Under the billable work in an accounting, law or consulting firm sits a layer of administration that fee earners and practice staff carry between them. Onboarding forms, identity checks, engagement letters, records requests, timesheets, WIP, trust ledgers and debtor lists move between the practice management system, email, the document system and the accounts, mostly by hand and often by the people with the highest charge-out rates. Client intake, year-end records, trust accounting, billing and collections, tenders, inboxes, privacy records, partner reporting and recruitment all sit in that layer. We automate the predictable steps as fixed, tested software that follows your checklists and billing rules the same way every time. AI is used only where something has to be read or drafted, such as a trust deed or a client’s reply, and its output is checked before anything leaves the system.
These are patterns we see in professional services firms, not client case studies. Your process gets its own map in the assessment.
Onboarding new clients under AML/CTF
Since 1 July 2026, accountants, lawyers and conveyancers who provide designated services have AML/CTF obligations, including customer due diligence. A new client or matter now means verifying identity, finding the beneficial owners of a company or trust, rating the risk and keeping records. Where the client is a company or trust, AI reads the trust deed, company extract or partnership agreement and lists the people who may be beneficial owners, and a staff member checks that list against the documents. Everything else follows your checklist, identically for every new matter, from the call to the ID verification service to the engagement letter and the record kept of each step. The risk rating and the decision to act stay with the responsible partner. We handle the records and routine steps, and the firm keeps responsibility for its obligations.
Turning recorded time into invoices
Fee earners record time late, with narratives like “emails” or “review file”. Before each billing run a partner or billing clerk works through the WIP, checks fixed-fee arrangements and disbursements, rewrites narratives and produces the invoices, often over a weekend. Each draft invoice is built from WIP and the engagement terms and checked for completeness on rules, flagging unbilled disbursements, matters with no time recorded against an agreed milestone, and fixed fees that have run over. AI plays a small part, reading narratives to flag entries a client would not understand and suggesting plainer wording for the fee earner to accept or change. Write-offs and the final invoice stay with the partner, who approves each one before it is sent.
Client records and requests
Each year the firm sends clients a list of what it needs, then spends weeks chasing bank statements, rental summaries and loan statements. They come back by email, by portal upload and occasionally in a shoebox. Automated, each client’s request list is built from last year’s file and sent, reminders follow on the schedule you set, and each item is ticked off as it arrives. AI sorts what comes in, reads each document to work out what it is and which period it covers, and files it against the request. Anything unreadable or unexpected goes to the accountant on the job. Workpaper preparation starts once the list is complete, and the professional judgement on the return stays with the accountant.
Accounting firms prepare periodic reports for clients’ banks, such as covenant compliance certificates, management accounts and annual reviews, each against the terms of a different facility agreement and a different due date. Missing one can put a client’s facility at risk. Fixed software pulls the figures from the client’s ledger, calculates each covenant using definitions your accountant has configured and checked against the agreement, diarises every due date and produces the report in the format the lender expects. Where a calculation shows a covenant close to breach, it goes to the client manager early. No AI is needed. The accountant reviews and signs each report, and conversations with the lender stay with the firm and the client.
Reception@, accounts@ and admin@ inboxes in a firm take in new matter enquiries, client documents, ATO and ASIC correspondence, court and tribunal notices, and invoices from barristers and other suppliers. Someone reads each one and forwards it, and a notice with a deadline can sit unopened under a pile of newsletters. AI reads each message and attachment, classifies it and extracts the client, the matter and any date it mentions. Routing rules send it to the right team or into the right process. Anything that looks like a deadline or hearing date is flagged for the responsible fee earner to confirm, and it never goes into the diary unchecked. Messages the AI cannot place with confidence wait in reception’s queue.
Collecting fees and reporting to partners
Chasing fees is awkward when the debtor is a client the partner has known for years, so it gets left. Your practice or accounting software can already send basic reminders. The work that remains is matching and the follow-up that does not fit a template. AI reads remittance advice and client replies and sorts them into paid, promised, disputed or asking for a payment plan. Around that, reminders go out on the cadence each partner sets, payments are matched to invoices, including one payment across several matters, and every contact is logged against the client. Anything off-template gets a drafted follow-up, which the relationship partner approves before it is sent. Disputes and hardship go straight to that partner.
At month end the practice manager exports WIP, billings, collections, lockup and utilisation from the practice management system, adds figures from the ledger, pastes it all into a spreadsheet and formats a pack for the partners’ meeting. It takes days and the figures are already stale. Automated, the data comes out of each system, the same calculations run every month and the pack is laid out in your format, so the numbers reconcile to the ledger before anyone reads them. AI drafts the commentary, noting where lockup has grown or a practice group is behind budget. The practice manager edits that commentary and decides what goes to the partners.
Trust money and privacy records
A law practice holding client money has to record every receipt, transfer and disbursement against the right client ledger and reconcile the trust account on the schedule its trust account rules set. An error is a serious matter for the practice and the principal. The work is entirely rules, so no AI is used. Fixed software matches bank transactions to trust ledger entries, prepares the reconciliation of bank statement, cash book and client ledgers, and lists anything that does not agree, such as a ledger in debit. The principal or trust account supervisor reviews and signs every reconciliation. The software keeps the records and does the routine matching, and the trust obligations remain the practice’s.
From 10 December 2026, APP entities must disclose in their privacy policy the kinds of personal information used in, and the kinds of decisions made by, computer programs that could significantly affect individuals. A firm using software to screen clients or triage enquiries needs to know where those programs are. A register records the systems, the personal information each holds and the decisions each supports, including any automation we build for you. Access requests are logged and their response dates tracked. AI helps find a person’s records across systems for a request, and a staff member reviews everything before release. The wording of your policy is a matter for you and your privacy adviser.
Winning work and hiring
Government and corporate panel applications, requests for tender and proposals for larger engagements ask for firm credentials, team CVs, professional indemnity details, methodology and rates, each in its own format and to its own deadline. Business development staff rebuild the same answers every time, pulling CVs that are out of date. We keep a library of approved answers and current documents, with a compliance matrix that maps every question to its response. Assembly and formatting run on rules. AI drafts answers to new questions from the approved content only and marks anything it could not source. Pricing, team selection and the decision to respond belong to the partners, and a named partner owns every submission.
A graduate program or a paralegal advertisement can bring in more applications than a small HR team can read, and candidates drift to other firms while they wait to hear back. Every application is acknowledged, interviews are booked against partners’ calendars and each candidate gets a status update at every stage. AI screening needs care. It can summarise each application against your published criteria for a reviewer, but HR staff and the partners read the applications and make every shortlisting and hiring decision, with attention to fairness and privacy. Any screening software belongs in the automated decision register described above, since privacy law will require disclosure of automated decisions from 10 December 2026.
One entry between practice management and the ledger
Where a firm runs a practice management system, a separate document system, a CRM and a general ledger, the same client details, matters and invoices get typed into more than one of them. A client changes its name or adds an entity, and the systems disagree for months. We measure all of that double entry as one process. Nearly all of it is fixed software, with mappings between systems, a scheduled sync, a reconciliation that compares them and an exception queue for records that disagree. AI appears only where a free-text note has to be mapped to a field, and that mapping is reviewed before it is saved.
If one of these costs you real money every month, describe it in four answers.
Thinking about what a process actually costs you?
The assessment is small, fixed-price, and tells you the real number, whether or not you ever build.
Worth a 30-minute conversation.