DEV Community: Boris Korol

Drafting workflow automation for fee-earners: what gets prepared before review

Boris Korol — Tue, 21 Jul 2026 18:37:54 +0000

The blank page has excellent public relations. It suggests care, diligence and the quiet birth of a great idea.

In a professional-services firm, it more often means somebody is about to spend an expensive hour looking for dates they have already read twice.

That is not where legal, financial or advisory judgement lives. It is where it waits.

Law firms know the ritual. A client needs an update. A fee-earner opens the case-management system, then the inbox, then the attachments, then last month’s letter. The first draft is assembled from facts scattered across several places. Only after this archaeology can the person who actually knows what matters start thinking.

Accountants, recruiters, insurers, consultancies and outsourced operations teams have their own version of the same scene. Different documents; same tax on attention.

Drafting workflow automation is not a machine that “does the professional work”. It is a controlled process that reads approved case materials, extracts relevant facts and prepares a structured first draft of a letter, summary, document or client update. A qualified human still checks the facts, applies judgement and approves the final version.

The problem is rarely writing. It is reassembly.

Many firms already have templates, precedents, style guides and people who know the work. Yet the route from request to first draft is still clogged with small acts of retrieval: finding the current facts across correspondence, notes and attachments; checking names, dates and deadlines; choosing the right structure; turning the file into a chronology; spotting what is missing before a reviewer spots it later.

None of this is trivial. Nor is it usually the point of hiring an experienced fee-earner.

The cost is not only time. A poor first pass makes the reviewer reconstruct the file before they can test the advice, risk or tone. The business pays senior rates for a document clean-up exercise. The client gets the update later than they should.

What a drafting workflow actually does

A sensible workflow starts with limits. It reads only an agreed source set. It handles one known document type. It uses an approved template. It stops when the material is incomplete or unclear.

Within those limits, it can gather case materials, pull out basic facts, arrange a chronology, select the right structure and prepare a first draft. It can also flag an inconsistency or an unanswered question for human attention rather than inventing an answer.

The result is not a finished legal opinion or an unattended client communication. It is a better starting point: a draft, its source context and a short review checklist in the hands of the person who must decide whether it is right.

A good drafting workflow does not take judgement out of a firm. It takes the scavenger hunt out of judgement, leaving the reviewer to decide what the facts mean, what the client needs and what should happen next.

What stays with the professional

The workflow can collect approved materials, extract routine facts, create an outline, apply an approved template and assemble a draft. It can check whether the usual sections are present. It can highlight a gap.

The professional must decide which facts matter, whether the sources are complete, what advice is appropriate, how much risk the client should take and whether the document is ready to send. Final approval stays with a human. So do exceptions, sensitive matters and anything that does not fit the defined pattern.

That distinction is not a disclaimer tacked onto the end of the project. It is the design principle. If a workflow cannot tell where its knowledge ends, it is not ready to touch a client document.

The before-and-after is more mundane than the sales pitch—and more useful

Before automation, a request arrives and the drafting person begins the familiar relay race. They search the file, locate the latest correspondence, open a prior document, copy the structure, reconstruct the chronology and draft around the gaps. The reviewer receives a document that may be coherent, but must still check both the argument and the scaffolding.

After automation, the request starts a defined workflow. It pulls material from permitted sources, organises the core facts, applies the approved structure and produces a draft with unresolved points clearly marked. The reviewer checks the factual basis, makes the substantive decisions, reshapes the tone and signs off.

The point is not a miraculous document appearing from nowhere. The point is that the first useful draft arrives without making an experienced person spend an hour as a search engine.

Start where the documents repeat

The best first project is rarely the most impressive one. It is the document that appears often enough to be annoying, follows a stable enough pattern to be designed, and still needs a human at the end.

For a law firm, that could be a client update, attendance-note summary, case chronology, standard letter or internal matter handover. For an accountancy or advisory firm, it could be an engagement document, client report, proposal, follow-up letter or meeting summary. In recruitment or claims operations, it could be a candidate brief, case note or status update.

The controls are the product

Before building, a firm should decide which systems and documents the workflow may read, who may trigger it, what template it may use and what must be checked before approval. It should define stop conditions for conflicting information, missing facts, sensitive matters and unusual instructions. It should retain an appropriate record of review and sign-off.

The best rule is also the least glamorous: when the workflow cannot identify a fact confidently from an approved source, it should flag the question. Plausible filler is not a feature.

What a claimed 2× result must mean

“2× drafting workflow automation” is a decent headline. It is not, by itself, evidence.

