How a 4-person accountancy practice took back 18 hours a week from the year-end chase.
The firm recovered 18 fee-earner hours a week, got 88% of clients filing on time, and returned £15,500 of advisory capacity to billable work — by automating the chase, not the judgement.
At a glance
| Before | After (one full year-end cycle) | |
|---|---|---|
| Fee-earner hours/week chasing documents (firm-wide, crunch) | 24 hrs | 6 hrs |
| Clients submitting documents on time* | 62% | 88% |
| Billable advisory capacity lost to chasing | £25,900/yr | down ~75% |
| Billable advisory revenue recovered for the firm | — | £15,500/yr |
| Clients slipping past the HMRC filing deadline | 14 | 3 |
*“On time” measures clients meeting the firm's internal cutoff. The HMRC deadline figure on the bottom row is a separate, harder measure.
Build time: 3 weeks · Payback: inside the first year-end cycle
The starting point
Arthur's firm is a 4-person accountancy practice handling year-end accounts and self-assessment for around 180 clients. Every year the same thing happened: the run-up to filing turned into a manual chase.
Fee earners — people on roughly £120/hr of billable advisory time — were spending their evenings tracking who still owed which bank statements, invoices, receipts and payroll records, then sending one-off chasing emails and making calls. There was no single view of who was outstanding. The same clients got chased repeatedly; a handful slipped through entirely and were dealt with in a deadline panic.
The cost wasn't just hours. It was the type of hours: senior, billable, advisory time spent on admin that nobody enjoyed and nothing got smarter from doing. Across the four fee earners, the firm was losing about 24 hours a week to the chase during the 12-week crunch — close to £26,000 of billable capacity a year.
Why this workflow, not another
We don't automate the loudest problem — we automate the highest-leverage one. The audit looked at four candidate workflows across the firm:
- Engagement letter automation — strong, but harder to prove out without exposing letter content.
- Bookkeeping query triage— valuable, but the outcome (“queries answered”) is fuzzy to measure.
- Onboarding / AML / KYC — high value, but regulatory complexity would have slowed the build.
- Year-end document collection — universal pain, dead easy to measure (hours saved per fee earner, % of clients on time), low regulatory risk, and buildable in three weeks.
Document chase won on every axis that matters for a first build: measurable, fast, low-risk, and felt by everyone in the firm.
What we built
A single system that owns the chase end to end — with humans kept exactly where their judgement is worth paying for.
- A live document checklist per client — exactly what each client owes for their year-end, in one place.
- A scheduled, personalised reminder sequence that escalates sensibly as the deadline approaches — and stops the moment a document arrives, so nobody gets chased for something they've already sent.
- Automatic logging of what's received, so the picture is always current without anyone updating a spreadsheet.
- A real-time dashboardshowing who's complete, who's outstanding, and days to deadline — the partners' single source of truth during crunch.
The human checkpoints — by design:
- A fee earner reviews any flagged complex or sensitive case before it gets escalated.
- Partner sign-off stays on anything non-standard.
- The “this client needs a phone call, not another email” judgement is surfaced to a person, never made by the system.
What we deliberately did not automate
This is the part that matters most.
We told Arthur's firm not to automate the partner-led conversations on complex or contentious returns, or the handful of long-standing clients who expect a personal call rather than a reminder. Where the relationship and the judgement arethe service, an automated nudge doesn't save time — it quietly erodes trust the firm has spent years building.
A chase reminder is a task. A conversation about a complex return is judgement. We automated the first and protected the second — and that line is exactly where the firm's value lives.
The results
After one full year-end cycle on the new system:
- Firm-wide chasing time fell from 24 to 6 hours a week during crunch — about 4.5 hours a week back per fee earner, redirected into billable advisory work.
- On-time client submissions rose from 62% to 88%.
- Average time from document request to receipt dropped from 21 days to 9.
- An estimated £15,500 of billable fee-earner time was recovered across the cycle (roughly 60% of the freed hours redeployed to chargeable work).
- Clients slipping past the filing deadline fell from 14 to 3 — out of 180.
“The chase used to own January. Now I open one dashboard and I can see exactly where every client is — and my team's doing actual accounting again instead of sending the same email for the fourth time. The part I didn't expect was that it told us whichclients still needed a real conversation, rather than just nagging everyone.”
— Arthur T., Managing Partner (firm name withheld at the client's request)What happened after the build
The system has been running on a halcroft management retainer since launch — halcroft handles hosting, monitoring, and the small adjustments needed each cycle, so nothing on the firm's side breaks when HMRC or the document providers change something.
Arthur's firm is now scoping its second build — bookkeeping query triage — from the same audit's 90-day roadmap. The audit was never a one-off recommendation. It was a sequence: chase first, queries next, advisory packaging after that, each one paid back by the last.
The shape of the engagement
Audit identified document chase as the highest-impact fit → we built the system in 3 weeks → here are the numbers after one year-end cycle → and here's the work we explicitly told them to keep human.
Want this for your practice?
Start with a free 45-min briefing. We'll tell you whether a paid audit makes sense — and what we'd expect to find if it does.