Idea 03 · AI Budgeting, a working name

Know what your AI costs, before you spend it.

Budgets, projections and model choices for AI work, in a tool a finance lead and a junior can both read.

Reconcile the month's invoices

A person£70
AI£180
Do this with a person £110 cheaper by hand

A scripted demo with illustrative numbers. Pick a task and the verdict changes; sometimes the person is cheaper, and knowing when is the point.

Nobody can say what the work actually costs.

Nobody can say what a task costs when an agent does it, or when a person should do it instead.

Model choice is folklore. Why the expensive model here, the cheap one there?

AI budgets get approved without projections, and reviewed without accountability.

How it works.

01

Describe the work

Tasks, volumes and the quality bar. Plain inputs a finance lead and a junior can both fill in.

02

See the real comparison

Human cost, AI cost per model mix, and where each one wins.

03

Get a budget you can defend

A budget, a model policy and a monthly variance view anyone can read.

Why this works.

Humans are often cheaper. Knowing when is the product.

The uncomfortable truth the tool is built on.

Model mixing is where the savings hide.

A cheap model for retrieval, an expensive one for synthesis. Today that decision is made by vibes.

It pairs with visible evals.

One model does the work, another checks it, quality and cost on the same dashboard.

The business.

Team subscription, priced to scale with the AI spend under management. A plain B2B tool.

Questions.

Which models do you cover?

Every major provider, plus your own rates for people.

Is this a cost tracker?

It projects before you spend, not just reports after.

Is it for finance or engineering?

Both. That is the gap it closes.

Humans are often cheaper. Knowing when is the product.

All ideas