Search for an AI consultancy in the UK and every firm appears to sell the same thing: strategy, delivery, "responsible AI". Underneath the identical language, the market has just been reshaped. In the twelve months to spring 2026, Accenture completed its acquisition of Faculty, London's best-known applied-AI firm; CGI absorbed the 2,400-person engineering consultancy BJSS; and Mesh-AI merged into Indicium AI. The market is consolidating around scale at exactly the moment buyers' needs are fragmenting by sector and by job.
This guide compares ten firms a UK buyer will realistically meet, grouped by what each is genuinely best at. It is not a ranking. Every fact below was checked against the firm's own published material, Companies House records or press coverage in June 2026, and one entry is our own practice, disclosed where it appears.
Key takeaways
- The UK AI consulting market consolidated sharply in 2025–26: Accenture completed its purchase of Faculty in March 2026, CGI absorbed BJSS, and Mesh-AI became part of Indicium AI. Check who owns any firm on your shortlist before you sign.
- No firm is best at everything. Global transformation, public-sector applied AI, data platforms, deep-tech product work, governance software and board-level governance are six different jobs, and they suit different firms.
- "We advise" and "we build" are different capabilities. Ask every candidate for systems running in production, not frameworks on slides.
- The UK has no single AI Act, so a consultancy's value rests on whether it can map regulator-applied principles to controls your board can actually show.
- Almost no firm publishes prices; scoped engagements are the norm. Insist on a defined deliverable and a named owner for any diagnostic phase.
How we assessed the field
Five criteria, applied to every firm including our own:
- UK regulatory fluency. The UK governs AI through principles applied by existing regulators, not a single statute. Can the firm map those principles — and ISO/IEC 42001, the NIST AI RMF and UK GDPR — to controls a board can evidence?
- Build versus advise. Does the firm ship working software, produce advice, or both? Neither is wrong, but a buyer needs to know which they are paying for.
- Sector depth. Generic AI capability transfers poorly into regulated sectors. We looked for named work in the buyer's world: health, government, financial services, housing, charities, professional services.
- Board-level focus. Whether the firm engages directors and audit committees, or only data teams.
- Evidence of shipped work. Public case studies, named clients, products in use — not announcements of intent.
No firm paid for inclusion or was consulted on its entry. We appear in the list ourselves; the disclosure sits on that entry.
QuantumBlack, AI by McKinsey
QuantumBlack is McKinsey's AI arm: a London-born analytics firm proven in Formula 1, acquired by McKinsey in 2015 and grown from a 45-person start-up into the firm's global AI and engineering capability. Its Labs unit has produced more than 20 internal AI products and over 140 sector accelerators across life sciences, retail, mining and financial services.
- HQ: London, within McKinsey & Company
- Best for: board-mandated AI transformation at global enterprise scale, where the AI work has to connect to a wider strategy engagement
- Strengths: scale, cross-sector benchmarking, and strategy, data science and engineering under one contract. This is a route for large enterprises; mid-market buyers will usually be better served further down this list.
Faculty — now part of Accenture
Faculty was founded in London in 2014 as a fellowship programme for PhD graduates and became the UK's most prominent applied-AI consultancy, known for public-sector work including pandemic-era NHS modelling to forecast hospital admissions and direct equipment such as ventilators. Accenture completed its acquisition of Faculty on 16 March 2026: more than 400 data scientists and engineers joined Accenture, Faculty's Frontier decision-intelligence product joined Accenture's portfolio, and co-founder Dr Marc Warner became Accenture's chief technology officer.
- HQ: London
- Best for: applied AI in government, defence, national security and health — now with Accenture's global delivery scale behind it
- Strengths: genuine technical pedigree and shipped public-sector systems. Buyers who valued Faculty as a boutique should ask how the engagement model and team continuity change inside Accenture.
IBM Consulting UK
IBM Consulting is the consulting arm of IBM, with a substantial UK practice and a distinctive position: the advisory work is paired with IBM's own watsonx product family. watsonx.governance, launched in November 2023, automates model documentation, monitoring and risk evidence across estates spanning IBM, AWS, Azure, OpenAI and other platforms, and IBM publicly cites AI strategy and governance work with clients including Nationwide Building Society.