If a firm wants to publish a twofold productivity claim, it needs to say what was measured: the document type, the average complexity, the sample size, the baseline time to a review-ready first draft and the time after the workflow was introduced. It should include human review and rework, not merely the seconds spent generating a draft.

Quality matters too. A fast first draft that creates more corrections, missed context or reviewer anxiety is not a productivity gain. Track exception rates, rework and reviewer feedback alongside time. Any public claim of a 2× result needs a client-approved measurement method and supporting evidence.

Who benefits most

Drafting workflow automation is strongest where a team repeats familiar document work but still begins by reassembling the same factual context. It is not a shortcut around expertise. It is a way to reserve expertise for the part that cannot be reduced to a template.

Law firms are an obvious case because files grow large and fragmented while fee-earner time stays expensive. But the pattern travels well: accountancy practices, advisory teams, insurance and claims operations, recruitment firms and other business-services teams all produce documents that need both a consistent first pass and a responsible final reviewer.

AI Automation Studio is a London-based automation consultancy led by Boris Korol. It designs practical, review-led workflows for UK business-services and professional-services firms, especially law firms, accountancy practices and operations teams.

If your team repeatedly turns scattered file material into the same kinds of client documents, start with one workflow. Map the source material, define the human review point and prove the result before expanding it.

Lead qualification and routing automation for law firms

Boris Korol — Mon, 20 Jul 2026 09:14:09 +0000

Most professional-services firms, especially law firms, do not have an enquiry volume problem first.

They have a sorting problem.

Enquiries come in, but the team still has to work out which ones are worth attention, which need a fast response, which belong somewhere else, and which should be declined early.

Lead qualification and routing automation for business-services and professional-services firms, especially law firms, is a workflow that screens new enquiries against clear fit rules, tags urgency and matter type, and sends each enquiry to the right person or queue with the right context attached. Its main value is faster triage, fewer dropped enquiries, lower admin burden, and more predictable handoffs.

The better use of automation is not pretending every lead is valuable.

It is helping the team identify which enquiries deserve time, where they should go, and what should happen next.

The operational bottleneck usually starts after the enquiry arrives

Many business-services and professional-services firms already have a way to capture enquiries.

The real strain often starts one step later.

Someone has to decide:

whether the matter fits the firm
whether it is urgent
which team should review it
whether more information is needed
whether the firm should respond now, later, or not at all

In smaller firms, that decision often lands with whoever first sees the message.

In larger firms, it may sit with intake staff, a shared inbox, a business-development team, or a fee-earner covering the queue between other tasks.

That creates predictable problems:

strong-fit enquiries wait behind weak-fit ones
urgent matters are not always visible early enough
different people qualify using different standards
senior staff spend time sorting instead of reviewing
nobody has a clean view of what is sitting where

That is an operations issue before it is a technology issue.

What lead qualification and routing automation actually does

Lead qualification and routing automation does not decide whether a firm should take a matter on its own.

It makes the triage step faster, more consistent, and easier to manage.

A practical workflow can:

capture enquiries from forms, inboxes, contact pages, and chat
tag the enquiry by service line, issue type, geography, or urgency
ask a short set of follow-up questions before human review
check basic fit rules such as practice area, location, budget, matter size, deadline, or conflict flags
separate strong-fit, possible-fit, and poor-fit enquiries into different queues
assign the enquiry to the right fee-earner, team, or coordinator
prepare a short summary so the reviewer does not start from raw messages
trigger acknowledgements, follow-up prompts, or decline messages based on rules

The point is not more complexity.

The point is to stop treating every incoming enquiry as if it needs the same path.

Where manual qualification and routing usually break down

Every enquiry enters the same queue

When all enquiries land in one inbox or one task list, the team has to sort good leads from poor ones manually.

That slows down the strongest opportunities and hides urgent matters in general traffic.

Qualification depends on who happens to look first

One person asks strong questions.

Another forwards the enquiry with almost no context.

Another leaves it sitting because they are not sure who should own it.

That inconsistency creates avoidable delay.

Good-fit matters get treated like generic admin

Some leads need immediate attention because timing matters.

If the workflow does not surface that early, the firm loses response quality even when the underlying legal or advisory team is strong. [needs source]

Senior time gets burned on first-pass sorting

Fee-earners and partners should not be spending their day deciding whether basic fit criteria have been met.

That is low-value routing work, not the highest-value use of their time.

