TL;DR. AI isn’t breaking one pricing model in professional services. It’s destabilizing the value metric every model attaches to. The emerging consulting advice to assemble “a toolkit of pricing models” is one level too low; the real decision is what the price attaches to. The Big Four already ran the most obvious answer (outcome pricing) through the 1990s and 2000s, and watched it collapse into attribution negotiations whenever the invoice mattered. The path that holds up moves from hours to named deliverable to methodology, picking the injection point at the boundary, where the firm’s work product crosses into the client’s operations and stops short of the customer’s downstream value creation.
Professional services firms integrating AI tools face a challenge that goes deeper than adjusting rate cards or switching to project fees. The current narrative holds that firms should move from one dominant pricing model to “a toolkit of pricing models matched to service archetypes.” That framing misses the core problem. AI doesn’t break individual pricing models. It destabilizes the value metric that every pricing model attaches to — a pattern visible across AI monetization in B2B software as a whole.
At SPP, we have priced three forms of intellectual property for decades: products, services, and insights. Most pricing advice handles two. The real AI pricing transformation isn’t about selecting new pricing models. It is about changing what the price attaches to, and picking the injection point carefully.
The Big Four already ran the opposite experiment through the 1990s and 2000s. They tried to charge on outcomes deep in the client’s value stream, and watched it collapse into attribution negotiations at every renewal. That history is the lesson the current “outcome-based pricing” conversation keeps forgetting.
Why AI makes traditional services pricing obsolete
How billable hours became the universal services metric
Every services pricing construct built today inherits from a single value metric: hours. Rate cards price per hour. Project fees reverse-engineer from estimated hours. Retainers spread hours across months. Time-and-materials charges directly for hours consumed.
AI tools compress the hours needed to produce the same output. Claude can draft strategy frameworks in minutes that previously required days. Internal workflow automation reduces research phases from weeks to hours. Specialized AI tools eliminate entire categories of manual analysis.
The metric’s pathologies were visible long before AI. Anyone who has opened a legal bill has seen them: line items that bear no relation to value delivered, surprise overages on scope that felt fixed, and a structural incentive for the firm to bill more hours rather than work faster. Clients tolerated it because there was no alternative metric to point at. AI removes that tolerance by forcing the question — what is the firm actually selling, if not the hours?
This creates an economic squeeze. Compressed hours mean lower revenue on the existing metric. Firms face what appears to be a binary choice. Raise rates on fewer hours (customers push back) or find new metrics to charge against.
Why the squeeze is asymmetric across competitors
The squeeze accelerates because efficiency gains are asymmetric. Every AI-using competitor compresses the industry’s delivery-time baseline. A firm that takes six weeks for market analysis competes against firms delivering the same scope in two weeks using AI research tools. The client’s expectation resets industrywide.
The software industry has run a similar restructuring before. When the per-license capital purchase became unworkable for buyers under the budget pressure that followed Black Monday in October 1987, the industry didn’t change the unit it counted — it changed how that unit was billed. SPP’s early financial-overlay work reframed on-premise license sales as subscription billing, shifting software from capital expenditure to operating expense and opening a different set of customer budgets. That restructuring laid the commercial groundwork for what later became SaaS. The lesson for services pricing under AI: restructuring how a unit is billed buys time, but eventually the unit itself becomes the bottleneck — and that is where hours now stand.
The false promise of “value-based” pricing transformation
Most pricing advice for AI-disrupted services suggests moving to “value-based pricing.” That generic framing fails because it doesn’t specify what value gets measured or how firms capture it.
A fixed fee reverse-engineered from partner hour estimates is still hour-based pricing. The accounting changes, but the underlying metric doesn’t. Clients still evaluate the engagement against time invested, and firms still scope projects using internal hour calculations. The disconnect shows up the moment a client pushes back on scope.
What we have observed across pricing engagements is consistent: when a metric becomes commoditized, pricing eventually has to migrate to a higher-order measurement that better reflects actual value creation. The pattern shows up across industries — compute-hours in cloud services, advisor-hours in professional services, machine-hours in manufacturing — and the firms that move first capture the new value while the laggards keep negotiating the old metric down.
Does Your Current Metric Connect to Customer Value Creation?
Generic value-based pricing fails because it doesn’t specify what value means for your customers. We assess whether your metric captures how customers actually derive value from your solution.
Why outcome-based pricing already failed once
The services industry has already run this experiment. Before AI, the Big Four firms and systems integrators spent the 1990s and 2000s trying to replace the billable hour with outcome pricing, charging on revenue uplift, cost reduction, shareholder-value delta, any metric that looked like it tied fees to engagement impact. It didn’t stick. Understanding why is what separates a pricing thesis that works from one that replays a twenty-year-old failure.
The EY billable-hour math: $1.4M billed, $15K bonus
I worked inside Ernst & Young earlier in my career. In one year, that work generated $1.4 million in billable fees and earned a $15,000 bonus. The math tells you exactly where the value capture sits in the billable-hour model, and it is not on the side of the people doing the work. That gap pulled me out and into running a B2B software company — which eventually became an SPP customer, and is how I ended up here.
The attribution negotiation pattern
The more instructive experiment wasn’t billable hours. It was the repeated attempts to move off them. Every Big Four firm and major systems integrator tried some version of outcome pricing during those two decades. The same failure pattern showed up every time. The moment the invoice crossed a meaningful threshold, clients would litigate attribution line by line.