- HQ: Armonk, New York, with its UK base in London
- Best for: large enterprises that want AI governance tooling and consulting from a single vendor across a hybrid estate
- Strengths: depth of tooling and long regulated-enterprise experience. The consultancy and the software come from the same shop — efficient for integration, worth pressure-testing for independence of advice.
BJSS — now part of CGI
BJSS is a Leeds-headquartered technology and engineering consultancy of more than 2,400 consultants with a long delivery record in health, government, retail, financial services and energy. CGI completed its acquisition in early 2025, taking CGI's UK headcount to around 9,000 across 26 locations, with a stated focus on consulting-led cloud, AI and data work.
- HQ: Leeds
- Best for: large-scale software and platform delivery where AI is one component of a wider engineering programme, particularly in the public sector and NHS
- Strengths: delivery muscle, UK regional presence, and engineering credibility that predates the AI cycle. AI here is a capability within a broad consultancy, not the whole firm.
Elixirr
Elixirr calls itself "the challenger consultancy": founded in London in 2009, listed on AIM since 2020, and built around a senior-led, non-pyramid model. Its FY2025 results reported revenue up 34%, with AI-related revenue up more than 260% and over 45 in-house AI tools in use across its workflows. Its data and AI engineering capability came partly through acquisition, including the US-based firm iOLAP.
- HQ: London
- Best for: senior-led AI and digital strategy for buyers who want partner attention throughout, without big-four engagement structures
- Strengths: senior staffing on the actual work, and listed-company transparency about its numbers. Engineering depth is newer than its strategy heritage; ask which entity delivers the build.
Mesh-AI — now Indicium AI
Mesh-AI launched in London in 2021 with $30m of backing from Columbia Capital, built specifically around data and AI for highly regulated industries — financial services and energy and utilities above all. In November 2025 it was acquired by Indicium, and the combined firm has operated under the Indicium AI brand since early 2026.
- HQ: London; the combined Indicium AI operates globally
- Best for: enterprise data foundations in regulated industries — the platform, engineering and operating-model work that has to exist before AI delivers anything durable
- Strengths: regulated-sector data engineering and a clear-eyed view that most AI failure is data failure. Buyers should confirm how UK delivery and account teams are structured after the merger.
Advancing Analytics
Advancing Analytics was founded in 2018 and has become one of the UK's most credentialed data-platform specialists: a Databricks Elite partner and Microsoft Data & AI solutions partner with more than 80 specialists across the UK, Europe and the US, offices in London and Manchester, and its own delivery accelerator, Hydr8.
- HQ: London, with a Manchester office
- Best for: Databricks and Azure lakehouse builds, data engineering and ML engineering — the technical layer underneath AI ambitions
- Strengths: an unusually deep certified bench, including Databricks Champions and Microsoft MVPs. A focused technical firm: it builds the platform, and you bring the governance and strategy.
Cambridge Consultants
Cambridge Consultants has done deep-tech product development since 1960 and now operates as part of Capgemini Invent, with around 800 people across Cambridge, Boston, Singapore and Tokyo and more than 140 laboratories. Its AI work lives where models meet hardware, including 2025 work on explainable AI for the UK Ministry of Defence's machine-speed command-and-control initiative.
- HQ: Cambridge
- Best for: AI inside physical products — medtech, telecoms, defence and industrial systems — where the model has to ship in a device, not a dashboard
- Strengths: sixty-five years of R&D engineering and genuine laboratory capability. Not the firm for enterprise transformation; the firm for when the AI has to be invented.
Holistic AI
Holistic AI was founded in London in 2020 by UCL researchers Adriano Koshiyama and Emre Kazim and now operates from London and Palo Alto. It is a governance platform first: software that inventories an organisation's AI systems, tests them for bias and security weaknesses, red-teams LLMs and generates compliance evidence. It appears in the UK government's published portfolio of AI assurance techniques and names enterprise clients including Unilever and Siemens.
- HQ: London and Palo Alto
- Best for: enterprises that need continuous governance software — monitoring and audit evidence across a large AI estate — rather than a consulting engagement
- Strengths: research pedigree and regulator-aware tooling. Buyers wanting hands-on board work will typically pair a platform like this with an advisory firm.