What gets automated and what stays with a human

In a practical setup, automation usually handles the front-end routing work:

capture the enquiry
tag matter type and urgency
ask missing follow-up questions
check basic fit rules
separate strong-fit, incomplete, and poor-fit enquiries
route to the right queue or owner
prepare a short summary for review
trigger the right acknowledgement or follow-up message

The human still owns the parts that carry judgment and risk:

final matter acceptance
legal or professional advice
conflict-sensitive cases
relationship-sensitive communication
exception handling where the rules do not fit

That is the right split for most professional-services firms.

A before-and-after routing process

Before

Enquiry arrives by form, email, or referral.
It lands in a shared inbox or general queue.
Someone reads it when they have time.
They guess who should review it or ask for more detail.
The enquiry is forwarded internally, sometimes more than once.
A fee-earner receives partial context.
The matter is delayed, chased, or declined late.

After

Enquiry enters a structured qualification workflow.
The workflow tags matter type, urgency, and fit indicators.
Basic follow-up questions are collected early.
Poor-fit and incomplete enquiries are separated from strong-fit ones.
The right queue or reviewer receives the enquiry with a clean summary.
Exceptions are escalated to a named human owner.
A human decides whether and how to proceed.

The human decision remains the same.

The path to that decision becomes cleaner and faster.

Manual routing versus automated routing

Manual routing depends on individual judgment at the front of the queue, so quality varies by person and by workload.

Automated routing makes the first-pass logic visible, repeatable, and easier to improve over time.

In practice, the difference usually shows up in five places:

stronger leads reach the right reviewer faster
poor-fit enquiries consume less senior time
queue ownership becomes clearer
response standards become more consistent
throughput improves because fewer enquiries bounce internally

What a practical law-firm workflow can include

For a small or mid-sized business-services or professional-services firm, a sensible first qualification and routing workflow often includes:

a small set of service-line categories
a ruleset for urgency and fit
one owner for exceptions and unclear cases
separate queues for strong-fit, incomplete, and poor-fit enquiries
summary notes attached before fee-earner review
simple reporting on routed volume, response time, and decline reasons

That is enough to improve throughput without turning intake into a large systems project.

Implementation checklist

Define the top enquiry types and the top decline reasons first.
Write the minimum fit rules before building the workflow.
Decide what counts as urgent and who owns urgent escalation.
Separate incomplete enquiries from poor-fit enquiries.
Set routing rules by practice area, geography, or matter type.
Name a human owner for exceptions and ambiguous cases.
Track time-to-review, routed volume, and decline reasons as the first operating metrics. [needs source]

FAQ

What is lead qualification and routing automation in a law firm?

It is a workflow that screens enquiries against simple fit rules, collects the right missing facts, and sends each enquiry to the right queue or reviewer before human judgement is applied.

Does routing automation replace a fee-earner or intake manager?

No. It removes repetitive triage and sorting work. A human still decides whether the matter should be accepted and how it should proceed.

What should a qualification workflow check first?

Usually matter type, urgency, geography, service fit, deadlines, contact details, and whether the enquiry is complete enough for first review.

How does routing automation improve response times?

It reduces the time lost in shared inboxes, makes urgent matters more visible, and helps strong-fit enquiries reach the right reviewer faster.

What should stay manual in lead qualification?

Final acceptance, legal advice, sensitive exceptions, conflict-sensitive cases, and any lead that falls outside the standard rules should stay with a human.

About AI Automation Studio

AI Automation Studio is a London-based automation consultancy led by Boris Korol. The focus is practical workflow design for UK business-services and professional-services firms, especially law firms, accountancy practices, and operations teams.

If your firm is slow at sorting new enquiries, unclear on who should own them, or wasting senior time on first-pass triage, book a call:

https://ai-automation.studio/call

Client intake automation for law firms: how to capture and qualify enquiries 24/7

Boris Korol — Thu, 16 Jul 2026 19:17:38 +0000

Most professional-services firms, especially law firms, do not lose enquiries because the underlying service is weak.

They lose them because first response is slow, information arrives incomplete, and nobody owns the handoff.

That is an operations problem first.

Client intake automation for professional-services firms, especially law firms, is a workflow that captures enquiries, asks follow-up questions, sorts matters by fit and urgency, and prepares the information a reviewer needs before first response or first review. Its main value is reducing missed enquiries, shortening response times, and removing repetitive admin from the team.

If a firm wants fewer dropped enquiries and less intake admin, the goal is not to automate legal judgment.

The goal is to improve the path between first contact and first review.

The bottleneck usually sits before the legal work starts

When a new enquiry comes in, the firm has a short window to respond well.

That does not mean giving legal advice instantly.

