Practice Management
in the Age of AI 2026 Edition
An honest, first-party read on how real accounting firms are using AI — what they'd trust it to do, and where they see the work heading.
firms want an AI agent to chase clients for missing documents — the #1 task they'd hand to AI.
What's inside
A first read, in eleven parts.
- 01Who responded — and how to read this03
- 02Foreword by Rain Allikvee04
- 03Key findings05
- 04How firms run today06
- 05AI in firms today07
- 06How firms feel about AI08
- 07What unlocks trust09
- 08The agentic future10
- 09The firms' own vision11
- 10Where software still falls short13
- 11What it means14
- —About Uku15
Who responded
Built from the people who decide.
This report draws on a survey of accounting firms in the Uku community, across 8 countries, fielded May–June 2026. It skews — deliberately — toward the people who decide: roughly 6 in 10 respondents are partners, owners, or managers. It also skews small-to-mid: most firms have 1–10 people. That's our base, and it's exactly who this report is about.
Firm size
- 1–3 people41%
- 4–1029%
- 21–5015%
- 11–2012%
- 51–2003%
Region
- Baltics50%
- Nordics24%
- North America12%
- UK & Ireland9%
- Rest of Europe6%
Role
- Partner / owner50%
- Individual contributor29%
- Manager12%
- Ops / tech / admin9%
Foreword
Publish it straight,
numbers and all.
Rain Allikvee
Co-founder, Uku
Every "best software" article in our category reads the same. Vendor copy, recycled claims, a feature list as long as your arm. We wanted the opposite: to ask real accounting firms what they actually think about AI — and then publish it straight.
So we asked the firms we know best: our own. What you'll read here isn't a forecast or a press release. It's what dozens of firms across eight countries told us about how they use AI today, what they'd trust it to do tomorrow, and where they think their work is heading.
First, firms want AI to chase, not to decide. The work they'd happily hand over is the chasing, the drafting, the flagging. The work they keep is judgment.
Second, trust is gated on control. Not a single firm told us it already fully trusts AI. What they want is simple: AI that proposes, and a human who approves.
Third, the bottleneck isn't features — it's time. The biggest daily pain is still missing deadlines. The firms that win the next few years won't be the ones with the longest feature list. They'll be the ones who finally have one clear picture of the work.
Key findings
Six things firms told us.
Firms want AI to chase, not decide
Nearly 7 in 10 say the #1 task they'd hand to an autonomous agent is chasing clients for missing documents — far ahead of anything touching judgment.
"Describe it and AI builds it" lands
More than 8 in 10 find it appealing to describe what they want in plain language and have AI build it — no developers.
The bottleneck is time, not features
The single biggest operational headache is missing deadlines and work slipping through — over billing, profitability, or visibility.
Trust is gated on control
What unlocks trust: a human approving first, security, privacy. Not one firm said it already fully trusts AI.
Optimistic, but clear-eyed
About 2 in 3 are excited or intrigued — but "intrigued" outweighs "excited." More measured than the headline market surveys.
Appetite for openness
About half say an open API and AI connectivity (for example, MCP) is critical or important — ahead of what most tools offer today.
How firms run today
They're not chasing intelligence.
They're chasing time.
Before any conversation about AI, it's worth remembering what firms are actually replacing. Half came from spreadsheets. A third switched from another practice-management tool. A quarter ran on email and shared folders. The "competition" for most firms isn't a rival product — it's the status quo of glue and memory.
Asked for their single biggest operational headache, firms didn't point at billing or analytics. They pointed at missing deadlines and work slipping through — the most-cited pain by a clear margin. Admin overload came next.
Our accountants would focus on adding value to clients, and all the recurring back-office entries would be handled by AI.
Partner — mid-size firm, Southern Europe
Biggest operational headache
Single most-cited pain
- Missing deadlines / work slipping32%
- Too much admin vs client work18%
- Covering absences / onboarding15%
- Not knowing if clients are profitable15%
- Slow or inaccurate billing15%
- No visibility into who's doing what6%
AI in firms today
Already in the building —
but lightly.
ChatGPT leads (roughly 6 in 10 firms), followed by Microsoft Copilot and the AI built into tools firms already use. What firms reach for it most is telling: client communication and writing top the list, then research and problem-solving.
What's notable is the restraint. Adoption of the big consumer tools is real but not yet deep, and AI for meeting notes is uncommon. The pull is strongest exactly where the daily grind is heaviest: communication, research, and documentation.
That contrast is the story: small European firms aren't behind on the idea of AI. They're just more deliberate about where it earns its place.
AI is the ability to conduct faster and deeper research so we can give better, faster advice to our clients.
Partner — small firm, Northern Europe
What firms use AI for most
Top uses (select up to 3)
- Writing & client communication76%
- Research & problem-solving59%
- Meeting notes & summaries29%
- Drafting SOPs & documentation29%
- Data entry & reconciliation24%
How firms feel about AI
Optimism with its eyes open.
About two-thirds of firms are excited or intrigued about AI — but the word that wins isn't "excited." It's "intrigued." A quarter are cautious, and a small group remains skeptical.
That restraint is healthy. Our firms are warmer than they are cold, but measured — and, crucially, they can tell you why they hesitate. Those reasons (control, security, privacy) are the subject of the next page.
For a vendor, measured optimism is the better signal. It means the appetite is durable, not hype — and the firms saying "intrigued" are still open to being shown.
