What is repetition costing you?
Short answer
The honest way to size AI ROI is to start from the cost of the time already being spent. Take the people doing the repetitive work, their loaded salary, and the share of their day the process eats. That gives you a ceiling on the opportunity.
Then discount it by what a system can realistically absorb. This calculator assumes 35%, and shows you that it does. Every assumption is on screen, because a hidden constant is what turns an estimate into a sales tool.
Three inputs, one number
Move the sliders. The arithmetic and the assumption behind it stay visible the whole time.
01
People
Everyone whose day the process eats.
02
Loaded cost
Salary plus everything sitting on top of it.
03
Share of day
Hours on the process, over an eight-hour day.
04
Absorbed
35%. The part a system realistically takes.
Result
Ceiling
The most a year of this work can be worth. Never a forecast.
Addressable annual cost
$341,250
This is the payroll cost sitting inside repetitive work that a production AI system can take over. It is the ceiling, not a promise. What you actually capture is what we agree in step 02.
Assumes 35% of that time is automatable and an 8-hour workday. We run this properly against your real process data.
What this number is, and is not
A ceiling on one side of the ledger. Here is the rest of it, in both directions.
Left out, and it lowers the number
- Build cost, and the weeks before anything runs
- Inference and infrastructure, monthly and forever
- Integration into the systems of record
- Evaluation: the graded set, and the work to keep it honest
- Maintenance, once the project team has gone
Left out, and it raises the number
- Response time that stops being a queue
- Errors that stop reaching a customer
- Work currently dropped because nobody has the hours
- Decisions made on today's data rather than last week's
How do you calculate AI ROI?
Start from the fully loaded cost of the time being spent, not from a vendor's savings claim. Multiply the number of people doing the work by their loaded salary, by the share of their day the process consumes, by the share of that work a system can realistically take. Then subtract build and running cost. The output is a ceiling on the opportunity, not a forecast.
What automation rate is realistic for AI?
This calculator assumes 35% of the repetitive time is absorbed. That is deliberately conservative. Well-scoped, narrow processes with good data access routinely exceed it, and poorly scoped ones fall far short. Anyone quoting 80% without seeing your process is selling.
What does this estimate leave out?
Build cost, inference and infrastructure cost, integration work, evaluation, and ongoing maintenance. It also leaves out the upside that is not time: faster response, fewer errors, and work that gets done at all rather than being dropped. It is a ceiling on one side of the ledger.
Why does the calculator show its assumptions?
Because a hidden assumption is the difference between an estimate and a sales tool. Every constant used here is on screen and can be argued with. If you disagree with the automation rate, change it and the number changes with it.
Turn the ceiling into a committed number
A 30-minute call. We map one process properly, replace the assumptions with your real figures, and put the target in writing before any code gets written.