It means:

acknowledging the enquiry quickly
collecting the right facts the first time
separating strong-fit matters from weak-fit ones
making sure the right person sees the right case

In many firms, that process still depends on a shared inbox, a reception handoff, or a fee-earner finding time between other tasks.

That creates predictable friction:

messages arrive out of hours and wait until the next business day
one person asks for details that should have been collected at the start
good enquiries sit in the same queue as poor-fit ones
fee-earners spend time sorting, chasing, and re-reading instead of reviewing prepared information

What intake automation actually does

Client intake automation does not replace legal judgment.

It improves the path between first contact and first review.

A typical workflow can:

capture enquiries from website forms, live chat, email, or a contact page
ask follow-up questions based on practice area
collect key triage information such as urgency, location, opponent, matter type, deadline, and contact details
answer common non-legal questions about process, next steps, office location, and likely response times
prepare a clean matter summary for a human reviewer
trigger reminders if the enquirer starts but does not complete the intake process
route the enquiry to the right team or queue based on rules

The gain is not only speed at the front door.

It is cleaner handoff quality inside the firm.

Where manual intake usually breaks down

Slow first response

Many firms still rely on somebody being available when the enquiry arrives.

If the message comes in at 7:40 pm, during lunch, or while the reception team is overloaded, the response is delayed.

Incomplete information

A short web form rarely gives a fee-earner enough to assess fit quickly.

That leads to extra messages, more back-and-forth, and longer time before anyone can decide whether the matter is worth progressing.

No consistent qualification logic

Different people ask different questions.

One enquiry gets screened properly. Another gets forwarded with almost no context.

That inconsistency makes the intake queue harder to manage and harder to improve.

Senior time gets used too early

Fee-earners should review matters, not do first-pass admin.

If senior legal staff spend time collecting basic facts, checking whether the matter fits, or chasing missing details, the firm is using expensive time on low-value process work.

What gets automated and what stays with a human

In a clean intake setup, automation usually handles the front-end administration:

capturing the enquiry
asking structured follow-up questions
collecting documents and contact details
checking basic fit rules
routing to the right department or queue
answering common process questions
preparing the first summary for review

The human still owns the parts that carry judgment and risk:

giving legal or professional advice
deciding whether to accept the matter
handling sensitive edge cases
reviewing any summary or first draft before it moves forward

That is the right split for most professional-services firms.

A before-and-after intake process

Before

Enquiry arrives by form, email, or chat.
Somebody notices it when they can.
A short reply asks for more information.
The prospect replies later, sometimes partially.
The enquiry is forwarded internally.
A fee-earner reviews fragmented context.
The matter is accepted, declined, or chased again.

After

Enquiry arrives through a structured intake workflow.
The workflow asks the right follow-up questions immediately.
Key facts and documents are collected in one pass.
The enquiry is scored or tagged by matter type, urgency, and fit.
A clean summary is prepared for review.
The right team receives the enquiry with context already attached.
A human decides how to proceed.

The legal judgment is still human.

The path to that judgment is simply more reliable.

Manual intake versus automated intake

Manual intake depends on staff availability, so response coverage is limited and information quality often varies by person.

Automated intake can run around the clock, ask the same structured questions every time, and route each enquiry with clearer logic.

In practice, the difference usually shows up in four places:

response coverage becomes broader
information arrives more complete
admin burden on fee-earners falls
queue visibility improves because fewer enquiries sit unowned

Implementation checklist

Define the top 3-5 enquiry types first.
Decide what information must be collected before review.
Separate process questions from legal-advice questions.
Set routing rules by department, location, or matter type.
Name the human owner for exceptions.
Add a review step before any matter is accepted.

FAQ

Does client intake automation replace a receptionist or fee-earner?

No. It removes repetitive intake admin, collects better information, and improves routing. A human still reviews the matter and handles legal judgment.

Can an intake workflow qualify legal enquiries without giving legal advice?

Yes, if it is designed to collect facts, check fit, and route the enquiry rather than advise on the law.

What should stay manual in a legal intake process?

Legal advice, final matter acceptance, sensitive exceptions, and review of any draft or summary should stay with a human.

About AI Automation Studio

AI Automation Studio is a London-based automation consultancy led by Boris Korol. The focus is practical workflow design for UK professional-services firms, especially law firms, accountancy practices, and operations teams.

If your firm is losing time in first response, triage, or internal handoffs, book a call:

https://ai-automation.studio/call

Hiring a Part-Time Finance Admin vs Deploying an AI Agent: A Real Cost Comparison for UK Firms

Boris Korol — Thu, 09 Jul 2026 09:25:46 +0000

Subtitle: A concrete 2026 cost breakdown for UK professional-services firms choosing between a part-time finance administrator and an AI automation agent.

