Five costed scenarios, assumptions on the table
ROI models built on public studies, with every assumption on display. Not client results: defensible orders of magnitude, designed to be discussed — and challenged — in the room.
How these figures are built
An AI ROI quoted without assumptions is worthless in an investment committee. So we apply the discipline of an industrial business case: bounded scope, explicit haircuts, ranges rather than single figures, and a public source behind every improvement rate we retain.
1. Bound the genuinely addressable scope
No study applies to 100% of your activity. We first strip out the share of tasks, hours or downtime the technology cannot touch. That haircut appears in plain sight in every table.
2. Use the published ranges, not the records
Gains come from peer-reviewed academic work or from consultancy publications, cited and dated. We take the low bound as the conservative scenario, never the record as the central case.
3. Three scenarios, never a single number
Conservative, central, high. If the decision only holds in the high scenario, it does not hold. The conservative case is what has to fund the project.
4. Deduct the full cost
Upfront investment, licences, integration, change management and the running cost of later years are all deducted. The return shown is net of those costs.
5. Show what would make the scenario fail
Every use case ends with its limits and its success conditions. A model with no failure conditions is a sales argument, not an analysis.
The studies cited measure productivity gains in specific settings. Transposing those rates to your organisation remains a hypothesis — which is exactly what a diagnostic is for.
What these scenarios are, and what they are not
These are models, built for illustration on public sector benchmarks. They are not results observed at Cardan-AI clients, and they constitute no commitment as to outcome. The improvement rates come from the studies cited; the choice of assumptions, the scope haircuts and the arithmetic are Cardan-AI's, and are shown in full so they can be challenged. Transposed to your organisation, these orders of magnitude can vary widely — in both directions.