ChatGPT prompt for auditing your software subscriptions and tools

This is the prompt we run every quarter to take back control of the software line item. It groups each subscription into four buckets and totals the possible saving. Its rule: a tool whose usage is uncertain goes into “needs investigation,” not “cancel.” The prompt never cuts blind.

The prompt to copy

Here's our list of software subscriptions with monthly cost and rough usage. Group into: keep, downgrade candidate, cancel candidate, needs investigation — with a one-line rationale each and total potential monthly saving. List: {paste}

Replace the field in curly braces with your list of tools, each with its monthly cost and a sense of real usage, then run.

When to use it

Every quarter, or whenever the software line in your accounts makes you wince. In ten minutes, you go from a list of charges nobody reviews to a clear grouping with a saving figure at the bottom.

What you get back

1. Each tool sorted into: keep, downgrade, cancel, needs investigation.
2. A one-line rationale for each proposed decision.
3. The total potential monthly saving, quantified.
4. An honest bucket, “needs investigation,” for anything with unclear usage.

Real example: subscription list → savings plan

Input: at Brightside Studio, an 8-person brand agency, ops manager Tom pastes the list of tools. An excerpt:

Figma — £45/mo — design team, daily
Survey tool — £29/mo — last used months ago?
Pro email suite — £60/mo — essential, everyone
Mockup generator — £19/mo — Priya thinks we've stopped using it

Output, the part that matters:

Tool	Grouping	Rationale
Figma	Keep	Daily use by the whole design team, core to production.
Pro email suite	Keep	Essential, used by everyone.
Survey tool	Needs investigation	Usage uncertain: check the last login before deciding.
Mockup generator	Needs investigation	Doubtful but unconfirmed: ask Priya before cutting.

Notice what the AI did not do: it did not cancel the survey tool or the mockup generator on the grounds that they “seem” unused. Tom flagged a doubt, not a certainty, so both go to “needs investigation.” That is the prompt working: it quantifies the possible saving without pushing you to cut a tool someone may still depend on.

This prompt is part of the library that ships inside the kit, with its usage note.