I'd be in the garden, not working as much.
Partner — small firm, United Kingdom · asked what a year from now looks like
Feeling about AI right now
Optimistic, but measured
- Intrigued38%
- Excited26%
- Cautious24%
- Skeptical or concerned12%
What unlocks trust
Capability is table stakes.
Control is the product.
Ask firms what would make them trust AI more, and the answer is strikingly consistent. Top of the list: a human approving before anything is sent or filed. Then strong data security. Then their data staying private and out of training sets.
The most revealing number is the one that's zero: not a single firm chose "nothing — I already trust it." Even the enthusiasts want a seatbelt.
The job isn't to build the most autonomous system. It's to build the most accountable one — visible, reversible, human-in-the-loop.
0%
of firms said they already fully trust AI. Every firm wants a check in the loop.
What would build trust in AI
Share of firms selecting each
- A human approves before it's sent / filed62%
- Strong data security56%
- Data stays private (not used to train)53%
- Higher accuracy50%
- A clear audit trail of AI's actions44%
The agentic future
Hand over the chasing.
Keep the judgment.
This is where our data is genuinely new. We asked what firms would most want an AI agent — one that can act on its own — to handle. The runaway answer: chase clients for missing documents (nearly 7 in 10). Then prepare client emails, then flag unprofitable clients.
Two more net-new signals. More than 8 in 10 firms find it appealing to describe what they want in plain language and have AI build it. And about half say an open API and AI connectivity (for example, MCP) is critical or important — ahead of what most tools offer today.
What firms would hand to an AI agent
Share selecting each task
- Chase clients for missing documents68%
- Prepare client emails & updates59%
- Flag unprofitable clients or jobs44%
- Assign, schedule & rebalance tasks32%
- Draft & send invoices26%
- Onboard new clients21%
like "describe it, AI builds it" — no developers
say an open API & AI connectivity (e.g. MCP) is critical or important
would keep it all human, with no agent at all
In firms' own words
What "AI done right"
looks like to one firm.
Overnight, transactions across all accounting platforms are already categorized. Anything ambiguous lands in a single exception queue — not scattered across five systems. I approve or correct in 20 minutes. Instead of closing books, I'm in front of clients. Headcount stays lean, but each person handles several times the client load.
The honest gap between today and that picture isn't the tools — most already exist. It's a unified exception layer, a feedback loop so corrections actually improve the system, and a knowledge base so the firm's expertise isn't trapped in people's heads.
Sovereign Solutions — full-service firm, Denmark
I see the AI agent working at night. I give instructions in the evening and check them in the morning — it posts entries, checks for missing documents, and prepares client reminders.
Specialist — solo firm, Estonia
I open my inbox and the new messages are already worked through, with a draft reply ready. Emails are my single biggest time sink.
Partner — solo firm, Estonia
The firms' own vision
A year from now,
a typical day.
At the golf course :)
Partner — small firm, United States
A shorter working day — but at least the same profit and turnover.
Partner — small firm, United Kingdom
Every accountant has an effective dashboard over their own work and client portfolio.
Robby Bobby — firm, Estonia
Cloud tools show me raw data and make me filter it — it feels like the tax client who brings everything in a shoebox.
Partner — solo firm, United States
Quotes shared with permission. Firms are named only where they consented; all others are identified by size and region only.
Where software still falls short
Reporting the gaps is what
makes the optimism credible.
We asked firms what they wish their software did better today. Here they are, in firms' own words:
- Billing still eats too much time — especially for fixed-fee and recurring clients.
- E-signatures, KYC and AML still live in separate tools. Firms want them built in.
- Track time without starting a timer — and show client profitability more clearly.
- Onboarding a new client (or covering an absence) is still too manual.
One source of truth for the whole firm.
Most of these gaps share a root cause: the work lives in too many places. Uku gives accounting firms a single source of truth for who's doing what — so deadlines stop slipping and nothing falls through.
See how Uku worksWhat it means
Four signals for whoever
builds the next firm software.
Build for accountability, not autonomy
Firms want AI that proposes and a human who approves — visible, reversible, auditable. The winning products make the human-in-the-loop effortless, not optional.
The first agent should chase, not decide
Nearly 7 in 10 firms told us the same thing: the highest-value agent chases missing documents. Start where demand is loudest and judgment risk is lowest.
Make it describable
More than 8 in 10 want to describe what they need in plain language and have it built. The future of firm software is configurable by the firm, not by a developer queue.
Time is still the headline
Underneath every AI question sits the same truth: firms are losing time to missed deadlines and admin. Solve that first, and the AI conversation gets concrete.
This is a first read. We're still collecting — the expanded v2 lands in August 2026, with deeper cuts by firm size and region.
Practice management,
built for accounting firms.
Uku gives accounting, bookkeeping and audit firms one source of truth for who's doing what across the team — so work gets done on time, billed, and paid. Beautiful, simple, and built for the recurring, deadline-driven reality of firm work.
- Workflow & deadline automation for recurring work
- Time tracking & automated billing
- A client portal your clients actually like
- Real-time client-profitability monitoring
- Vacations & automatic cover for absences
- CRM, calendar, tasks & monitoring in one place
Founded 2017 by Rain Allikvee & Jaanus Lang, Co-founders.
© Uku 2026 · getuku.com/ai-report-2026