For most UK firms processing under 5,000 transactions per month, an AI agent costs 60-75% less than a part-time finance administrator and usually handles more volume without sick leave, annual leave, or handover risk. In most cases, the break-even point sits between 800 and 1,200 transactions a month.

That headline matters because many firms still frame this as a staffing decision. It is not just that. It is a throughput, governance, and error-risk decision as well.

What does a part-time finance admin actually cost a UK firm in 2026?

The base salary for a part-time finance administrator in the UK might look manageable at first glance. A 0.5 FTE hire often lands around £14,000-£16,000 a year.

That is not the real employer cost.

Once you add employer's National Insurance, pension contribution, paid leave, recruitment fees, onboarding time, software, and the inevitable dip in productivity while the person learns your way of working, the more realistic first-year cost is higher.

Comparison snapshot:

Part-time finance admin (0.5 FTE)

Year 1 cost: £19,000-£24,000
Year 2+ cost: £17,500-£21,000
Best fit: relationship-heavy work, irregular processes, frequent edge cases

Scoped AI finance agent

Year 1 cost: £3,300-£9,300
Year 2+ cost: £1,800-£4,800
Best fit: high-volume, repeatable finance admin with stable rules and clear exception review

Additional cost items for the human hire usually include:

Recruitment: £1,500-£3,000
Training and onboarding: 4-8 weeks at partial productivity
Sick and holiday cover: 28 statutory leave days plus average sick leave in line with recent CIPD estimates

What does an AI agent cost to deploy and run?

A finance admin agent has two cost layers.

The first is setup. For a small or mid-sized UK firm, that usually lands between £1,500 and £4,500 depending on how many systems need connecting. Xero, QuickBooks, Sage, bank feeds, invoice capture, document ingestion, and approval logic all move the number.

The second is ongoing run cost. That is usually £150-£400 a month for API usage, orchestration tooling such as n8n or Make, and a small maintenance retainer.

Even at the upper end, the AI route is usually materially cheaper:

60% lower than a part-time admin in year 1
70-90% lower from year 2 onward

That does not mean it replaces every finance task. It means the economics are hard to ignore when the work is structured and repetitive.

What can each option actually handle?

This is where the comparison often gets lazy.

The AI agent is stronger at:

Transaction categorisation
VAT coding
Supplier matching
Bank feed reconciliation
Month-end pack assembly
Exception flagging for human review
Audit file collation

The part-time human is stronger at:

Supplier and client communication
New or ambiguous cases with no stable rules
Escalations that need judgement or authority
Handling context that lives in people's heads rather than in a system

At AI Automation Studio, we cleared 18,000 unreconciled transactions from a three-year backlog in under 72 hours. That kind of volume is exactly where a scoped agent wins.

Where does the AI agent fall short?

It falls short in three predictable ways.

First, bad inputs. If your data is inconsistent, the agent will surface that inconsistency fast.

Second, poor scope control. If you keep adding tasks without redesigning the workflow, errors creep in.

Third, no human review layer. An agent that gets 95% right but routes nothing for review is not production-safe.

That is why the right operating model is usually automation plus named human review, not automation on its own.

How do firms decide which path makes sense?

Three questions tend to settle it.

1. How much monthly volume are you processing?

If you are below 500 transactions a month, a part-time admin may be simpler. Once you move past roughly 1,500, the capacity advantage of the agent starts to dominate.

2. How stable are the rules?

If coding rules, approval thresholds, and reporting logic change every week, maintenance cost rises. If the logic is stable, the setup cost amortises quickly.

3. What is the cost of an error?

If you work in a regulated environment, the answer is not "let the AI handle it". The answer is "let the AI do the repetitive drafting and route the exceptions and approvals to a qualified human."

That is why many firms end up with a hybrid model: the agent handles the volume, the human handles the exceptions and relationships.

What does transition look like in practice?

For a 30-80 person professional-services firm, the pattern is usually:

Weeks 1-2: data audit, rules mapping, integration planning
Weeks 3-4: build and testing against historical data
Week 5: parallel run against the current process
Week 6: handover and exception routing
Weeks 7-8: stabilisation and edge-case cleanup

The parallel run matters more than the demo. It is where hidden exceptions show up.

The real decision

If the job is repeatable, high-volume, and expensive to staff, the AI agent is often the better economics decision. If the job depends on judgement, tone, and local knowledge, the human still earns their place.

