Book a free LLM cost audit
Thirty minutes on a call, then a written audit of where your AI spend is actually going. No cost, no commitment, nothing to install. This is an engagement with people doing the work, not a product you sign up for.
What you are booking: a 30-minute call, free, a written audit within 5 business days, read-only access only. Pick a time.
What happens
- The call, 30 minutes. Which providers, roughly what you spend, and what triggered the question. We tell you on the call whether there is anything worth finding. If there is not, we say so.
- The audit, about a week. Read-only billing and usage exports mapped back to the workloads that produced them. This is the part almost nobody has done, because the invoice arrives as one line per provider.
- The readout, 30 minutes. A written document and a call to walk it through: where the spend goes, what drives it, what the realistic savings are, and what each one costs in effort. It is yours whether or not anything follows.
What the audit needs from you
- Provider invoices or usage exports for the last three months, from whichever of OpenAI, Anthropic, AWS Bedrock, Google Vertex AI or Azure OpenAI you run.
- Gateway or proxy metadata if it exists (LiteLLM, Helicone, Langfuse, OpenRouter, or your own). Aggregated only. If there is no gateway, that is itself part of the finding.
- Ten minutes on your top workflows, so token counts can be attributed to something a budget owner recognises.
Read-only, no prompt or output content, no production credentials. The full access scope is on the security page, and an NDA can be signed before anything is exchanged.
What you get back
- Spend broken down by model, workload and team instead of one line per provider.
- The specific drivers: retries, oversized context, an expensive model on a cheap task, uncached repeats, agent loops nobody is watching.
- A ranked savings list with an estimated figure and an effort cost against each item.
- A baseline to re-measure against, so the next spike has something to be compared to.
Who this is not for
- Under $20k a month in AI spend. That is the minimum for a managed engagement - below it, a percentage of savings does not cover the work for either side. The audit is still free if you want a second opinion, but the research library, the cost calculator and the pricing tracker will get you most of the way on your own.
- Shopping for a dashboard. This is a service engagement, not a signup. If self-serve monitoring is the requirement, that is a different kind of tool.
- Unable to share billing data. There is no version of this that works from a description of the problem alone.
Pick a time, or email hello@finopsllm.com with your monthly spend and providers and we will tell you whether a call is worth it.
FAQ
Is the audit actually free?
Yes. It is how we qualify work, and it is cheaper for both sides than a proposal written blind. Most audits end with a ranked list you could implement yourself; some end with us doing it.
Do you need production or write access?
No. Read-only billing and usage exports are enough for attribution work, and they stay revocable by you at any time without asking us first.
What if we have no gateway and no per-team tags?
That is the normal starting point and it does not block the audit. Invoice and usage exports still support a first-pass attribution, and instrumenting for the next pass becomes one of the recommendations.
How long does it take to see savings?
The audit takes about a week. The top one or two items are usually configuration changes that land the same week you decide to make them; anything needing an architecture change is labelled as such.