Let's be honest: Time & Material billing was always a hack. A workaround for a world where effort and output were tightly coupled. Where you needed ten engineers to do what ten engineers do.
That world is over.
The 100x Engineer Is Not a Myth
In 2024, GitHub reported that developers using Copilot completed tasks 55% faster than those without. That was the early days. Today, with autonomous agent frameworks — Cursor, Devin, Claude Code, custom multi-agent pipelines — a single senior engineer with the right setup can produce what previously required entire squads.
McKinsey's own research points to 45% productivity gains across development teams using AI augmentation. Some firms report 50–80% faster delivery timelines. Teams that have fully embraced agentic development talk about 10x–100x leverage on specific workstreams.
Think about what that means for T&M.
If you're billing by the hour, and your engineer suddenly ships in 10 hours what used to take 100 — you've just destroyed 90% of your own revenue. The incentive structure is catastrophically misaligned: the better your tooling, the worse your billing.
The Perverse Logic of Selling Hours
T&M made sense in the industrial model of software development. Bodies in seats. Hours logged. Outputs roughly proportional to input.
AI agents break this entirely. The production function has changed. The marginal cost of additional output is collapsing toward zero. And yet most agencies, consultancies, and freelancers are still billing like it's 2015.
Here's the uncomfortable truth for any CTO buying services: if your vendor is on T&M and using AI tools internally, they're either not using AI (and falling behind), using AI and pocketing the efficiency (while charging you full hours), or using AI and discounting honestly (but killing their own margins).
None of these is a healthy equilibrium. The model is broken.
What Value Based Pricing Actually Looks Like
VBP isn't new. McKinsey has linked roughly 25% of its global fees to outcomes as of late 2025. Management consultants have charged for transformations, not slide decks, for decades. But in software and tech services, it's been slow to catch on — until now.
1. Outcome Fees for Defined Deliverables
Instead of "200 hours at €150/h", the engagement is structured around outcomes:
- "We'll reduce your infrastructure costs by 30% within 90 days. Fee: 20% of the savings achieved."
- "We'll deliver a production-ready AI onboarding flow. Fixed fee: €25,000. We eat any overruns."
Real example: Boutique AI consultancies are already packaging solutions at fixed outcome fees — a dental clinic network paid €12,000 flat for an AI-driven patient intake system, 40% above the T&M market rate, because the outcome was clear and risk was shared.
2. Revenue Share & Success Fees
When the value created is measurable in revenue terms, take a cut.
Klarna famously replaced the equivalent of 700 human support agents with an AI system. The value wasn't in the hours it took to build — it was in the €40M+ in annual cost savings. A vendor pricing that on T&M would have captured maybe €500K in fees. On VBP? The upside scales with the impact.
E-commerce AI optimisation firms increasingly price at "X% of incremental revenue attributed to our AI" — trackable, defensible, and tied entirely to client success.
3. Per-Outcome / Per-Unit Pricing
This is value pricing at the infrastructure level. Vercel charges per build, per request. Zapier charges per task. These are not T&M — they are commoditised outcome pricing. The model works because the value per unit is real and measurable.
Legal tech firms like Luminance charge per document analysed, per risk flagged — not per hour of lawyer time. The client pays for the outcome (a cleared contract stack), not the process.
4. Retainer + Performance Bonus
A softer entry point for existing relationships: a base retainer covers availability, a performance bonus triggers when KPIs are hit — deployment velocity, uptime SLAs, conversion improvements, cost-per-ticket reductions.
This hybrid is how many established agencies are transitioning away from pure T&M without losing client trust overnight.
Why Now? The Agent Teams Tipping Point
The reason this shift is accelerating in 2025–2026 is not philosophical — it's structural.
Agent teams change the economics permanently. A single engineer orchestrating specialised AI agents (one for code review, one for testing, one for documentation, one for deployment) can deliver a sprint's worth of output in hours. The bottleneck is no longer developer bandwidth — it's clarity of requirements and quality of oversight.
When throughput decouples from headcount, T&M loses its anchor. You can't charge 500 hours for work that took 50. Well — you can, but your clients will eventually figure out you're arbitraging their ignorance. And in an era where CTOs are increasingly sophisticated about AI tooling, that window is closing fast.
What CTOs Should Be Demanding
If you're buying tech services right now, these are the questions worth asking:
"How are you using AI in your delivery process?" — Any vendor not using AI tools is already burning your hours inefficiently.
"If your team becomes 3x more productive, does our contract price drop?" — Under T&M, yes. Under VBP, no. The honest answer tells you which model you're actually in.
"Can we structure this as a fixed fee or outcome-based fee?" — If the vendor refuses, ask why. Either they don't believe in the outcome, or they know they're inefficient.
"What's your upside if this project wildly succeeds?" — Under T&M, none. Under VBP, aligned incentives. This answer tells you how much skin they have in the game.
The Transition Is Happening — With or Without You
BCG recently published that B2B software pricing in the AI era is fundamentally changing — moving from seat-based and time-based models toward usage-based and outcome-based structures. The same dynamic is hitting services.
Early movers — agencies and consultancies that have already restructured around VBP — are reporting 3x revenue expansion in 12 months, because they're capturing a fraction of the value they create rather than pricing by the clock.
The 100x engineer isn't a future hypothetical. They're deploying right now, at firms that have already made the transition.
T&M is the old model. It served us well. But the world has changed — and the pricing model needs to change with it.
The most dangerous thing you can do as a CTO right now is sign a T&M contract with a vendor who's silently 10x-ing their output via AI agents — and keeping every cent of that margin for themselves.