AI Cost Defensibility Assessment
If a client, an investor or an auditor asked what a matter cost in AI, would your number hold up?
A section is a part of the one Assessment, not a choice between two: every section opens at the close of your Baseline.
Nothing routes through us.
Add
One line beside the provider client you already use.
Report
The envelope of each call, on the side, after it returns.
Keep
Your prompts, your replies and your keys never leave your process.
In the report.
For firms that bill someone else for work agents did.
- What one is for you: a matter, a claim, a case. Declared by you at intake.
- All-in AI cost per one, over the window. The middle, the tail and the worst.
- One real matter end to end: every call carrying its tag, and what it prices at.
- Where it would not hold: calls under another tag, models with no published price, code with no tap line.
- The rule that assigned a shared call to one unit, written so a third party can reject it.
- Five moves, in order, to start on Monday.
Every section is computed from the calls your own application reported.
The other section of the same Assessment: Agent Exposure Assessment
It ends in one offer: the plan that fits your band, monthly or annual.
- The price
- The price of a plan is its published band price, and it never depends on what your Assessment found.
- No trial
- A plan started from your Assessment has no trial, because your Baseline week was one.
The bands and their prices are on the pricing page.
What’s already true
- An assessment never asks for a provider key.
- The metering schema has no field that can hold a prompt or a reply.
- Receipts are on by default, on every plan — Free included.
- Admission runs before your provider does.
- Designed to fail open.
- Unlimited seats on every plan.
- We take no money from a party we rate.
- You’ll have a reply within one business day.