The client argument was always the same. Yes, revenue went up. But your team only contributed to part of it. Our soil, our people, our market, our execution. You don’t get to claim the full outcome. The fee doubles. The fee is now five times the hours worked. The client argues their own contribution was decisive, the consulting firm’s contribution was marginal, and the right number is something closer to cost-plus. That dissection (what’s your value, what’s our value, how do we split it) is always, functionally, a negotiation to reduce the bill.
This matters for every services firm now considering outcome-based AI pricing. The same dynamic is already playing out with outcome-based software pricing. Companies that picked an outcome metric too deep in the customer’s value stream are finding their pricing becomes a renegotiation surface every renewal.
The fix isn’t to abandon outcome-proximity. The fix is to pick the injection point carefully.
This is the central question of The Value Metric Decision — too upstream and you charge for inputs the customer hasn’t acted on yet (per-hour, per-seat, per-API-call); too downstream and you’re claiming a share of the customer’s own harvest, which provokes attribution fights at every renewal. The defensible point is at the boundary — where the river leaves the mountain — where what your firm produced becomes a thing the customer can hold in their hand.
Where that boundary lands looks different for every firm. The work to find it is the same: extracting the repeatable parts of the engagements you already run, so the outputs of that work — whatever form they take — can become what the price attaches to, while inputs (hours, FTEs, time-on-task) move to internal cost-tracking rather than the external pricing basis. Picking outcomes too deep (customer ARR lift, retention improvement, shareholder value) replays the 2000s failure. Picking outputs too shallow (hours, FTEs, time-on-task) keeps the firm on the eroding metric AI is collapsing. The winning point is somewhere in between, and where it lands inside that range is a firm-specific question.
Building durable AI pricing for services: the crystallization path
The sustainable direction of travel is from charging for inputs (hours, bodies, time-on-task) to charging for insights (the answer, the framework, the defensible recommendation). What that movement looks like in practice isn’t a playbook anyone can run from the outside. Every firm’s engagement shapes, buyer relationships, and IP position create a different translation. But the shape of the movement is consistent.
Hours → Named deliverable → Methodology
It typically moves through three recognizable stages, in whatever sequence makes sense for the firm. Engagement patterns that used to be priced in hours get named and scoped, so the client buys a defined deliverable, and hours become the internal cost-tracking mechanism rather than the external pricing basis. Over time, repeated versions of the same named deliverable reveal an underlying framework the firm is actually selling beneath the surface of each engagement, and that framework, once it is stable and defensible, becomes sellable IP in its own right. From there, the productization path opens. The framework can be licensed, embedded in tools, or delivered as standardized work. None of those prices on hours. EY’s Fusion methodology is a publicly visible example of a firm that travelled this arc. It is not a template others can copy, but it illustrates that the arc is real and that it ends somewhere commercially durable.
Why the capture point matters more than the sequence
What makes the movement work isn’t the sequence. It is that each stage moves the capture point slightly downstream from inputs, but stops at the boundary where the firm’s work product crosses into the client’s operations. The assessment crossed. The framework is bought and used. The productized insight is licensed. Each is a concrete thing the client received and paid for, not a claim on what they then did with it. That is the injection point that defends the invoice. Our three-decisions framework forces firms to confront that question (what the price attaches to, and where along the value stream to capture it) before designing how to package or structure payments around it.
What makes this work, and why it isn’t a template
The sequence matters more than the timeline. Firms that succeed confront the value metric before they touch packaging, and touch packaging before they rewrite the pricing model. Skipping straight to “let’s try outcome pricing” or “let’s productize a fixed-fee offering” without settling what the price attaches to is how firms end up with repackaged hourly work. The billing structure changes; the internal scoping doesn’t. A 1,000-hour engagement gets sold as a $200K fixed-fee project, but the math underneath is still hours × rate. Clients can tell. At renewal they pull the time math back up — “how many hours did your team actually spend?” or “if next year’s scope is 70% of last year’s, why isn’t the fee 70%?” — and the conversation reverts to time-based evaluation regardless of what the contract says. The firm ends up defending its invoice with timesheets, which is exactly the trap the repackaging was supposed to escape.
What transfers across firms: the questions, not the answers
The actual decisions inside each step (which metric, what packaging, what price points hold under pressure) aren’t transferable between firms. A strategy consultancy and a research-intensive advisory firm can share the same framework and land on completely different metrics, because their engagement shapes, client relationships, buyer sophistication, and IP position are not the same. Published “how to move to value-based pricing in 12 weeks” content routinely underestimates this. Firms following a generic roadmap end up with generic outcomes.
What does transfer across firms is the set of questions worth confronting, and the order to confront them in. Getting the order right matters because each decision shapes the one below it. Picking a metric before understanding how buyers currently evaluate the engagement usually produces a metric buyers can’t estimate. Designing packaging before settling the metric produces packages that don’t differentiate on what the customer actually pays for. The frame holds across firms. The answers don’t.
The practical question for services firms isn’t whether AI will compress the metric under their current pricing. It will. It is whether the firm pivots to a defensible new metric before competitors do, or spends the next two years discounting against an eroding base.
That pivot is a value metric decision, not a pricing-model decision. It’s the kind of problem SPP has been working on with services firms across research, strategy, and specialist advisory practices for decades, and it is the work that gets operated continuously through LevelSetter, not as a one-time repricing project.
If your firm is watching AI tools compress the hours you used to charge for and you are not sure what metric to move to next, talk to an expert. We will walk through the crystallization path against your specific engagement shapes, and show how our approach handles the licensing-model decision that most pricing advice skips.
