Free — no purchase required
Two working Excel models
These are not lead-magnet summaries. They are the models, with live formulas, worked examples and every assumption documented. Use them whether or not you ever buy the manual.
The AI Cost Model
In the FinOps Foundation's 2026 practitioner survey, 98% of practices reported a mandate to manage AI spend — up from 31% in 2024. The most requested capability was granular monitoring of tokens, model requests and GPU utilisation. Commercial tooling has not delivered it.
- Model rate card, with your negotiated rates
- Use-case unit economics — cost per conversation, document, employee
- Allocation to business capability that reconciles to source
- Volume scenarios at 2×, 3× and 5×
- Model-substitution savings, quantified
- Dashboard: the four questions most organisations cannot answer
Software Capitalisation Evidence Pack
IAS 38 assumes projects with phases. Agile delivery has neither. This resolves the gap without asking a single engineer to record hours — evidence attaches to the backlog item, not to the person.
- Draft policy with twelve activity types classified
- Sprint tracker with classification dropdown and allocation
- Asset rollup with amortisation and sanity checks
- Audit trace — eight questions and where each answer comes from
- Judgement log for contemporaneous records
Send them to me
Both files, one email, no sequence of nurture messages. We will occasionally write about technology finance; unsubscribe in one click.
We ask for sector and reporting framework so we can tell you when the US edition is ready. Nothing else is inferred from it.
The rates in the AI Cost Model are illustrative placeholders, not quotations. They vary by region, commitment and negotiation. Replace them with your own before presenting any output from the model.