AIImplementationCost

Guide

Measuring AI return without fooling yourself

Updated

The benefit half of an AI business case is usually weaker than the cost half, and it fails in a predictable way.

Talk to a specialist

Saved time is not saved money

If a task takes twenty minutes instead of an hour and the person simply has an easier day, the business has spent money and gained nothing measurable. The saving becomes real when the time is redeployed to something valuable or when the headcount requirement genuinely changes.

Decide which of those two you are claiming, before the project, and say so out loud. Projects that skip this step cannot be evaluated afterwards.

Measure the before

You cannot show an improvement you never baselined. Time the task, count the errors, measure the cycle time, before anything is installed.

This takes days and it is the difference between a defensible result and an anecdote.

Beware the quality trap

Faster and worse is easy. Any benefit measure should be paired with a quality measure, or the project optimises the thing you counted at the expense of the thing you wanted.

Decide what a wrong output costs and how you would detect one, because that determines how much checking the process needs and therefore how much time is really saved.

Count internal time honestly

Your own people working on the project are not free. Cost their days at the same rate you would pay externally, or the business case is comparing a real number with an imaginary one.

This single correction reverses a surprising number of enthusiastic business cases.

Want this costed against your actual stack?

The model uses multiples. What it really costs depends on your data, your systems and what you are trying to automate. Tell us the shape of it and implementation partners will scope it.

Tick the box above to send your enquiry.

  • Free, and no obligation
  • Your details go to at most three partners
  • We do not sell software and take no vendor commission

Budget four lines, not one

Licence, data, integration and change, plus the benefit test that decides whether saved time is worth anything.

Talk to a specialist Open the calculator

Sources

This site's own datasets, every figure quoted from its source and dated:

  • The median published figure in the AI Adoption and Spend is 19.95 percent across 3 sources' own pages (range 19.80 to 35 percent), read 12 October 2026.
  • For chatbot, the median published figure in the AI Project Cost Index is 20,000 USD across 11 figures from the sources' own pages (range 1,500 to 300,000 USD), read 12 October 2026.
  • The AI Talent Pay Benchmark carries 2 published figures read from 2 sources' own pages, from 120,230 to 140,300 USD, read 12 October 2026.
  • The GPU Compute Price List carries 15 published figures read from 5 sources' own pages, from 0.48 to 12.42 USD, read 12 October 2026.
  • The Model API Price List carries 3 published figures read from 3 sources' own pages, from 0.02 to 0.50 USD, read 12 October 2026.
  1. NIST, AI Risk Management Framework
  2. SBA, hire and manage employees
  3. FTC, advertising and marketing business guidance
Talk to a specialist