From Billable Hours to Outcomes: The Rise of AI-Native Consulting

A few weeks ago we got pulled into a LinkedIn discussion with a handful of people in our network who, like us, have spent years inside the traditional consulting model and know its gaps intimately. The thread kept circling back to one thing: incentives, and specifically billable hours.

Talk to anyone at any firm and the messaging is identical. "We deliver outcomes, not invoices." But look closely at how the billable hour model works and the incentives point the other way. It rewards time spent, not necessarily value delivered, which means the steadiest path to revenue is more hours rather than better results. That is not a knock on the people doing the work. It is just what the structure pays them to do.

And yet this model, with roots going back more than a century, has until now been the best tool we have had for turning human expertise into predictable, billable revenue. This is starting to change, and AI is the reason.

The History: How Time Became the Product

Surprisingly, the billable hour model is not as old as most people think. For most of legal history, lawyers charged flat fees or worked off minimum fee schedules set by state bar associations.

The shift traces back to a Harvard-trained attorney named Reginald Heber Smith, who took over the Boston Legal Aid Society in 1913. Influenced by Frederick Winslow Taylor's scientific management movement, Smith built a timekeeping system that logged work in six-minute increments so he could see exactly where his team's hours went. The point was internal efficiency, not pricing.

Firms eventually realized the same way they used to manage and track efficiency could be used to bill, and Smith's timesheets and hours-and-tenths approach became the template. The legal community resisted at first, since many lawyers found accounting for their time beneath them, but by the 1950s and 1960s corporate clients had embraced hourly billing for the transparency and control it gave them.

From law, the model spread into financial advisory and general consulting for one obvious reason. When the thing you are selling is human expertise, time is the easiest input to count. How else do you price the work it takes to research, analyze, and ship advisory deliverables?

It Works Great, Until It Stops Fitting

For decades billable hours was the best available answer. Firms needed a way to price the labor behind advisory work, and clients who did not get what they wanted could walk to a competitor. That churn is the free market's way to punish firms who don't deliver. It still functions this way today.

There are exceptions worth naming. In corporate financial advisory especially, some firms win business almost entirely on reputation and scale. "No executive ever got fired for hiring this firm" is itself an outcome a buyer is paying for, and that is an enviable position to be in. No shade intended. That said, we have talked to plenty of executives who are tired of writing seven-figure checks and not getting the result they hoped for. The structure of the billable hour does not make that any easier to solve. (That is a thread worth pulling on its own, so we will save it for another post.)

The obvious alternative is to price on the outcome instead of the clock. The problem is measuring the outcome. Attribution is considered difficult now and it was harder in the 1950s and 60s, particularly for human capital work. Physical products are easier to monitor success with. Advisory is murkier.

The Catch With Outcome-Based Pricing

Software has been trying to crack outcome-based pricing for years, and most attempts have struggled for reasons that are well documented.

The first is incentive alignment. A vendor billing on outcomes still has to decide whether it is optimizing for the client's result or its own invoice, and different stakeholders inside the same buyer often define the "outcome" differently, which makes the metric subjective before anyone signs.

The second is predictability. Outcome pricing produces volatile revenue, which is a nightmare for the vendor's finance team and the buyer's budgeting process alike. Revenue volatility is consistently flagged as the single biggest risk in these models, especially for smaller firms that cannot absorb a bad quarter.

The third is adoption. If you only get paid when the client achieves a result, but the client never rolls out the technology or changes the workflow, you eat the loss for something outside your control. Attribution in environments where outcomes depend on tools, process, and human effort all at once is exactly where these deals fall apart.

This is why the monthly per-seat subscription won. It gives the vendor predictable recurring revenue and rewards it when a happy customer expands by adding seats. That logic is the backbone of the modern SaaS playbook, captured well in *Predictable Revenue* by Aaron Ross, which is still one of the better reads on the subject.

The per seat model is not a clean escape either, and its cracks are worth a post of their own. We are watching SaaS fatigue settle in the same way subscription fatigue hits consumers. Billy Howell described it well when he coined "SaaS fatigue": there is simply too much software, each tool solves too narrow a slice of the problem, and businesses burn hours stitching subscriptions together instead of doing the work. The data backs the shift. Seat-based pricing dropped from 21% to 15% of companies in a single year, because when one AI agent can do the work of ten seats, charging per seat means the better your product gets, the less you earn.

The Shift AI Actually Creates

Here is what changes the math. AI decouples the value of an outcome from the human hours spent producing it.

To be clear, this does not mean firing the experts. Systems built on AI still need a person in the loop to keep the work reliable and accurate, and the research is unambiguous on what happens when that breaks down. In the Harvard and BCG field study of 758 consultants, the people using AI on tasks inside its capability range did meaningfully better. On tasks that fell outside that range, consultants leaning on AI were 19% less likely to reach the correct answer than those working without it. The expert is not optional. The expert is what keeps the speed honest.

What AI compresses is the timeline. The months of research, market analysis, documentation, and slide production that used to define high-end advisory work can now happen in days, sometimes hours. And once the work is fast and the system is producing structured output, attaching a real KPI to it (time saved, error rate cut, throughput gained) gets a lot more practical than it was when everything ran through a human's calendar.