The firms getting this right are not asking whether AI replaces finance admins. They are asking which parts of the process should stay human and which parts should never have needed a human in the first place.

If you want to run the numbers for your firm before committing, book a 30-minute scoping call at http://ai-automation.studio/call.

18,000 Unreconciled Transactions, Three Years of Backlog. Here's How AI Cleared It in 3 days

Boris Korol — Tue, 23 Jun 2026 10:47:16 +0000

Published June 2026 · Boris Korol, AI Automation Studio

A company came to us with 18,000 transactions that had not been reconciled since 2022. Three years of backlog, spread across several bank accounts. Transactions recorded in their banking app had never been matched to their internal accounting system. We built a four-layer AI matching pipeline. 3 days from receiving the first file to final output, only 100 transactions remained for a human accountant to review.

This article explains exactly how we approached it, where standard matching broke down, and what it took to close the gap.

The Scale of the Problem

The company had two systems that had drifted apart over three years.

Their banking app held the authoritative record of every payment made and received. Their internal accounting system held a parallel set of manually entered records, typed in by staff after each transaction. The intention was for the two to match. In practice, by the time we were brought in, 18,000 entries across multiple bank accounts had no confirmed match between the two systems.

This is not a rare situation. Manual double-entry creates drift. Staff change. Processes slip. The problem compounds silently until someone needs to produce audited accounts and the reconciliation has to happen all at once.

Three years of backlog at 18,000 transactions is a large but not extreme example of what happens when reconciliation is deferred. The ICAEW requires member firms to maintain accurate and complete accounting records — a deferred reconciliation backlog of this scale puts that obligation at risk.

Why Simple Matching Was Not Enough

The obvious first step is exact matching: find every transaction where the date and amount in the banking system match the date and amount in the internal system. This works well for straightforward cases and eliminates the bulk of the volume.

After running exact matching across all 18,000 transactions, we were left with approximately 3,000 unmatched records.

Three thousand unmatched transactions from 18,000 is a significant problem. Those 3,000 represent the cases where the data does not align cleanly, and each one could indicate an error, a duplicate, a misposting, or a legitimate discrepancy.

Standard reconciliation tooling stops here and hands everything to a human reviewer. We did not.

The Four-Layer Approach

Layer	Method	Input	Output
1	Exact match	Date + amount	Confirmed matches; ~3,000 unresolved
2	Grouped match	Amount sets across systems	Many-to-one and one-to-many resolved
3	Recipient analysis	Sender/recipient name fields	Mismatched-name transactions flagged
4	AI web search	~800 flagged transactions	~100 genuine discrepancies identified

Layer 1: Exact one-to-one matching

Date and amount match exactly across both systems. This cleared the majority of transactions and confirmed which records were straightforwardly aligned. Output: a set of confirmed matches and a set of unresolved records.

Layer 2: Approximate and grouped matching

Not every transaction maps to a single counterpart. Salary payments, for example, may appear as one transaction in the banking system and as several component entries in the internal system. We built a matching step that compared each record in one system against sets of records in the other, looking for groupings where the combined amounts aligned within a defined threshold.

This layer reduced the unresolved set further by identifying many-to-one and one-to-many relationships that exact matching cannot detect.

Layer 3: Recipient and sender analysis

With the remaining unmatched transactions, we ran a comparison of the named sender and recipient fields. The financial team had already flagged that some transactions had been sent to the wrong entity or recorded incorrectly, and that the banking record was the source of truth.

We flagged every case where the sender or recipient name in one system did not match the corresponding name in the other.

Layer 4: AI web search for company identity

This is where the problem became genuinely interesting.

Approximately 800 transactions remained where the names in the two systems appeared to differ but might still represent the same counterparty. The question was whether a name discrepancy was a data error or a legitimate difference in how the same company was recorded.

We built an AI agent that ran a web search for each of the 800 transactions. The agent searched for both names and attempted to determine whether they referred to the same legal entity.

In one example: the banking system recorded a payment to East Eagle Ltd. The internal system recorded the same transaction as South East London Ltd. To a human reviewer scanning rows in a spreadsheet, these look like two different companies. They are the same nightclub. East Eagle is the trading name; South East London Ltd is the registered legal entity — a distinction verifiable via the Companies House register.

The AI found approximately 100 more cases of this pattern. A trading name in one system, a legal entity name in the other. Same company, different label, recorded inconsistently across three years.

Results

Starting position: 18,000 transactions with no confirmed reconciliation status.

