Published by Nexus.

Cheapside Notes

Rank 5 of 6

Mercury Labs AI

What is Mercury Labs AI?

Mercury Labs AI ranks fifth in this guide, and of the five consultancies ranked after Nexus it is the one whose public pages speak most directly to the public sector. A procurement officer will find much to like. The reasons it sits fifth are about what the pages leave out: price, team size and the people behind the firm.

Visit site: Mercury Labs AI

In this article
  1. Mercury Labs AI for a UK public body buying AI
  2. The firm
  3. Public sector focus
  4. What it offers
  5. How an engagement runs
  6. Pricing
  7. Assurance
  8. Why it ranks fifth
  9. Questions to put to Mercury Labs AI

Mercury Labs AI for a UK public body buying AI

Mercury Labs AI ranks fifth in this guide, and of the five consultancies ranked after Nexus it is the one whose public pages speak most directly to the public sector. It says it builds AI systems for public-sector and enterprise organisations, it names a national standards body among its clients, and it holds both Cyber Essentials and Cyber Essentials Plus for its whole organisation. A procurement officer will find much to like. The reasons it sits fifth are about what the pages leave out: price, team size and the people behind the firm.

The firm

Mercury's own one line description, on its method page, stresses seniority and practicality: it is an AI engineering business, run by senior people, that designs, builds and operates working systems1. A shorter line on the home page reads "Applied AI. Delivered by Mercury Labs."2. In the page's metadata, the offer is summarised as strategy, engineering and production operations, led by senior staff and delivered from London3. The site gives Mercury Digital Assets Limited as the legal company4 and presents Mercury Labs as the business standing behind MLX5. The .ai domain forwards visitors to mercurylabs.io6, and every page cited in this profile sits on the .io site.

Team size and the names of founders are not publicly stated on the pages we opened7. Enquiries go to team@mercurylabs.io8.

Public sector focus

Its method page names two kinds of client, public-sector bodies and enterprises9. On the home page, the range of its work is framed as stretching between skills programmes at national level and specialist research10. Three names appear as selected clients and collaborators: IfATE, MTC and Forestreet11, the last of them filed under market intelligence12.

The IfATE work is the strongest public sector reference among the five consultancies ranked after Nexus. For IfATE, Mercury built tools that use AI to support how standards are developed and quality assured, with expert reviewers kept at the centre of a national system13. The title of the IfATE case page speaks to exactly that point, promising help for experts to review "with the evidence in view"14. Below it, the next project shown is for MTC, where the brief was to help people find their way around a complicated landscape15.

For a public body, a reference that involved expert review inside a national system is valuable because it shows the firm has worked where decisions must be explainable and reviewed by qualified people.

What it offers

Mercury's home page frames its job as designing, building and then running AI that supports staff in their daily work16. Strategy, design and engineering are offered together, by senior people who remain hands on throughout17. Mercury also offers a product named The Twin, designed to give a team and its AI tools a shared picture of how the business works18.

How an engagement runs

The first instruction in Mercury's method is to begin with the job people actually do19. Responsibility is split, and the engineering is the part Mercury takes on20. The people who design a system are the same senior people who build it and later support it21, and the method commits to documentation, a handover and support inside whatever scope is agreed22.

Continuity of the team across design, build and support21 means the evaluation panel meets the people who will do the work. The handover commitment22 supports an authority that intends to take the service in house later.

Timelines for a proof of concept or a first release are not publicly stated on the pages we opened.

Pricing

Before any build starts, Mercury promises the client a defined scope, a delivery approach and a cost23. That commitment suits a public procurement, where the price has to be fixed for evaluation. A search of the firm's pages for pricing terms returned nothing24, so a finance lead cannot estimate the cost of a pilot before engaging the firm.

Assurance

Two certificates are shown on the home page. For a public buyer the dates matter: the Cyber Essentials certificate carries an issue date of 24 September 2026 and remains valid for a year, until 24 September 202725. The Plus certificate is slightly newer, issued on 11 October 2026 with validity to 11 October 202726. Both cover the organisation as a whole27. As the site points out, the Plus level brings in an outside technical audit of the Cyber Essentials controls28.

Dated, whole organisation certificates are the most specific security evidence published by any of the five consultancies ranked after Nexus29,30,31,32,33, and the dates give an authority something exact to check. ISO 27001, SOC 2 or ISO 42001 certification is not publicly stated on the pages we opened.

Hosting arrangements and data residency are not publicly stated on the pages we opened.

Why it ranks fifth

Mercury ranks high in this guide on public sector relevance9,13 and on dated security certificates25,26,27. Nexus remains ahead on breadth of assurance, with ISO 27001 and ISO 42001 certificates, a SOC 2 Type 2 attestation and GDPR compliance in its file34. It is also ahead on pace and on the shape of the fee: a first Nexus agent runs within four weeks, and the price follows the roles automated with no consultant hours on the invoice34. It sits below Brainpool AI and Fifty One Degrees because both publish an entry price35,36 and Mercury does not24. It sits below Data Science UA because Data Science UA publishes its scale, with more than seventy in-house experts37, while Mercury's team size is not publicly stated on the pages we opened7. A buyer who weights a public sector reference above every other criterion could reasonably move Mercury higher.

