The consulting business has been a beautiful little money machine for a long time.
A client has a hard problem. A firm puts smart people on it. Those people spend hours researching, analyzing, building slides, running meetings, and polishing recommendations. Then the firm bills for those hours.
AI makes that machine awkward.
The obvious version is about consultants. The more interesting version is about everyone whose work is still measured by time, presence, utilization, or the sacred corporate ritual of looking busy in a calendar block.
First up, the TL;DR
The old professional-services bargain was simple: smart people spent hours on hard problems, then billed clients for those hours.
AI makes that bargain weird.
WSJ reported that consulting firms are trying to move away from hourly billing as AI makes some work faster, cheaper, and harder to meter the old way. Business Insider reported a similar shift: clients increasingly want firms to put “skin in the game” through fees tied to results.
Here’s what happened:
- Deloitte reportedly showed consultants a chart suggesting traditional labor-based consulting could shrink sharply as a share of the market by 2035.
- AI agents are expected to become a much larger part of professional services.
- Firms are testing fixed-fee pricing, where clients pay a set amount for a defined project.
- They are also testing outcome-based pricing, where pay depends on agreed results.
- McKinsey says more than 30% of its global fees already come from pricing tied directly to client outcomes.
Why this matters: Consulting has a math problem. If AI lets a team finish a 40-hour project in 10 hours, clients will eventually ask why they are still paying for 40.
The replacement models are messy. Fixed-fee work can crush margins if the project drags on. Outcome-based pricing can trigger fights over what “success” means. Big Four firms also face audit rules that limit how much they can tie pay to client results.
But this is bigger than consultants. Billable hours punish fast workers because efficiency reduces the thing being sold. Salaries can do the same thing in softer form: finish your work early, and you often get more work, more meetings, or suspicion.
Our take: AI probably kills the cleanest part of the old professional-services model: selling time as a proxy for value.
The next model will likely look more like a base retainer plus upside for measurable results. Clients still want predictable costs. Workers still want predictable paychecks. The trick is building a system where speed creates shared upside instead of quietly becoming unpaid extra productivity.
The billable hour was built for a world where time was a decent proxy for effort. Agent swarms make that proxy look ancient.
How the billable hour became the default scoreboard
Most work gets paid in one of a few ways.
Hourly workers sell time directly. Salaried employees sell availability, role responsibility, and a rough bundle of expected output. Sales teams add commission. Startup employees may accept lower cash for equity. Consultants, lawyers, and some doctors built a prestige version of the same time machine: the more expert hours involved, the more expensive the work.
That worked because time used to be a tolerable stand-in for effort.
If a lawyer spent 10 hours reviewing a contract, the client understood the invoice. If a consulting team spent six weeks building a strategy deck, the client may have grumbled, but the logic was legible. Human labor was the scarce input.
AI changes the scarce input.
The scarce thing becomes judgment: what problem to solve, what data to trust, what answer to reject, what risk to own, and how much accountability the provider accepts when the work leaves the room.
That is why this feels bigger than a consulting pricing story. Consulting is simply where the contradiction is most visible because the invoice says the quiet part out loud.
What changed
According to the WSJ report, Deloitte showed consultants a projection that traditional labor-based, hourly-rate consulting work could shrink significantly as a share of the total market by 2035.
That same report said firms are trying to operate more like software or product businesses, selling fixed-fee subscriptions or packaged solutions instead of renting out human time.
Business Insider added a cleaner taxonomy. Consulting firms are increasingly using two buckets:
- Fixed-fee pricing: the firm scopes a project upfront, estimates the work, and charges one agreed price.
- Risk-based or outcome-based pricing: the client pays a base fee, then additional compensation depends on agreed results, like cost savings, efficiency gains, or revenue impact.
That second model is the one clients love in theory. It sounds fair: if the consultant creates value, the consultant gets paid. If the AI-powered miracle deck turns into a very expensive PDF with vibes, the client does not eat the whole bill.
The problem is that “value” is slippery.
