediation is triggered.

• The payment trigger moves from "work performed" or "milestone reached" to "outcome delivered and verified." You pay when the result lands and clears the bar, not when a timesheet is submitted.

None of this requires you to throw out your procurement function. It requires you to point it at a different target. (For a firm-side view of why vendors are restructuring to be able to sell this way, see the operating-model piece linked above.)

How to Tell a Genuinely Value-Based Vendor From an Hourly Shop in Disguise

The market is moving faster than the integrity of the marketing. Many firms have discovered that "outcome-based" and "value-based" are excellent words to put on a slide, and have changed almost nothing underneath. The tell is always the same: a firm whose cost structure is still fundamentally a roster of billable people cannot genuinely price on value, because its own costs scale with time. It can only dress up time-and-materials in outcome language. Four questions cut through it.

1. "Will you put real money on the outcome?" The single most revealing question; everything else is commentary on it. A vendor that genuinely produces results will tie a meaningful portion of its fee to those results. A vendor that insists on billing only for access, only for hours, or only for a fixed fee that quietly maps back to a headcount estimate does not trust its own delivery enough to stand behind it. Watch what they do with the downside, not just the upside.

"Show me the price per outcome — and the math behind it." A real value-based vendor can name a unit and a number: this much per resolved ticket, this share of documented savings. A disguised hourly shop deflects to "it depends on scope" and, when pressed, produces a number that is transparently hours × rate with a new label. Ask them to derive the price from the value to you, not their cost to deliver. If they cannot reason about your value, they are pricing their time.

1. "What is your marginal cost of one more unit?" You will rarely get a precise answer and do not need one — you are listening for whether the vendor understands the question. A genuinely agentic model has a marginal cost that trends toward compute; a staffing model's marginal cost is another person's salary. A vendor for whom every additional unit means another body is the old model with new vocabulary.
2. "What happens to my bill if you get faster?" Under the old model, vendor efficiency was the vendor's gain and often your loss — the faster they worked, the more they billed elsewhere while your rate held. Ask directly whether improvements in delivery speed flow to you as lower per-outcome cost or are simply pocketed. A value-based partner's incentives are aligned with yours by construction; an hourly shop in disguise has the opposite incentive and will struggle to answer honestly.

A vendor who answers all four with specifics and a straight face is rare, and worth a serious relationship. A vendor who fails two or more is selling you the label.

A 90-Day Buyer Action Plan

You do not need to renegotiate every contract this quarter. You need to be deliberately positioned, so that the next renewal and the next RFP are conversations you run rather than react to. Ninety days is enough to get there.

Days 1–30: Map your spend by its unit of cost. Mark every service relationship — consultancy, agency, support, specialist firms, retainers — by how much of its price is anchored to time and effort versus delivered outcomes. The hour-heavy, high-spend relationships are simultaneously your biggest savings opportunity and your biggest exposure to a vendor that may be structurally unable to change. This map prioritizes where value-based buying pays off first.

Days 31–60: Define one outcome precisely, and run a value-based pilot. Pick a single engagement where the result is unambiguous and countable — a category of tickets, an invoice workflow, a development deliverable, a lead-qualification target. Write the outcome, quality bar, and measurement before you talk about price. Then engage a vendor on an outcome, gainshare, or hybrid basis and run it. The point is evidence: a real data point on cost, quality, and what verification the outcome requires.

Days 61–90: Rewrite one template and apply the four-question test. Update one contracting standard end-to-end — a performance-based SOW with an outcome SLA and a payment-on-result trigger. Get counsel involved while the stakes are a single pilot, not your core operations. Then bring the four-question test to your next renewal or RFP, and start sorting your relationships into "can genuinely price on value" and "selling me the label."

At the end of ninety days you will not have transformed your whole vendor stack. You will have something more useful: proof that value-based buying works in your own house, one rewritten template you can reuse, and a test that tells you which of your incumbents can come with you and which cannot.