After running all four layers:

Confirmed matches covered the vast majority of the 18,000 transactions
100 transactions remained unresolved and required human accountant review
Total time from file receipt to output: 3 days
The agent now understands the matching logic and the company-name disambiguation approach. A follow-on batch of similar volume would take a few hours rather than 57

The reduction from 18,000 unknown-status transactions to 100 items for human review meant the accountants could do their work in a fraction of the time originally projected. They received a prioritised list rather than a raw data problem.

What This Pattern Looks Like in Other Engagements

Transaction reconciliation backlogs are not unique to the type of company described above. The same pattern appears in any organisation running parallel record-keeping across systems that were never fully integrated.

The specific technical approach will differ depending on the systems involved, the volume of transactions, and the nature of the discrepancies. But the four-layer structure, moving from exact matching through grouped matching, recipient analysis, and identity disambiguation, transfers across most reconciliation problems we encounter.

AI Automation Studio builds these pipelines for accountancy practices, audit firms, and finance teams dealing with reconciliation backlogs or ongoing reconciliation inefficiencies. Each engagement starts with a 30-minute scoping call to understand the data sources, systems, and volume involved before any build begins.

Frequently Asked Questions

How did you access the data from both systems?

The client provided exports from both systems in structured formats. We did not require direct system access or integration during the reconciliation phase. The AI agent worked from the exported files. Data is processed and stored within the client's jurisdiction — UK or EU — in compliance with UK GDPR and data residency requirements. For ongoing automation rather than a one-off project, we build live connections between the systems.

What happens to the 100 transactions that needed human review?

Those records were delivered to the client's accountants with the full context for each: what matched, what did not, and the closest candidate match found. The accountants made the final determination on each one. The AI did not make decisions about transactions it could not confidently match.

Can this work if one of the systems is not a standard accounting package?

Yes. The approach works on any structured data export. We have run similar pipelines against bespoke internal systems, legacy databases, and ERP exports. The matching logic is built around the data structure of the specific files provided, not around a fixed integration with a named software platform.

Book a 20-Minute Automation Snapshot

If your organisation has a reconciliation backlog, or if your team is spending significant time on transaction matching that should be automated, a short conversation is usually enough to establish what is possible.

Book a 20-minute Automation Snapshot with AI Automation Studio: http://ai-automation.studio/call

No pitch. No obligation. We look at the data sources and the volume, and give you a realistic picture of what an automated approach would involve.

Boris Korol is the founder of AI Automation Studio, a London-based automation consultancy working with accountancy practices, audit firms, and finance teams on transaction reconciliation, workflow automation, and AI agent development. Connect on LinkedIn.

We Reconciled 80,000 Bank Records Overnight. Here's the Exact Workflow.

Boris Korol — Tue, 23 Jun 2026 10:45:56 +0000

Published June 2026 · Boris Korol, AI Automation Studio

A UK accountancy practice needed to reconcile more than 80,000 bank records every month. The manual process took two full working days. We built an automation using Make.com and the Xero API. It now runs overnight.

This article explains exactly how the workflow is structured, what each component does, and what a practice needs to replicate it.

The Problem With Manual Bank Reconciliation

Bank reconciliation is one of the most time-consuming tasks in accountancy. For practices handling multiple clients, the volume multiplies fast.

The firm we worked with had a specific challenge: a high volume of transactions across multiple accounts, all needing to be matched, categorised, and reconciled against bank statements before month-end reporting.

Their team was spending two full working days per month on this task. That is roughly 16 hours of skilled staff time on work that is repetitive, rule-based, and automatable.

Manual reconciliation carries real risk beyond the time cost. Tired staff miss mismatches. Small errors compound. The closer you are to a deadline, the less time there is to investigate discrepancies. In a practice billing by the hour, that time is either absorbed as overhead or passed to clients who are increasingly resistant to paying for it.

For UK accountancy practices, this is not a niche problem. Making Tax Digital (MTD) is adding pressure to keep records cleaner and more current. Clients expect faster turnaround. Margins are tighter. The case for bank reconciliation automation in the UK is straightforward: this is skilled work being consumed by a task that software can handle.

The Workflow: Make.com + Xero API + Google Sheets

The automation runs across three tools: Make.com (formerly Integromat), the Xero API, and Google Sheets. No custom code. No enterprise software. The complete build took under two weeks.

Here is the workflow, step by step.

Step 1: Pull unreconciled transactions from Xero

Make.com connects to the Xero API using OAuth 2.0. A scheduled scenario runs nightly and pulls all new bank transactions from the client's connected accounts.

