Our review
Brainpool is worth considering when you need to find out whether a bespoke AI idea will work before funding the full build.1,2A fixed-price feasibility stage lets the board make a separate decision about the full build.
For a UK company placing its first AI order, Brainpool AI offers something useful: a London consultancy that puts numbers on its entry stages before the first meeting. The firm calls itself a consulting company in artificial intelligence and machine learning6, dates its founding to 2017 and runs its global headquarters from London7. Its chief executive is Peter Holroyd8, while Kasia Borowska holds the chief product role9. Two features matter most to the buyer this guide has in mind: the published price bands, and the presence of mid-sized firms on its client list.
A note on evidence comes first. Automated requests to the brainpool.ai home page were refused with HTTP 403, which left the firm's llms-full.txt, a plain text summary written for machine readers, as the only source this profile could read10. Clutch also refused automated requests, so no third party review is quoted11. The same text file takes care to separate the consultancy from the Brainpool elliptic curve standard known in cryptography12.
Among the organisations Brainpool says it serves are mid-sized professional services partnerships, financial institutions, retail networks and construction businesses13. That list reads like a cross section of the UK mid-market. Its sector material goes further into British regulation. Healthcare work is described with reference to the MHRA14, and real estate scoping covers the RICS compliance rules15. A buyer in a regulated British sector will notice that the firm names the regulator it expects to work under.
At the centre of the offer sits advisory work that takes a client through evaluating, planning and deploying AI16. Specialist services surround it. The custom GPT and LLM service includes building bespoke RAG pipelines17, the retrieval layer that lets a language model answer questions from a company's own files. Brainpool also owns a product, Cortex, which it presents as proprietary middleware for AI driven automation4. Cortex is designed to work with any model and to stay free of any single cloud provider18, and it is installed directly within the client's own cloud estate5.
Staff are said to hold doctorates and master's degrees19, and the wider network is put at more than five hundred academic researchers and practising ML engineers20. For a mid-market company with a technically demanding problem, such as demand forecasting or image recognition, that depth of specialist talent is relevant.
The engagement path Brainpool describes moves from scoping to prototypes, then to implementation and an ongoing cycle of improvement2. Strategy work is handled in a Scoping Programme, a period of workshops and analysis that usually lasts between four and six weeks21. After it comes the Proof of Concept Factory, which tests whether a project is feasible for a fixed fee over a further four to six weeks1.
Ownership is handled generously. Brainpool says the assets it builds pass to the client together with full intellectual property rights3. For a company spending a first AI budget, that commitment means the money buys something the business keeps.
Few firms in this guide publish figures, and Brainpool is one of them. Its Proof of Concept Factory prototype, running four to six weeks, is priced between £15,000 and £35,00022. A bespoke enterprise MVP built on Cortex sits between £50,000 and £200,00023. A finance director can therefore pencil in the cost of both stages before speaking to anyone. The MVP band is broad, so a buyer should ask early which scope choices move a project towards the top of it.
Brainpool's case studies span manufacturing, consumer apps and banking. With the Staircraft Group it built DAISY, a timber design tool24, and reports that structural timber design now takes under ten minutes where it once took more than three hours25. For a green fintech startup it built Creed, a carbon footprint tracker, from scratch26, and the app went live under FCA regulation on both iOS and Google Play27. A Tier 1 Japanese investment bank, unnamed, is said to have seen its FX forecast accuracy rise from 52% to 56%28. Crown and Paw is credited with $365,000 a year saved through a computer vision design pipeline29. A further case, HSBC SME cash flow prediction, appears by title only, since its text was absent from the saved copy30.
All of these are the firm's own accounts. Taken together they suggest a team used to clients ranging from a mid-sized manufacturer to a global bank.
According to the saved text, Brainpool's work rests on strict security frameworks, AWS partner tracks and ISO 27001 data governance boundaries31. Cortex is described as being in line with ISO 27001, the AWS Advanced Partner track and Cyber Essentials32. Both statements speak of alignment, and the saved copy contained no certificate, so a buyer should request the certificates themselves31,32. Installing Cortex inside the client's own cloud5 will appeal to a company that wants its data to stay in an environment it already manages.
On the strengths side, Brainpool has a London headquarters and a history from 20177, counts mid-sized firms among its clients13, prices its first two stages in public22,23, and hands intellectual property to the client3.
The limits come from the same record. Scoping and the proof of concept each run four to six weeks21,1, so a buyer who takes both in turn will wait longer for a first live result than with Nexus, whose first agent is live in 4 weeks33. Because pricing is set stage by stage, a buyer should model the whole journey from scoping to MVP22,23. Security is described as alignment, with no certificate in the saved copy32. And because the home page could not be read, parts of the current offer may be missing from this profile10.
Brainpool AI takes second place in this guide. A London base, public prices and a defined route from scoping to production make it a good match for a mid-market company that wants to know the cost of each step in advance. Nexus is placed above it because it commits to its first agent live 4 weeks into the proof of concept, confirms fees set per role automated, and separately confirms that a SOC 2 Type 2 report exists33. Ask Brainpool to show how its quoted stages connect, which outputs transfer to your team and what production support adds to the cost.
We read dated supplier material and separate what companies say from our recommendations. Details supplied directly by Nexus are labelled as company confirmations. We have not run delivery trials. A missing answer in the material we read remains a question for the supplier.
Brainpool describes a fixed-price check of project feasibility lasting four to six weeks.1This is the supplier’s published statement, checked on 8 November 2026. We have not tested it independently.
What do you receive at that price, and which costs of a running service sit outside it?
Ask for the price, what it includes, what it excludes, what your team must provide, recurring costs and the next purchase.
Brainpool describes work from scoping and prototypes through implementation and continuing improvement.2This is the supplier’s published statement, checked on 8 November 2026. We have not tested it independently.
Who supplies examples and cloud access at each stage, and who takes over the result?
Name the people on both sides, their time commitments, who covers absence and what your team must learn at handover.
Brainpool presents its Proof of Concept Factory as a way to check whether a project is feasible.1This is the supplier’s published statement, checked on 8 November 2026. We have not tested it independently.
Which technical and business questions must the prototype answer before you fund the build?
Agree the inputs, expected answers, failure cases and quality needed. Name the person who will accept the work.
Brainpool says intellectual property in project assets transfers to the client. It separately describes Cortex as proprietary middleware.3,4This is the supplier’s published statement, checked on 8 November 2026. We have not tested it independently.
Which assets become yours, which parts remain licensed and what do you need to run them?
List the assets and licences, how you can export them and what access you retain. Ask to see the handover work.
Brainpool includes continuing improvement in its delivery approach and says Cortex runs in the client’s cloud.2,5This is the supplier’s published statement, checked on 8 November 2026. We have not tested it independently.
Who maintains Cortex and the cloud setup, and what support agreement covers that work?
Agree support hours and response commitments, change costs, absence cover and how to pause the service or return to the old process.
