RESEARCH · AI ADOPTION IN PROFESSIONAL SERVICES
The quiet adopters: how the UK's small professional service firms are really adopting AI
The Change Innovation 2025 insight report is based on in-depth research with nine owners and partners of solo and small professional service firms across law, finance, consulting, research and psychology.
​
Every participant reported real benefits from AI. However, none of them used it in client-facing work without human review first.
​
This page shares what they told us - and what it means for your firm.
Change Innovation Insights
The Headline:
adoption is pragmatic, incremental and cautious
“There's absolutely no way I could have built two businesses at the same time without GenAI… we'd have to employ someone — there's just not the funds for that.”
— Owner, small consultancy
Small firm leaders are not waiting for permission (but they are not gambling their reputations either).
​
They are using AI where the value is immediate and the risk is low, and holding back where their credibility and professional reputation is on the line.
​
All nine firms in the research saw real benefits. Cost was a barrier for no one. And nothing reached a client without human review.
Finding 1:
The benefits are real - and bigger than efficiency
Why This Matters
Thinking partner and capacity are two forms of value that established adoption theory doesn't capture. But they matter hugely to firms with deep expertise and limited resources.
For a solo or small firm, the case for AI is not just about doing the same work faster; it is about doing work that would otherwise not get done at all.
Every participant saw real value in AI, and all nine cited efficiency.
But the more interesting result is what sits behind it: for the smallest firms, AI is providing capacity they simply could not afford to buy any other way.
As one owner of a small research firm put it: “It's like having a team.”
Participants also valued AI as a creative aid and as a thinking partner - a sounding board that challenges assumptions and surfaces blind spots - which matters when you work alone.
Finding 2:
The invisible line - nothing reaches a client without human review
“Nothing more embarrassing than being caught out… it has to pass the red face test.”
— Solo Consultant
These professionals were reporting clear benefits and found the tools affordable, but they keep AI away from client-facing work unless a human has reviewed and reworked the output first.
Professional service firms sell judgement and trust, and being caught out using AI badly would damage the thing the business runs on: reputation.
​
Participants described discomfort when AI output didn't sound like them, and human review was essential.
Disclosure was handled carefully, because client attitudes were split: some clients dislike AI use, others expect it.
Finding 3:
Regulation decides who can move - and how fast
“We can't use AI for many functions because we're a regulated firm — there is no regulator policy on AI, so we couldn't have AI doing any prescribed role.”
— Partner, small finance firm
The gap between the most and least adventurous firms wasn't attitude, age or budget — it was sector.
Consulting and creative firms experimented more freely; law, finance and psychology moved carefully because the rules around them are unclear.
In regulated sectors, ambiguity is the barrier: with no regulator policy on what safe AI use looks like, a misstep carries serious consequences, so cautious firms stay cautious.
Finding 4:
Peers teach what training courses don't
“I started by watching others use it.”
— Solo Consultant
Almost no one mentioned formal training. Again and again, the route into confident AI use was informal: watching a peer use it, a crash course from someone in your network, one-to-one support from a person you trust.
Early attempts were often “intimidating and tricky” and what turned it around, consistently, was a person, not a course.
Larger firms have IT departments and structured programmes; small firms have networks. In this research, networks worked. And habit compounded the gains: confidence followed daily use.
Finding 5:
The fear is real - but it's projected onto others
“I worry about it for more junior people — they will miss those opportunities to get good.”
— Owner, small consultancy
Ask experienced professionals whether AI threatens them, and most say their judgement, experience and specialism protect them.
Ask who they worry about, and the answer is immediate: the juniors. The early-career work that AI can do (such as research, first drafts, routine analysis) is how professionals have always learned their craft.
If AI absorbs it, where does the next generation's judgement come from?
Meanwhile, the emotional arc for adopters themselves ran from early scepticism to confidence and even delight. Participants described AI as “fun”, “exciting”, and feeling “re-energised” in their work.
How adoption happens:
the model
Why This Matters
Two things distinguish small professional firms from the organisations most adoption theory was built on: reputation acts as a brake, and peers act as an accelerator.
The full one-page diagram is in the report.
Drawing the results together, the research proposes a model of AI adoption specific to solo and small professional service firms. Perceived benefits including efficiency, capacity, creativity and thinking partnership create the intention to adopt, a belief shared by all nine participants.
​
But four moderating forces determine what actually happens: regulation and professional standards; professional credibility and reputation; culture and practical conditions, especially peer influence; and psychology.
The result is cautious, selective use in regulated sectors, broader experimentation in consulting and creative firms, and internal tasks first(client-facing only with human review).
Questions the research answers
Are small UK professional firms actually using AI?
Yes: quietly and pragmatically. In Change Innovation's 2025 research with nine solo and small professional service firms, every participant was using generative AI and seeing real benefits, particularly efficiency and capacity. But use stayed concentrated on internal tasks (eg drafting, planning, research and admin) and nothing reached a client without human review.
Is AI worth it for a small consultancy or solo business?
For the firms in this research, clearly yes (and cost was not a barrier for them). The biggest value wasn't just saving time; it was capacity. One consultancy owner said they could not have built two businesses at the same time without it, because hiring wasn't affordable. For a solo or small firm, the case for AI is not just about doing the same work faster; it is about doing work that would otherwise not get done at all.
Do small law firms and finance firms use tools like ChatGPT?
Cautiously, and far less than consultancies. The research found regulated professions (law, finance, psychology) held back not by attitude or budget but by ambiguity: with no clear regulatory expectations or professional body policy on what safe AI use looks like, a misstep carries serious consequences, so firms restrict AI to low-risk internal tasks. Clearer guidance would change adoption in these sectors more than any new tool.
Should I tell clients I use AI?
There's no settled norm yet. Participants in the research reported split client attitudes (some clients dislike AI use, others expect it) and most disclosed only when asked or when tenders required it. The practical advice from the research: decide your disclosure position deliberately, before a client asks, rather than improvising in the moment.
Will AI replace consultants, lawyers and other professionals?
The experienced professionals in this research didn't think so. They saw their judgement, experience and specialism as protection. Their genuine concern was for junior professionals, because AI now does the early-career work through which judgement has traditionally been built.
What's the best way for a small firm to start with AI?
Based on the research: start with internally focused work and low risk use cases (drafting, planning, background research), decide what you will and won't use AI for before habit decides for you, make human review the condition of anything client-facing, and seek out peer-to-peer style learning rather than a generic course. The full five moves are in the report.
Methodology
Nine semi-structured interviews (approx. 60 minutes each) with owners and partners of established UK solo and small professional service firms across law, finance, consulting, research and psychology, conducted mid-2025 as part of MSc research at the University of Sussex, with university ethical approval. Interviews were analysed thematically. All participants are anonymised; quotes are lightly edited for readability and attributed by broad sector and firm type only. The results are qualitative and reflect a small sample at a specific point in time; they describe patterns, not statistical generalisations.