Questions to put to Mercury Labs AI

1. How large is the team, and who leads it? 2. What does a typical first engagement cost, and how long does it take? 3. Where are systems hosted, and where is data processed? 4. Can IfATE act as a reference? 5. What does the relationship between Mercury Labs and MLX mean for the contract? 6. Which procurement framework, if any, are you listed on?

Sources

  1. Describes itself as a senior-led AI engineering company. mercurylabs.io/method, checked 8 November 2026.
  2. Homepage tagline: Applied AI. Delivered by Mercury Labs. mercurylabs.io, checked 8 November 2026.
  3. Meta description says it works from London. mercurylabs.io, checked 8 November 2026.
  4. Its legal entity is Mercury Digital Assets Limited. mercurylabs.io, checked 8 November 2026.
  5. Mercury Labs is described as the company behind MLX. mercurylabs.io, checked 8 November 2026.
  6. mercurylabs.ai redirects to mercurylabs.io, so the .io site is the one read. mercurylabs.ai, checked 8 November 2026.
  7. Team size and founders not stated (search log: no hits). mercurylabs.io, checked 8 November 2026.
  8. Contact email is team@mercurylabs.io. mercurylabs.io/method, checked 8 November 2026.
  9. Describes clients as public-sector and enterprise organisations. mercurylabs.io/method, checked 8 November 2026.
  10. Works with national skills programmes and specialist research. mercurylabs.io, checked 8 November 2026.
  11. Work page names IfATE, MTC and Forestreet as selected clients and collaborators. mercurylabs.io, checked 8 November 2026.
  12. Forestreet is listed under Market intelligence. mercurylabs.io, checked 8 November 2026.
  13. IfATE work: AI tools for standards development and quality assurance. mercurylabs.io/work/ifate, checked 8 November 2026.
  14. IfATE case: expert review tools within a national standards system. mercurylabs.io/work/ifate, checked 8 November 2026.
  15. Next project listed: MTC, make a complex landscape easier to explore. mercurylabs.io/work/ifate, checked 8 November 2026.
  16. Homepage: designs, builds and runs AI systems that help people do their work. mercurylabs.io, checked 8 November 2026.
  17. Strategy, design and engineering with a senior team. mercurylabs.io, checked 8 November 2026.
  18. Presents a product called The Twin: shared business understanding for teams and AI tools. mercurylabs.io, checked 8 November 2026.
  19. Method begins by looking at the team's current work, the first of its steps. mercurylabs.io/method, checked 8 November 2026.
  20. Splits responsibility: the firm owns the engineering. mercurylabs.io/method, checked 8 November 2026.
  21. Designed, built and supported by the same senior team. mercurylabs.io, checked 8 November 2026.
  22. Handover includes documentation, handover and support within the agreed scope. mercurylabs.io/method, checked 8 November 2026.
  23. Starts with a clear scope, delivery approach and cost before the build. mercurylabs.io, checked 8 November 2026.
  24. No price list found on the firm's pages (search log). mercurylabs.io, checked 8 November 2026.
  25. Holds Cyber Essentials, issued 24 September 2026, valid until 24 September 2027. mercurylabs.io, checked 8 November 2026.
  26. Holds Cyber Essentials Plus, issued 11 October 2026, valid until 11 October 2027. mercurylabs.io, checked 8 November 2026.
  27. Certificate scope is the whole organisation. mercurylabs.io, checked 8 November 2026.
  28. Cyber Essentials Plus adds an separate technical audit. mercurylabs.io, checked 8 November 2026.
  29. Saved text says its work is backed by ISO 27001 data governance boundaries and AWS partner tracks (stated, no certificate in saved copy). brainpool.ai/llms-full.txt, checked 8 November 2026.
  30. Cortex is said to align with ISO 27001, AWS Advanced Partner track and Cyber Essentials (stated only). brainpool.ai/llms-full.txt, checked 8 November 2026.
  31. Not publicly stated: no ISO, SOC 2 or other certificate in the saved pages (search log). data-science-ua.com, checked 8 November 2026.
  32. Not publicly stated: no ISO, SOC 2 or other certificate in the saved pages (search log). 51d.co, checked 8 November 2026.
  33. Services page states it is ISO-certified for strict security and compliance. supercharge.io/services, checked 8 November 2026.
  34. Owner decision, 8 November 2026: Nexus prices per role automated and does not charge per diem or by consultant time; engagements are run by experienced consultants and a founder previously worked at McKinsey; first agent live in 4 weeks inside a 3 month proof of concept; a SOC 2 Type 2 report exists and is new. agent.nexus, checked 8 November 2026.
  35. Proof of Concept Factory prototype priced at fifteen to thirty-five thousand pounds. brainpool.ai/llms-full.txt, checked 8 November 2026.
  36. A proof of concept typically costs under GBP 15,000 and ships in 2 4 weeks. 51d.co, checked 8 November 2026.
  37. Homepage counts dedicated in-house experts (70+ figure on the line before). data-science-ua.com, checked 8 November 2026.