About
We built this for ourselves first
Rise10x is a UK AI cost engineering firm. The gateway, the cost telemetry and the evaluation gate we bring to client work are the ones we run on our own products — which is a materially different thing from having read the documentation.
Rise10x started from an uncomfortable question about our own software. We were running several AI products on Azure, the model bill was growing faster than the usage behind it, and although we could read the invoice we could not explain it. Nobody could say which feature was responsible for what share, or whether the expensive parts were the ones customers valued.
The tooling to answer that did not exist in a form we could buy. Observability products told us what calls cost. Product analytics told us what customers achieved. Nothing joined the two, and without the join the only available metric was cost per token — which is precisely the metric that tells you a cheaper model was a success while your retry rate quietly doubles.
So we built the join: a gateway that records what every call cost, an outcome event from the application saying whether the customer was actually served, and a ledger keyed on a shared task ID. Two independent writers, so neither could flatter the number. Then a replay worker to test cheaper configurations against real traffic offline, and an evaluation gate in CI so a saving could not quietly become a quality regression three sprints later.
That system is what this firm does for other people now. The order of events matters: we did not start a consultancy and then look for a method. We had a problem, built the instrument, and found that almost everyone running LLMs in production has the same problem and no instrument.
Background
Who does the work
Rise10x is led by a founder with more than seventeen years in software, cloud and product delivery, and it is deliberately small. Cost engineering is senior work — the value is in judgement, not headcount.
Software delivery, not slideware
Seventeen-plus years building and running business-critical systems. The recommendations come with an implementation, a runbook and a rollback path, because we have been the ones on call for them.
Azure-native by practice
Our own estate runs on Azure — Container Apps, Azure OpenAI, Postgres, Terraform, CI. Azure-specific cost work is the part we know from operating it, including the parts the documentation does not mention.
Product discipline
We build and operate our own products, including Verdian, a legal workflow platform. Shipping to real users is what makes us fussy about evaluation gates, canaries and one-line rollbacks.
Position
What we are, and what we are not
What we are
- A specialist firm that reduces the cost of running LLMs in production.
- Independent — no reseller relationships, no provider commission, no incentives.
- Measurement-first. Every recommendation arrives with the evidence that produced it.
- Finite by design. The retainer is optional and has to keep earning renewal.
What we are not
- An AI strategy consultancy. We do not run workshops about what AI could be.
- An observability vendor. There are good products; we will point you at one.
- A staff augmentation shop. The scope is cost and the guardrails that protect it.
- A firm that quotes savings before measuring. Any percentage offered before someone has seen your traces is a sales number.
Details
The boring but necessary facts
- Registered name
- RISE10X AI LTD
- Company number
- 16411324
- Registered office
- 128 City Road, London EC1V 2NX
- Jurisdiction
- England and Wales
Everything we would say on a first call is already written down in the guides, and the method is set out step by step in how it works.
FAQ
About the firm
How big is Rise10x?
Deliberately small. Cost engineering is senior work — the value is in judgement about which lever is worth pulling and how to prove it, not in the number of people on the call. Engagements are staffed by the people who do the work, and we would rather turn one down than staff it thinly.
Do you have case studies?
Not published ones, and we would rather say that than dress up a number. Client cost data is commercially sensitive and most of it is under NDA. What we can do on a call is walk through the method against your architecture and show you the tooling running on our own estate, which is a more useful hour than reading a sanitised case study.
Why should we trust a firm that publishes its whole method?
Because the method is not the moat. Every technique on this site is documented somewhere in a provider's own guidance. The hard part is finding which of a dozen levers matters in your specific estate, sizing each one honestly, and proving the risky ones before a customer is affected — and that is work, not a secret.
What is Verdian, and how does it relate?
Verdian is a legal workflow platform we built and operate. It is one of the AI products running behind our own gateway, and it is where a good deal of this tooling was first proved out under real production traffic rather than in a lab. Product-building discipline is what makes us fussy about evaluation gates and rollback paths.
Find out what your AI actually costs per completed task.
Two weeks, a fixed fee, and a ranked savings plan with the quality risk of every move stated up front. If the numbers say an audit is not worth it for you, we will say so on the first call.