Frequently asked questions

What is value-based pricing in professional services? Value-based pricing sets the price of an engagement by the worth of the outcome to the client, not by the hours or effort the vendor spends producing it. Instead of "this will take 200 hours at our rate," the conversation becomes "this result is worth X to your business, and here is what it costs to deliver it." It has always been the theoretically superior model; what changed is that agentic delivery finally makes it practical to run.

Why is the billable hour dying now, specifically? Because AI broke the link between time and value that the hour depended on. The hour was only ever a proxy for value, useful while output genuinely tracked human time. When agentic systems compress a multi-hour task into minutes at near-zero marginal cost, the hour stops measuring anything worth paying for. Industry data already shows the migration — for example, hourly billing among US accounting firms fell to under 4% in 2025 (Ignition's 2025 pricing benchmark).

Isn't fixed-fee pricing already the answer? Fixed-fee fixes your budgeting problem but not the alignment problem. Under a fixed fee, every hour the vendor doesn't spend is their margin, so the incentive is to do the minimum that passes acceptance. Value-based pricing is the only model where the vendor wins by making the outcome better, not by spending more or fewer hours — your interests and theirs point at the same thing.

How do agents make value-based pricing possible when it wasn't before? Value-based delivery requires cheap iteration — many rounds of "is this the value we wanted? adjust" until the outcome clears the bar. Under human-hours economics, each round was expensive, so vendors protected margin with the hour and change-orders. An agentic harness — the loop, guardrails, and quality gates around the agents — makes each round of feedback one more turn of the loop instead of a budget event, so the vendor can afford to iterate to genuinely good. Cheap iteration is the unlock.

What is an "agentic harness," and why should a buyer care? A harness is the infrastructure around the AI agents: defined roles, the tools each agent can use, the order of operations, human checkpoints, and the quality gates an output must pass to advance. Buyers should care because the harness is what makes outcome pricing safe — it is the mechanism that drives the work toward your defined outcome, catches bad direction in a single iteration rather than a quarter, and keeps a human steering at every checkpoint.

How do I start without renegotiating every contract? Run one pilot. Pick a single engagement with a countable outcome, define "done" and "good" and how you'll measure them before discussing price, and contract it on a hybrid (modest base plus an outcome or gainshare component). Use it to generate real evidence on cost, quality, and verification — then reuse the rewritten SOW/SLA template on your next renewal. Ninety days is enough to prove it in your own house.

The bottom line

For a century, the unit of cost in professional services was the hour, because time was the only honest proxy anyone had for value. Agentic delivery breaks the proxy: when output no longer tracks time, the hour stops measuring anything, and the determinant of cost moves to where it always belonged — the value of the outcome delivered.

That shift is not happening to you. It is leverage handed to you. For the first time, you can scope by the result you actually want, pay for the value you actually receive, and push delivery risk onto the party that controls it. And the deeper change underneath the savings is alignment: value-based pricing is the first model where your vendor wins only when you do. The vendors who built their cost structure on billable people will resist this because they cannot survive it — and they are often the ones charging you the most today. The vendors worth keeping will meet you on it, because their model was built for it.

At Facet, we have spent fifteen years guiding mid-market companies through structural shifts like this one — and we are running the same hours-to-outcomes transition inside our own house, pricing our work on the value it creates and delivering it inside agentic harnesses built to iterate toward exactly that. We are not selling you a change we have not made ourselves. The four foundations we have always built on — analytics, tooling, process, and automation — are exactly what let an outcome be defined, measured, iterated toward, and priced in the first place.

If you want to map your services spend by its true unit of cost — and design your first value-based engagement — request a procurement readiness session. We will show you where you are still paying for time you no longer need to buy, and help you build the outcome scope, pricing model, and contract terms your next vendor conversation requires.

---

Sources: Ignition 2025 pricing benchmark (US accounting firms: hourly billing under 4% in 2025, down from ~8%; fixed-fee ~54%). Futurum / IDC (Zendesk per-resolution AI agent pricing, ~$1.50 per autonomous resolution). McKinsey, The economic potential of generative AI (2023) — generative AI could add roughly $2.6–4.4 trillion annually across use cases. GitHub / Microsoft developer productivity research — developers completed a coding task ~55% faster with AI assistance. Editor: link exact reports before publish.