Value
If the initiative cannot reduce cost, increase revenue, or create capacity, it should not be funded.
56%
of CEOs say AI has delivered neither higher revenue nor lower cost so far.
PwC, 29th Global CEO Survey1Business cases for AI often stack optimistic time savings, ignore supervision, and treat the launch invoice as the only cost. The review then becomes a negotiation about the story, not a decision about the work.
How I see it
Fund the work only if the economics clear the bar
An ROI assessment should be conservative on purpose. Use the time people actually spend on the workflow, not the time they wish they spent. Apply an addressable share that leaves room for exceptions. Price the human review you will still need.
Include the cost to build, integrate, evaluate, and operate. Token spend is rarely the number that decides the bet. Integration, change, and the people who will own the queue usually are.
Payback is more useful than a five-year NPV theater. If the conservative case does not return the investment on a timeline the business can live with, the idea is not ready. That is a successful assessment.
The output is a decision memo: fund, reshape, or stop. It should be short enough that a COO can read it once and know what to do.
How the engagement runs
What you actually buy.
A service is a sequence with an artifact at each step. It is not a transformation program you purchase as a bundle.
Baseline the current cost
Time, volume, error, and capacity as they are today — not a target invented for the deck.
An owner already manages the number.
A benefit story with no starting point.
Build a conservative case
Addressable share, implementation cost, and run cost. If it is not material, stop.
A number that survives a skeptical operating review.
An ROI slide that ignores the cost to operate.
Get the win into the operating numbers
The result has to show up in cost, revenue, or capacity someone already reports.
The win is managed after the project team leaves.
Adoption of a tool as the only metric.
Do not buy this for
These engagements fail for predictable reasons.
Using best-case time savings
If every person saves 40 percent on day one, the model is a wish. Use a conservative slice of a measured baseline.
Ignoring the cost to keep it alive
A cheap pilot with an expensive operating tail is a bad investment.
Treating ROI as a justification document
If the project is already approved, you need a kill test, not a prettier spreadsheet.
Before we start
Questions that decide whether to engage.
What is the baseline in units a finance partner will accept?
Hours, cases, error rates, and loaded cost. Not sentiment.What share of the work is honestly addressable in the first release?
If the model needs 80 percent to work, the bet is fragile.What will it cost to run, review, and fix the path for twelve months?
If that number is unknown, the ROI is incomplete.How long until payback on the conservative case?
If payback requires a second transformation phase, do not fund the first.What result would make you stop after the first slice?
No kill criteria means the ROI will not be used.
How we size it
Conservative ROI(annual conservative value − annual run cost) ÷ (build + change + first-year extras)
If the ratio only works with aggressive adoption and zero exceptions, do not fund it.
This is the wrong engagement if
Do not start this engagement yet.
- Nobody will share cost or volume data.
- The initiative is already a committed vendor purchase.
- Leadership wants a large positive number more than a decision.
- The workflow has no baseline and no plan to take one.
A useful next step
Bring the initiative you are about to fund.
We will pressure-test the value, the cost, and the payback, and tell you if the honest answer is no.
Discuss an AI opportunity