Outcome pricing

From Hour Producing Factory to Decision Power-as-a-Service

Hourly billing is dying. Clients want instant answers and outcome pricing is becoming the norm. Build one micro-agent, add a human review layer, track time-to-insight, and sell decision power-as-a-service instead of hours.

Bart van Maarseveen

14 Aug 2025 — 4 min read

Van Nelle Fabriek, Rotterdam (NL). © Bart van Maarseveen, 2006

Deadline: 31 December 2026. By then, outcome pricing will be the norm, and you’ll be competing with software that never takes a break.

Monday morning, 09:00

Daan, managing partner at a mid-size consultancy, pours his first coffee when Slack pings.

“Can we have a scenario this afternoon for how we should deal with the EU AI Act?” asks the CFO of a listed client.

Five years ago, Daan would have put a team of juniors and managers on this for eight weeks. Now he knows: if he needs more than a few hours, the client will hire an AI agent from a competitor tomorrow. The question isn’t whether this new reality will hit his business model, but how fast.

That single ping captures a broader crisis: clients expect immediate answers, margins shrink under AI automation, and talent opts out of repetitive work. The classic pyramid model —partner > manager > “army of juniors”— is cracking at every seam.

The breaking point

McKinsey: from army to core team

A Wall Street Journal case (August 2025) shows how deep the cut goes:

12,000 AI agents replaced routine tasks; headcount fell from 45,000 to 40,000 FTE in two years.
40% of revenue is already AI-driven; 25% of engagements are billed on concrete results.
Partners are bonused on AI adoption; the vision is one agent per consultant.

The classic revenue model (hours × juniors) is evaporating; McKinsey is now selling scalable decision capacity.

Monks: creativity under power

In the same year, Victor Knaap, founder of the creative network Monks, said that four out of five agencies without an “AI backbone” will lose their right to exist. Monks chose not to wait: copy, editing, and media planning were automated; clients now buy tooling plus embedded teams. The agency earns from licenses and scalable assets instead of linear campaign hours.

Together, these cases show that no professional service is safe: whether you deliver strategy, audit, or creative concepts, AI eats the repetitive work first, and after come the margins.

Why this hits your firm too

Clients want immediate impact. They experience real-time insights elsewhere and expect the same from you.
Margins erode. Every slide in a PowerPoint can soon be generated by an agent.
Talent wants meaning. Smart starters won’t accept “slide slavery”; they want to build agents, not spreadsheets.

You don’t need to be a futurist to see the pyramid model crumbling. The partner who bills juniors today will bill GPU minutes tomorrow—or lose the relationship.

How mid-size consultancies make the shift

To avoid paralysis from the scale of the change, Daan’s firm started with one pilot question:

“What is the impact of the EU AI Act on our operational risk processes?”

Three weeks later, a micro-agent was running that —fed with internal cases— delivering the analysis in under five minutes. Two senior consultants added a review layer; clients paid a subscription that included access to the agent plus two hours of human insight per month.

Results after three months

Time-to-first-insight: from 12 days to 4 minutes
Net Promoter Score: +18 points
Margins: +22% thanks to license revenue
Junior satisfaction: +30% because they became “agent trainers” instead of spreadsheet fillers

The firm now books this as a digital asset on the balance sheet: Scalable without extra FTEs. More importantly, it has a story new clients immediately understand.

From intention to action: your 30-day roadmap

Week 1 — Map the real problem.
Have every consultant log all client questions for a week. Anything that comes back three times deserves automation.

Week 2 — Build a decision tree.
Pick the top issue and draw the decision tree. Bring legal in early for privacy, IP, and bias checks.

Week 3 — Develop and test the agent.
One senior sketches prompts, a data analyst feeds cases, a conversation designer optimizes outputs. Test internally for latency < 1 second.

Week 4 — Go live with two pilot clients.
Measure time-to-insight (TTI), net promoter score (NPS), and annual recurring revenue (ARR). Let the numbers speak before you scale.

This sprint forces focus, limits risk, and produces an immediate business case.

But… compliance?

Precisely because AI projects can move fast, governance is essential. A simple pre-launch checklist prevents later pain:

GDPR impact assessment completed
IP rights on training data documented
Bias scan performed; mitigations documented
SLA for latency and uptime signed

Those who formalize this pull legal forward and accelerate adoption.

The confrontation

“Anyone still billing hourly rates on 1 January 2027 is faxing in a 5G world.”

Do you dare put that proposition on tomorrow’s leadership agenda? Deadlines create momentum; committees do not.

What you can do differently starting tomorrow

Log recurring questions systematically. Within a week you’ll see which 60–70% can be automated.
Start small, scale fast. One convincing agent builds more trust than ten half-baked pilots.
Price on decision capacity. Clients buy speed and certainty.
Measure value, not hours. Build dashboards for time-to-value, AI volume, and MRR.
Reposition juniors. Make them owners of data quality and prompt engineering.

The result: less project overhead, higher margins, and a scalable AI asset.

The promise of Decision Power-as-a-Service

McKinsey shows “AI-first” can work at global scale; Monks proves creatives aren’t immune. Between those extremes lies your opportunity: package your thinking into agents and sell decision power instead of hours. Build an asset that keeps working—even when the team sleeps.