Governance AI
Disclosure: Governance AI is our own practice — we include it because the comparison is incomplete without the governance-specialist category; judge the criteria for yourself.
Governance AI is a Birmingham-based consultancy founded by Dr Karl George MBE, creator of the tgf Governance Code — endorsed by the late Sir Adrian Cadbury — and Partner and Head of Governance at RSM UK. The firm does both halves of the job: board-level advisory, from the AI Wake-Up Call literacy session to the GovernIQ™ diagnostic (from £3,950), and build — shipped systems including a public-sector evidence workspace, an insolvency intelligence engine and an AI operations system for a property manager, each with governance controls written into the code: read-only database access, append-only audit ledgers, confidence floors and human approval gates. The operating model behind that is set out in how an AI-native consultancy works.
- HQ: London and Birmingham, United Kingdom
- Best for: board-level AI governance in regulated and purpose-driven sectors — housing associations, charities, local authorities and professional services — where the work has to stand up to a board, an auditor or a regulator
- Strengths: one team designs the governance and writes the code, so nothing is lost in the handoff between advisor and developer. This is not the firm for thousand-engineer transformation programmes; that category sits at the top of this list, and we'd say so in the first call.
The field at a glance
| Firm | Best for | Sector coverage | Advise / build |
|---|---|---|---|
| QuantumBlack (McKinsey) | Global AI transformation tied to strategy | Cross-sector enterprise | Both, at scale |
| Faculty (Accenture) | Applied AI in government, defence, health | Public sector, defence, health, enterprise | Both |
| IBM Consulting UK | Governance tooling plus consulting, one vendor | Cross-sector enterprise, financial services | Both, anchored to watsonx |
| BJSS (CGI) | Engineering delivery at programme scale | Health, government, retail, financial services | Build-led |
| Elixirr | Senior-led AI strategy outside the big four | Financial services, consumer, private equity | Advise-led |
| Mesh-AI (Indicium AI) | Data foundations in regulated industries | Financial services, energy and utilities | Both |
| Advancing Analytics | Databricks / Azure platform and ML builds | Cross-sector, mid-market to enterprise | Build-led |
| Cambridge Consultants | AI in physical products and deep tech | Medtech, telecoms, defence, industrial | Build (R&D) |
| Holistic AI | Governance software across a large AI estate | Cross-sector enterprise | Platform, with advisory |
| Governance AI | Board-level governance in regulated sectors | Housing, charities, local authorities, professional services | Both, governance-first |
How to choose
Start from the job, not the brand. "We need AI" is not a brief; "we need our complaints-handling triage governed well enough to defend to the Housing Ombudsman" is. Once the job is named, most of this list eliminates itself.
Then apply four tests. First, check ownership: three firms on this list changed hands in the last eighteen months, and the partner who pitched may not own the relationship by delivery. Second, demand shipped evidence in your sector — a named system in production, not a methodology diagram. Third, ask what governance artefacts you will hold at the end: a risk register, control mappings, audit evidence your board can show a regulator. If the answer is "a report", you are buying advice, which is fine only if advice is what you scoped. Fourth, meet the people who will do the work, not the people selling it.
A structured diagnostic is the cheapest way to test a firm before committing to a programme — ours is the GovernIQ™ diagnostic, and most serious firms offer an equivalent first phase.
What should boards look for in an AI consultancy?
Three things, in order. Evidence of shipped work: systems in production, with the firm willing to explain the controls inside them. Named accountability: who, on their side and yours, owns each governance artefact — because the UK's regulatory approach makes your board, not your supplier, answerable for how AI is governed. And the willingness to say no: a consultancy that has never told a client something should not be automated is selling enthusiasm, not judgement. Fluency in ISO/IEC 42001 and the NIST AI RMF matters too, but as a means of producing evidence, not as logos on a slide.
If you want to know where your own organisation stands before speaking to any firm on this list, the free Board AI Scorecard takes a few minutes and gives you a defensible starting point.
Last reviewed: 12 June 2026.
We are one of the ten firms above, and we have tried to be as honest about ourselves as about everyone else. If your need sits in our category — board-level AI governance in a regulated organisation, advised and built by the same team — start with the free Board AI Scorecard, read how we work, or book a short conversation about where your AI exposure sits.