If a firm helps improve sales, was it the new process, the market, the client’s team, better pricing, or a lucky quarter? If an AI transformation saves money, who gets credit: the consultants, the software vendor, the internal team, or the fact that finance finally stopped exporting seven spreadsheets named FINAL?
Not that any of us would ever do that. Obviously. Spiritually, yes. Legally, no.
The old model punishes speed
The deepest problem with hourly billing is simple: it punishes the person who gets good enough to be fast.
A junior person might need 12 hours to do a task. A senior person might need two. Under a pure hourly model, the junior person produces more billable time, even if the senior person produces better work.
AI exaggerates that absurdity.
A consultant with good AI workflows can compress research, analysis, drafting, and slide production. A lawyer can summarize case law faster. A marketer can generate 20 campaign angles before lunch. A finance analyst can reconcile data in minutes instead of a day.
The old system then asks a ridiculous question: should faster work be cheaper, even if the output is better?
Clients will say yes. Firms will say, “Well, hold on.” Workers will quietly wonder why learning AI made them responsible for twice as much work.
That last part matters.
Because inside companies, salaried employees face a softer version of the same trap. Many knowledge workers are paid for a role, but managed by visible effort: office hours, meeting attendance, Slack responsiveness, and whether their calendar looks sufficiently doomed.
A fast employee often earns more work. An efficient employee may get fewer boundaries. Someone who uses AI well can produce more, but the reward is frequently a bigger pile.
That is the corporate version of the billable-hour problem: speed creates value, but the value does not automatically flow back to the person who created it.
Why outcome-based pricing sounds obvious and gets messy fast
Outcome-based pricing has a clean moral logic.
Pay for the result. Reward the provider for impact. Stop pretending hours are the product.
McKinsey says more than 30% of its global fees now come from pricing tied directly to client outcomes, according to the WSJ story. BCG’s CEO told the Journal that three-quarters of its largest AI cases use variable-fee structures, according to Business Insider.
The appeal is obvious.
AI projects are uncertain. Clients are spending a lot and often struggling to prove ROI. Tying fees to outcomes shifts some risk to the firm promising transformation.
The messy part is execution:
- Measurement fights: Both sides must agree on what success means before the work starts.
- Attribution fights: A result may depend on market conditions, client execution, or internal politics.
- Cash-flow risk: Providers still need predictable income while waiting for outcome payments.
- Quality risk: Faster AI-assisted output can still be wrong, shallow, or hallucinated.
- Regulatory limits: Big Four firms face independence rules that restrict outcome-based fees in audit work.
That means the billable hour will probably survive for some work. Messy, high-stakes, ambiguous projects are hard to price by outcome because nobody can guarantee the outcome.
A trial lawyer cannot promise a verdict. A doctor cannot promise a recovery. A cybersecurity consultant cannot promise that nothing bad ever happens again.
But routine research, drafting, analysis, reporting, implementation, and project management are much harder to defend as pure time-based work once AI compresses the labor.
The legal world is seeing the same thing
The legal industry is a preview of the broader professional-services fight.
In Bloomberg Law, Ford general counsel Steven Croley argued that AI agents decouple time spent from market value. His point was blunt: AI can perform research, legal problem-solving, and writing at high speed, while still requiring lawyers to scope, direct, and validate the work.
That changes the buyer’s expectations.
Ford’s legal team, he wrote, plans to reward external counsel that uses AI fastest for productivity. It may also reduce reliance on firms that lag.
A separate Bloomberg Law piece quoted Cooley CEO Rachel Proffitt saying the billable hour will lose its central role over time, even if it does not disappear tomorrow.
That is the pattern to watch across industries.
AI does not erase expertise. It changes the packaging around expertise. The expert still matters, but the invoice needs a new explanation.
So what replaces time?
The honest answer: probably a messy bundle of models, not one perfect replacement.
1. Base retainer plus outcome bonus
This is the most plausible near-term model for consultants and agencies.
The provider gets a predictable monthly fee for availability, judgment, systems, and delivery. Then they earn upside for measurable results.