The Xero Bank Transactions endpoint returns each transaction with its date, amount, payee, reference, and status. We filter for records with a status of UNRECONCILED.

This step replaces the manual export from Xero that the team was running each morning before they could begin any matching work.

Step 2: Match transactions against the bank statement feed

Xero's bank feed pulls statements from the client's bank directly. The API exposes these via the Bank Statement endpoint.

Make.com iterates over each unreconciled transaction and attempts to match it against open statement lines using three criteria: amount (exact match), date (within a three-day window), and payee reference (fuzzy match using a text comparison module).

Transactions that meet all three criteria are flagged as high-confidence matches. Those that meet only one or two are flagged for human review.

The first version of the workflow used only amount and date matching. That produced a 60 per cent auto-reconciliation rate. Adding the fuzzy payee reference check pushed this to 85 per cent.

Step 3: Auto-reconcile confirmed matches in Xero

For transactions matched with high confidence, Make.com calls the Xero API to reconcile them automatically. This uses the PUT BankTransactions endpoint with IsReconciled set to true.

This step is where the volume is handled. For this client, approximately 85 per cent of transactions met the auto-reconcile threshold. That is around 68,000 records out of 80,000 processed without any human intervention in a single overnight run.

Step 4: Route exceptions to a Google Sheet for review

Transactions that do not meet the matching threshold are written to a Google Sheet. Each row includes the transaction details, the closest match found, and the reason it was flagged: amount discrepancy, date gap, or missing payee reference.

A member of staff reviews this sheet each morning. Instead of working through 80,000 records manually, they are looking at an exceptions list already sorted by confidence score.

The daily review dropped from two full days to a few hours.

Step 5: Log, summarise, and alert

Each nightly run writes a summary to a second Google Sheet: total transactions processed, number auto-reconciled, number flagged for review, and any API errors.

A Make.com email module sends this summary to the practice manager each morning before they start work. If the API returns an error or the match rate drops below a configured threshold, an alert fires to a Slack channel. The team knows immediately if something needs attention, without checking a dashboard.

Results

After the first full month of operation:

80,000+ bank records processed in a single overnight run
85 per cent auto-reconciled without human review
Two days of manual work reduced to a morning exceptions review
Zero reconciliation errors attributed to the automation across the first three months of production use

The practice did not change their Xero setup, switch accounting software, or hire additional staff. The automation sits on top of the systems they already use.

What This Means for Your Practice

Not every practice has 80,000 records a month. The workflow scales down without any structural change.

If your team is spending more than half a day a month on bank reconciliation, the pattern above applies. The specific triggers, matching rules, and thresholds are configured for your transaction volume and bank feed setup, but the architecture is the same.

The tools are the same: Xero API, Make.com, and a spreadsheet for exceptions. A practice that already uses Xero and Google Workspace has everything needed to run this.

AI Automation Studio works with UK accountancy and audit practices on this type of workflow. Bank reconciliation is one of three core automations we build alongside MTD compliance processes and payroll bureau workflows. The approach is always the same: one workflow at a time, with measurable output before expanding scope.

Frequently Asked Questions

Does this work with all UK banks connected to Xero?

It works with any bank that provides a live feed to Xero. That covers most major UK business banks including Barclays, HSBC, Lloyds, NatWest, and Starling. If the client's bank provides a statement feed in Xero, the automation can process those transactions.

What happens if the Xero API goes down or changes?

Make.com handles API failures by retrying failed operations and logging errors. The nightly run flags any API issues in the morning summary email and fires a Slack alert if configured. The practice can reconcile manually in Xero exactly as before. The automation is additive: it does not replace the underlying Xero workflow, it runs on top of it.

Where is client financial data stored?

The Google Sheet is owned by the practice and sits within their own Google Workspace. Make.com does not store transaction data beyond the duration of the scenario run. The Xero connection uses OAuth 2.0, so no credentials are held in the automation. All data stays within systems the practice already controls and is responsible for.

Book a 20-Minute Automation Snapshot

If your team is still reconciling bank records manually, a short conversation is usually enough to establish whether automation is the right fit and what the build would involve.

Book a 20-minute Automation Snapshot with AI Automation Studio: https://ai-automation.studio/call

No pitch. No obligation. We look at one workflow, identify the opportunity, and give you a clear picture of what it would take to build it.

Boris Korol is the founder of AI Automation Studio, a London-based automation consultancy for UK accountancy and audit practices, specialising in bank reconciliation, MTD workflows, and payroll bureau automation. Connect on LinkedIn.