The Gap Will Close on Traditional Consulting

Right now, a lot of incumbent firms look stronger than ever, and for good reason. They are pairing a proven model with AI that collapses the timelines underneath it. AI is not yet priced into the advisory model, so much of that efficiency gain flows straight to the bottom line.

The numbers show how far the pricing lags behind the technology. The Big Four and the top consultancies have poured more than $10 billion into AI since 2023. McKinsey has added roughly 25,000 AI agents to its workforce in under two years and now counts them alongside its 40,000 people. And yet even at an AI-forward firm like McKinsey, only about 25% of fees are tied to outcomes. The rest is still traditional billing. The revenue model has not kept pace with the tooling, largely because partners are understandably reluctant to put fees at risk and the whole comp structure is built around hours.

That is the opening. The market tends to catch up. When it does, firms whose economics lean heavily on billable hours will face pricing pressure from entrants who can deliver a comparable outcome faster and at lower cost. It is simply harder for a firm built over decades around billable hours to become AI-native and outcome-based than it is for a new firm to start that way. And plenty of executives in the enterprise and mid-market would rather pay for the result than the hours it takes to get there.

This is the structural bet Emergence Capital laid out in its essay "The Death of Deloitte." The Big Four pulled in roughly $200 billion in revenue in a single year, but their economics run on human labor and hourly billing, which Emergence argues can be "turned on its head by AI-native vendors." Software incumbents are adapting to the AI wave without much trouble. Services incumbents have the harder problem, because their product is people, and you cannot refactor people the way you refactor a codebase.

There is also a new accountability pressure forming across the industry. One widely reported case saw Deloitte Australia refund AU$440,000 after AI-generated deliverables did not hold up, a reminder that unverified AI work is a profit-and-loss issue and not just a quality one. That applies to anyone moving fast with these tools, ourselves included, and it is exactly why expert oversight matters, a point we have made in our post Vibe Consulting.

The Rise of AI-Native Consulting

AI-native consulting is what fills the gap. The firms that win will be the ones that pair real talent with the right tooling and frameworks so a small team can deliver quality outcomes faster and more reliably than a large one running the old playbook. And when the work runs through systems instead of timesheets, outcome and value-based pricing finally becomes measurable, because the system can report on what it actually did.

The productivity ceiling here is high. In the same study, consultants using AI completed 12.2% more tasks, finished about 25% faster, and produced work rated roughly 40% higher in quality. The largest gains landed on the lower-performing consultants, who improved by around 43%. AI raises the floor of an organization, which is exactly what you want when you are trying to deliver consistent outcomes at speed.

We will not turn this into a pitch, but it is worth being concrete about what this looks like in practice, because it is how we run our own delivery. We are building internal tooling that generates proof-of-concepts directly from conversation transcripts, drafts scopes and quotes from engagement history, and stands up customer-feedback in sandbox environments before a line of production code ships. A human reviews every output and checks the AI-generated work. That oversight is not friction. It is the part that makes the speed trustworthy. The net effect is that our time to delivery continues to drop, without sacrificing quality.

When you can ship quality work that fast, the incentives flip on their own. You stop needing to defend hours and start being able to price on value.

It also changes the relationship. If you help an organization become self-sufficient in their own AI strategy, you become their AI plumber, the trusted resource they call when something needs fixing or building. Your incentives and the client's success line up by design rather than by slogan.

This is also where AI-native services start to do something SaaS never could. A typical SaaS product solves maybe 30% to 40% of a real problem, and that partial solution is easy to replicate with AI. Tailored solutions, built fast, solve the whole thing. SaaS won the last two decades on speed to scale. The next edge is speed to deploying solutions that actually fit. When services can scale the way software did and still deliver the specific outcome a buyer wants, that combination is hard to compete with. We dive into this deeper in our post Be like Water: the firms that adapt to the shape of the problem, instead of forcing the client into a fixed product, are the ones that take the market.

The same shift points to where the openings are clearest. The most disruptable work is repeatable and execution-heavy with outcomes you can actually measure, and the richest targets tend to be services that used to fail constantly. Emergence points to mainframe-to-cloud migration, where roughly half of the projects run by services firms failed before AI made that old code tractable. AI-native services do not just do the existing job faster. They make work feasible that was not feasible before.

The Second Monetization Revolution

The first monetization revolution in professional services was the billable hour. It gave the industry a clean, defensible way to price expertise when time was the only input anyone could reliably count. It served its purpose for a hundred years.

The second is starting now. For the first time, outcomes can be tracked, attributed, and measured well enough that value-based pricing is not just a nice idea but an operational reality. AI is what made that possible, and the firms built around it from the start will be the ones that benefit.

The hard part for incumbents is not the technology. It is the reinvention. Rewiring a comp structure and a partner culture that both assume billable hours is incredibly difficult, and it is more expensive than it looks from the outside. For an AI-native firm, none of that has to be undone. It just gets built right the first time. That asymmetry, between the cost of reinvention and the cost of building from the ground up, is the whole game. The future of consulting belongs to the firms on the right side of it.