That preserves the thing everyone likes about fixed costs: planning. Clients can budget. Providers can pay salaries. Workers can count on income.
Then the bonus layer gives speed somewhere to go.
2. Productized services
Some consulting work becomes more like software.
A firm packages a repeatable outcome: AI readiness audit, workflow automation buildout, sales ops cleanup, legal contract review, finance reporting overhaul.
The client pays for the package, not the hours. The provider wins by improving the system behind the scenes.
This model rewards firms that build reusable tools, templates, agents, and data pipelines. It hurts firms whose main product is “many smart people spent many expensive hours on this.”
3. Capacity subscriptions
Some professional work may look like a subscription to a team plus agent stack.
The client pays for access to a certain level of managed capacity: weekly strategy calls, ongoing analysis, agent-run reporting, human review, and escalation when something gets weird.
This feels especially likely for AI operations, legal ops, finance ops, and marketing ops, where the work never fully ends.
4. Equity or upside sharing
Silicon Valley already solved part of this culturally: workers accept equity because they are helping build enterprise value, not simply trading time for cash.
That model could spread in smaller ways. Consultants might take success fees. Employees might get profit-sharing tied to team-level productivity gains. Internal AI builders might get bonuses for automations that save real money.
The danger is obvious. Upside pay can become a lottery ticket if the worker has little control over the outcome. Equity also does not pay rent unless there is liquidity.
Still, the principle matters: when AI makes output scale faster than hours, compensation needs some connection to the value created.
5. Results-based internal comp
This is the hardest version because companies love salaries.
Salaries are simple. They are predictable. They make mortgages and HR systems work.
But AI will pressure companies to rethink what the salary buys. A team that uses agents well may produce the output of a much larger team. If the company captures all of that upside, employees eventually notice.
One alternative is a salary plus project-outcome bonuses. Another is shorter workweeks when output goals are met. Another is team-based profit sharing. Another is explicit “automation dividends,” where employees who build durable AI workflows share in the savings.
The exact model matters less than the principle: efficiency should create shared upside, not just higher expectations.
The agent-swarm problem
The compensation debate gets even stranger once agents do meaningful work in the background.
Today, most AI productivity still feels like acceleration. You ask. It answers. You edit. You move on.
Agentic work is different. You might assign research, monitoring, QA, reporting, outreach, or analysis to a group of AI systems that operate while you do something else.
Your embodied time goes down. Your responsibility does not.
You still need to define the task, inspect the output, catch errors, make decisions, and own the consequences. But the work is no longer neatly attached to your sitting time.
That breaks the moral intuition behind time-based pay.
If you manage five agents that produce useful work overnight, did you “work” those hours? If a consultant builds a system that keeps saving the client money for six months, should payment stop when the implementation meeting ends? If an employee automates 30% of a workflow, should they get time back, more pay, or a larger job description?
Companies are about to face those questions everywhere, not only inside McKinsey decks and law-firm invoices.
Our take
The billable hour will fade where work becomes measurable, repeatable, and AI-compressible.
It will survive where outcomes are ambiguous, risk is high, and human accountability is the real product.
The better question is who gets the AI dividend.
Clients want lower costs. Firms want to protect margins. Employees want better tools without being punished for using them. Managers want predictable budgets. Nobody wants to replace one broken incentive system with a more complicated broken incentive system wearing an “outcomes” badge.
The likely answer is a hybrid: predictable base pay for responsibility, plus upside for measurable value, plus explicit protection against turning every efficiency gain into more work.
That could mean retainers plus success fees for consultants. Salaries plus team bonuses for employees. Equity or profit-sharing for people building durable systems. Shorter weeks when output holds. More pay for people who can orchestrate agent teams and validate the work.
The companies that figure this out will have an underrated recruiting advantage. They will be able to tell ambitious AI-native workers: when you make the work faster, you share in the gain.
The companies that do not will discover a fun little management puzzle: their best people used AI to do more work, then realized the reward was simply… more work.
