How do success-fee AI transformation engagements work?
In a success-fee engagement the provider finds and builds improvements, the client's team verifies the value on its own numbers, and only then is a share of that value invoiced. What has to be true for that to work.
A success-fee AI transformation engagement is one where the provider is paid only from the value it creates, after the client's finance team has verified that value on its own numbers. Three mechanics make it work: a baseline agreed per opportunity before the build, sign-off by the client's team before any invoice, and a fixed term long enough for the value to reach the books.
Why the risk belongs with the provider
The base rate for AI projects is poor. IBM's 2025 CEO Study (May 2025), covering 2,000 chief executives in 33 countries, found only 25 percent of AI initiatives had delivered the expected return. BCG's Where's the Value in AI? (October 2024) found 74 percent of 1,000 surveyed companies had yet to show tangible value. Deloitte's State of AI in the Enterprise 2026, surveying 3,235 leaders between August and September 2025, found 74 percent hoping to grow revenue through AI against 20 percent already doing so.
Under a day rate, every one of those failed projects was paid in full. A success fee moves the loss to the party that chose what to build.
Every opportunity gets a baseline before anything is built
The engagement starts with discovery. The provider hears the whole organisation, ideally everyone rather than a sample of managers, and writes up every opportunity with a number: hours removed, cost removed, contract renegotiated, licence retired. The number is built from the company's own hours and cost figures.
If provider and client cannot agree what the invoice process costs today, they cannot agree what was saved. So the baseline is agreed per opportunity before engineers start. Opportunities the client does not want pursued are dropped, and cost nothing.
The client's team verifies the value before any invoice
Once a fix is running, the saving is measured on the client's books, by the client's team. The provider proposes the measurement. Finance checks it against the baseline. Only when finance signs does the value count, and only then is a share of it invoiced.
The provider therefore carries the risk of a fix that works technically and does nothing financially. A faster invoice intake feeding a four-day approval chain is a technical success and a financial zero, and under this model it earns nothing.
The term has to match how value shows up
Value from process redesign takes months to appear. A retired licence shows up at the next renewal. Removed hours show up when the headcount plan changes.
For owners the timing matters more than it used to. Bain's Global Private Equity Report 2026 (February 2026) calculates that a typical deal now needs 10 to 12 percent annual EBITDA growth to return 2.5 times capital over five years, against 5 percent in the 2010s.
That is why these engagements run about a year. A shorter term pushes the provider toward quick wins that fade. An open-ended term becomes a retainer by another name.
What is different from a day-rate engagement
| Day rate or fixed fee | Success fee | |
|---|---|---|
| When you pay | Monthly, from the start | After value is verified by your team |
| What you pay for | Time or deliverables | A share of measured value |
| Provider's incentive | More scope, more days | Fewer, larger, provable opportunities |
| Who defines success | Contract at signing | Baseline per opportunity, agreed before the build |
| Risk of a fix that changes nothing | Client's | Provider's |
| Typical term | Project length | Fixed 12 months, optional maintenance after |
Our view: attribution is only hard if you skip the baseline
The standard objection to outcome-based pricing is that attribution is impossible. It is true only when the baseline is set after the fact. Fix the process cost before the build, in the CFO's figures, define the measurement then, and attribution becomes arithmetic. A saving finance cannot verify should not be invoiced.
The model has a cost worth admitting. It only works where the accounts can separate cost by process and the finance team has time to verify, and the fit is weak below about $20M in annual revenue, where the opportunities are usually too small for a year of senior operators and engineers.
Questions people ask next
What stops the provider from inflating the savings figure?
The client's finance team signs off every figure against a baseline agreed before the build. If finance does not sign, nothing is invoiced.
What if the saving depends on a decision the company does not take?
A renegotiation the company chooses not to pursue, or a role it decides to keep, does not count as value and is not invoiced.
Is there really nothing to pay if nothing is delivered?
Nothing, if the contract is a genuine success fee. Watch for a minimum fee, a mobilisation charge or a right to re-baseline, each of which turns the deal back into a retainer.
What happens to the systems after the 12 months?
They were built into the client's own stack by the provider's engineers, who stay until they run. An optional low-cost Maintain phase follows.
What does the client have to provide?
Access to the cost data, time from finance to verify, and permission for the whole organisation to be interviewed.
Lightbloom AI starts every engagement with the discovery this model depends on: Transcript builds the world model of the company and puts a number on every problem before anything is built. How we work is at /service-offering.
References
- IBM Institute for Business Value and Oxford Economics, CEOs Double Down on AI While Navigating Enterprise Hurdles (2025), https://newsroom.ibm.com/2025-05-06-ibm-study-ceos-double-down-on-ai-while-navigating-enterprise-hurdles
- Boston Consulting Group, AI Adoption in 2024: 74% of Companies Struggle to Achieve and Scale Value, from Where's the Value in AI? (2024), https://www.bcg.com/press/24october2024-ai-adoption-in-2024-74-of-companies-struggle-to-achieve-and-scale-value
- Deloitte, State of AI in the Enterprise 2026 (2026), https://www.deloitte.com/us/en/what-we-do/capabilities/applied-artificial-intelligence/content/state-of-generative-ai-in-enterprise.html
- Bain & Company, Private equity resurgence gathers steam as new era challenges firms to enhance value creation, Global Private Equity Report 2026 press release (2026), https://www.bain.com/about/media-center/press-releases/2026/private-equity-resurgence-gathers-steam-as-new-era-challenges-firms-to-enhance-value-creationbain--company-global-pe-report/
Keep reading.
All notes- How do you capture tribal knowledge before someone leaves? Tribal knowledge is how the company actually runs, held by the people who run it. Why handovers fail, what works, and why it is the first layer of a world model. Read →
- Why AI agents need a world model of your company AI agents stall inside companies because nobody has described how the work runs. What a company world model gives an agent, and why to build it before you deploy. Read →
- What is a company world model? A company world model is a living picture of how a company really works: every process as it runs, what hurts most, and who everything depends on. Why AI agents need it. Read →