AI rollups in the UK: what they mean for accountancy and lettings
An AI rollup buys small services firms and uses AI to take work out of each one, so profit rises after the acquisition instead of depending on the multiple paid at exit. The model started in US accounting, IT services and property management. In 2026 it reached UK lettings, and UK accountancy consolidators look like the next test.
If you run or back a UK buy-and-build platform, the question has shifted. You used to be asked how many firms you could buy. Now the next buyer will ask how much margin you added to the ones you own, and "we integrated them" won't be enough of an answer.
This guide covers what an AI rollup is, who is running them, how the UK accountancy and lettings markets compare, where the savings come from inside a practice, and what a consolidator should do before its next deal. Every figure is dated and linked to its source.
Key TakeawaysGeneral Catalyst has allocated $1.5 billion to AI-enabled roll-ups, targeting services businesses where AI agents can automate 30 to 70 percent of workflows (Capital & Clarity, April 2026).Thrive Holdings is committing $1 billion to buying US accounting practices through Current, which reported a 31 percent time saving on tax returns after running 7,000 of them through its AI (Footnote, July 2026).Dwelly brought the model to UK lettings: £69m raised in February 2026, £32m of it equity led by General Catalyst, with eight letting agencies bought by then.UK accountancy consolidators grew mostly by acquisition. Sumer grew 163 percent and Cooper Parry 61 percent in filed accounts to August 2026, and Xeinadin, built from 122 firms, was reported to be seeking around £1 billion at 15 to 17 times EBITDA and struggling to find a buyer.BCG found in January 2026 that most private equity firms cannot show meaningful returns from AI in many portfolio companies. The platforms that do will be the ones that change the work inside each practice, not the ones that buy licences.
What is an AI rollup?
An AI rollup is a buy-and-build strategy in which the acquirer uses AI to cut the labour in each firm it buys, aiming to raise margins across the group. A traditional roll-up makes most of its return from multiple arbitrage: buy small firms cheaply, combine them, and sell the larger group at a higher multiple.
The difference shows up after completion. A traditional platform integrates back office, brand and banking, then keeps buying. An AI rollup also rebuilds how the work gets done in each practice, so the tenth acquisition starts from tools already tested in the first nine.
| Traditional buy-and-build | AI rollup | |
|---|---|---|
| Main source of return | Buying at a low multiple and selling the group at a higher one | Margin added inside each firm after acquisition, plus the multiple |
| What changes after a deal | Brand, finance, IT, some shared services | The same, plus how the admin and production work is done |
| Stated margin ambition | Scale savings on overheads | General Catalyst's thesis: from 5 to 10 percent EBITDA towards 30 to 40 percent |
| Main risk | Paying too much and integrating slowly | Promising AI savings the practices never adopt |
Who is running AI rollups in 2026?
Three names set the pace, and one of them is already buying in the UK.
General Catalyst has allocated $1.5 billion to what it calls AI-enabled roll-ups, according to Capital & Clarity (7 April 2026). The companies it has created or backed include Long Lake in homeowner association management (raised $670m and bought 18 businesses), Titan in IT managed services ($74m raised alongside its acquisition of RFA), Eudia in legal ($105m raised), Accrual in tax preparation (launched with $75m) and Dwelly in lettings.
Thrive Holdings, spun out of Thrive Capital, is committing $1 billion to buying US accounting practices through Current, formerly Crete Professionals Alliance. Footnote reported on 14 July 2026 that Current had pulled together close to 50 practices, ran 7,000 tax returns through its AI last season and reported a 31 percent time saving on returns. OpenAI took a stake and lent a team to train AI agents on tax work.
Dwelly is the clearest UK example. On 26 February 2026 it announced £69m of funding, a £32m equity round led by General Catalyst and a £37m debt facility from Trinity Capital, having already bought eight letting agencies (Letting Agent Today). It integrates its AI operating system into the agencies it buys, and each new agency gives it more data to train its automations. In July 2026 it bought the lettings business of AP Morgan.
| Operator | Sector | Capital (as reported) | Scale reported | Source and date |
|---|---|---|---|---|
| General Catalyst | Several services sectors | $1.5bn allocated to AI roll-ups | Portfolio incl. Long Lake, Titan, Eudia, Accrual, Dwelly | Capital & Clarity, April 2026 |
| Thrive Holdings / Current | US accounting | $1bn committed | Close to 50 practices; 7,000 returns through AI | Footnote, July 2026 |
| Dwelly | UK lettings | £69m (equity and debt) | 8 agencies by February 2026 | Letting Agent Today, February 2026 |
How are UK accountancy consolidators different?
UK accountancy consolidation was built on the traditional model, and the exit market is starting to test it. Filed accounts to August 2026 show Sumer up 163 percent, BK Plus up 85 percent, Moore Kingston Smith up 69 percent and Cooper Parry up 61 percent, with most of that growth from acquisition rather than organic wins (The Accounts).
The multiples explain the model. Mid-sized traditional practices change hands at around 4 to 6 times EBITDA, according to The Finance Story (5 February 2026). The same piece reported that Xeinadin, built from 122 UK and Irish firms and backed by Exponent, has more than £100m of revenue and was seeking around £1 billion, 15 to 17 times its estimated £60m EBITDA, with Evercore running the sale, and that buyers were balking at the price. Sumer, backed by Penta Capital, had completed 34 deals and reached 2,400 staff.
The US shows where this goes. Between 2015 and 2025, 177 direct private equity investments in US accounting firms set off another 875 roll-up acquisitions underneath them, and by early 2026 roughly half of the 30 largest US firms carried private equity money or an alternative practice structure (Footnote). The US platforms now have an AI-first competitor buying the same practices.
