The AI Costs Nobody Puts in the Business Case
The licence fee is the number everyone quotes and the smallest one that matters. Here are the seven lines that decide whether it pays.
Erin Moore
Fractional Chief AI Officer
The subscription price is usually the smallest line in an AI business case, and the only one anyone quotes. Total cost of ownership includes integration, data preparation, the internal time nobody bills, ongoing maintenance, the review cycle, the exit, and the cost of being wrong — and it is the sum, not the licence fee, that decides whether the thing pays.
The seven lines
1. Licence or usage. The visible number. Watch for usage-based pricing where you cannot forecast the driver — that is an open-ended commitment dressed as a per-unit price.
2. Integration. Connecting the tool to your systems, handling errors, and putting the output somewhere useful. Frequently exceeds the first year's licence and almost never appears in the pitch.
3. Data preparation. Getting the inputs into a usable state. If your data readiness is weak, this is the dominant cost and it arrives before any benefit.
4. Internal time. The hours your team spends specifying, testing, correcting and re-training. Loaded cost, not salary. This is the line most business cases omit entirely, and it is often the largest.
5. Maintenance. Models drift, dependencies change, and someone must own the thing for as long as it runs. Budget a meaningful fraction of build cost annually or plan to abandon it.
6. Review and governance. Someone maintains the register, reviews the outputs, and answers the customer question about how data is handled. Small per tool, real across a portfolio.
7. Exit. What it costs to leave — data extraction, retraining people, rebuilding the workflow. Negotiate this at purchase, when you still have leverage.
The cost of being wrong
Separate from the seven: what happens if the output is wrong and nobody notices. For an internal summarisation tool, close to nothing. For anything touching billing, contracts or customer communication, potentially far more than the entire programme.
Size this deliberately rather than assuming it away. It is also the reason to start with reversible use cases — the argument in how to prioritize AI use cases.
Doing the sum honestly
The ROI calculator forces the baseline and the ongoing cost into the open, which is most of the work. The discipline that matters is including line 4 — internal time — because a business case that ignores it will show a return that never materialises in the accounts.
And whatever the sum says, record the baseline you measured it against. Without that, next year's review becomes an argument rather than a calculation. How to measure AI ROI covers the mechanics.
Frequently asked questions
What is the total cost of ownership for an AI tool? Licence plus integration, data preparation, internal time, maintenance, governance review and exit — plus the cost of the output being wrong. The subscription is usually the smallest component.
Which cost is most often missed? Internal time. The hours staff spend specifying, testing and correcting rarely appear in a business case, yet they are frequently the largest single line.
How much should we budget for maintenance? Enough that someone genuinely owns the system for its whole life — commonly a meaningful fraction of build cost annually. If nobody is funded to maintain it, plan for it to be abandoned.
Should exit costs really be in the business case? Yes, and negotiated at purchase. Data extraction and workflow rebuilds are cheap to agree before signature and expensive to discover at renewal.
Further reading
Tagged with:
Ready to Automate Your Business?
Let's discuss how AI automation can deliver measurable ROI for your organization in 90 days or sooner.