About AI Spend Audit
A hands-on audit service for teams whose AI and dev-tool bills have outgrown anyone's attention
Where the money usually goes:
Unwatched API Spend
LLM API bills grow month over month with no one checking which models, prompts, or retries are driving the increase.
Unused Licences
Premium coding-tool seats allocated to developers who haven't opened them in months.
Duplicate Tooling
Two or three tools bought by different teams that do substantially the same job.
Unmanaged AI Adoption
AI tools adopted organically, without governance, cost controls, or visibility into what's actually delivering value.
AI Spend Audit is a concierge audit service. Every audit is carried out by hand and reviewed personally before it reaches you — there is no dashboard to log into and no automated scan behind the scenes.
We run two audits, covering the two ways engineering teams tend to overspend on AI:
API Cost Audit
Your OpenAI, Anthropic and other LLM API billing. We look for wasted tokens, inefficient prompts, missing caching, retry storms and model-routing choices that cost more than they need to.
Dev Tool Spend Audit
Your Cursor, GitHub Copilot, Claude Code and Windsurf seats. We look for tier mismatches, dormant premium licences, overage leakage, duplicate coverage and credit-pool waste.
How it actually works
- You enquire: tell us about your stack and roughly what you're spending. No account needed.
- We confirm scope: we reply with payment details and exactly which billing exports or usage data to send.
- We review it by hand: your data is read and analysed personally — not fed through a scanner.
- You get a written report: findings, prioritised fixes and worked code examples, delivered within 2 working days of receiving your data.
Because every audit is done by hand, we take a limited number at a time. If we can't find meaningful savings in your data, we'll tell you that plainly rather than pad a report.