For a UK consolidator the arithmetic is simple. Buying at 5 times and selling at 15 times only works if the next owner believes the margin will hold. Practice-level AI is one of the few levers that raises margin without another acquisition.
What does lettings consolidation look like?
Lettings is more fragmented than accountancy and is consolidating in a different shape. About 20,000 firms operate in UK lettings, and the top 100 hold under 30 percent of the market (Letting Agent Today, February 2026).
Research by The Business Acquisitions Group covering deals from 2024 to August 2026 found that buyers now prefer lettings books to whole agencies, and that private-equity-backed platforms have become more prominent. Active buyers include Lomond, Campions (the UK arm of Emeria) and Dwelly (Property Industry Eye, 9 September 2026).
A lettings book suits an AI rollup well. The recurring work is predictable: tenant enquiries, maintenance requests, rent chasing, gas and electrical certificates, and the tenancy changes the Renters' Rights Act brought in. A platform that runs that work on one system can absorb a new book with far less extra headcount.
Where does the AI saving come from inside a practice?
The saving comes from the dull, repeated admin inside each practice, not from a board-level AI strategy. BCG's January 2026 review found that most private equity firms cannot show meaningful returns from AI in many portfolio companies, because distributing licences does not change how the work is done (BCG).
In our own survey of 167 UK small business owners in March 2026, respondents lost an average of 8 hours a week to repetitive admin, and 47 percent had never automated a single task. Those are the firms platforms buy. In accountancy and lettings, the hours sit in a short list of jobs:
| Accountancy practice | Letting agency |
|---|---|
| Chasing clients for year-end, VAT and payroll documents | Answering tenant enquiries and triaging maintenance |
| New-client onboarding and AML identity checks | Rent chasing that escalates in the right order |
| Quarterly Making Tax Digital updates for sole traders and landlords | Tracking gas, electrical and licensing certificates |
| Monthly management reports from Xero or Sage | Producing landlord statements and updates |
Treat the headline numbers as ambitions until your own practices prove them. General Catalyst's 30 to 70 percent is a thesis about workflows, and Current's 31 percent is a reported result on one task, tax returns. Your number depends on how standard your practices' work is and how many of your staff use the new systems.
Take a hypothetical platform that owns six accountancy practices. If it measures admin hours in each before it builds anything, it can show the board that document chasing fell by a set number of hours per practice. If it skips the baseline, it can only say the tools are live. The second answer does not move a valuation.
What should a UK consolidator do before the next acquisition?
Build the AI integration into the acquisition process, not the other way round. Five steps cover most of it:
- Measure admin hours in each practice at completion, before anything changes, so every later saving has a baseline.
- Pick one standard for practice management and client communication across the group, so each build can be reused.
- Build once and deploy per acquisition. The tenth practice should start from automations already tested in the first nine.
- Keep a person approving anything that goes to a client, especially tax figures, deposit deductions and AML decisions. Each practice stays responsible for its data under UK GDPR.
- Report hours saved and the resulting margin to the board every quarter, practice by practice.
This is the work we do inside individual practices. For platforms, we run a pre-deal AI review of a target and a 100-day sequence in each acquired practice, using our AI Core Audit and AI Core Build. The details are on our page on AI for private equity buy-and-build platforms.
What are the risks of an AI rollup?
The main risk is the gap between the model on paper and the practices in real life. Four problems come up most:
- Adoption. Staff in a newly acquired practice keep their old habits. An automation nobody uses saves nothing.
- Regulated output. A wrong tax figure, deposit deduction or identity check has consequences an email reply does not. Those workflows need a human check before anything goes out.
- Client data across merged firms. Moving client records between systems raises UK GDPR questions each practice has to answer.
- Overclaiming. A margin target in a fund presentation is not a result. Buyers will ask for the baseline.
None of these argue against the model. They argue for doing it at practice level, with measurement from day one.
Frequently asked questions
What is an AI rollup?
An AI rollup is a buy-and-build strategy where the acquirer uses AI to cut the labour in each firm it buys, aiming to raise group margins after acquisition. It differs from a traditional roll-up, which relies mainly on buying firms at a low multiple and selling the combined group at a higher one.
Who started AI rollups?
General Catalyst is the most prominent backer, with $1.5 billion allocated to AI-enabled roll-ups across sectors including homeowner association management, IT services, legal, tax and lettings. Thrive Holdings runs a parallel strategy in US accounting through Current, with OpenAI as a shareholder.
Are there AI rollups in the UK?
Yes. Dwelly, backed by General Catalyst, raised £69m in February 2026 and had bought eight UK letting agencies by then, integrating its AI operating system into each one. UK accountancy consolidators such as Sumer, Cooper Parry and Xeinadin grew mainly through traditional acquisition.
Do AI rollups work?
It is too early to say for most. Current reported a 31 percent time saving on tax returns, while BCG found in January 2026 that most private equity firms cannot yet show meaningful AI returns in many portfolio companies. The difference is whether the work inside each firm changes.
Where this leaves UK consolidators
AI rollups have moved from a Silicon Valley thesis to UK lettings in a year. The US accounting market already has an AI-first buyer with OpenAI on its shareholder register. UK accountancy platforms built on traditional roll-up maths now have to show the next buyer margin as well as scale.
The lever sits inside each practice: document chasing, onboarding, tenant communications and reporting, measured from the day of acquisition. If you run or back a platform buying accountancy practices or lettings books, see how we work with buy-and-build platforms, or start with an AI Core Audit in one practice.
Richard Thomas-Pryce is the founder of HeyBRB, a London-based AI consultancy that builds automations for accountancy practices, letting agencies and other firms of 1 to 20 staff. Last updated 1 October 2026.