Preparing for a Trust Account Audit Without Panic

The letter arrives on a Tuesday. It is polite, it is short, and it gives you a date. Somewhere between two and six weeks from now, depending on your regulator, someone is going to sit in your conference room and ask to see how you have handled other people's money. The partners read it twice, forward it to the bookkeeper, and then the firm quietly stops doing anything else for a fortnight.

That fortnight is the tell. A firm that panics when the examination notice lands is a firm that has been keeping its trust records in a state that requires assembly. The records exist, scattered across a bank portal, a spreadsheet, a shoebox of deposit slips, and the memory of whoever has been at the firm longest, but they do not exist as a record. They exist as raw material for a record, and someone now has to build the thing under time pressure while also practising law.

The firms that get through examinations without disruption are not better at preparing. They are firms where preparation is not a distinct activity, because the record the examiner wants is the same record the firm uses on an ordinary Wednesday to answer the question "how much of this client's money is sitting in the account right now." This piece is about what examiners actually ask for, why reconstruction is the real risk rather than dishonesty, and which specific documentation gaps convert a routine review into a written finding.

What an Examiner Is Actually Testing

The instinct in most firms is that the examiner is hunting for theft. Almost never. Deliberate misappropriation is a small fraction of what these reviews surface, and when it happens it is usually already obvious to everyone involved. What the examiner is testing is whether your records can prove, at any arbitrary point in the past, that every dollar in the trust account belonged to an identifiable client and that the total of all client balances equalled the money actually in the bank. That is the entire question. Everything else is a route to answering it.

This matters because it changes what "being ready" means. You are not preparing a defence. You are demonstrating that a control existed continuously, not that a number happens to be correct today. An examiner who sees a perfectly balanced account with no evidence of how it stayed balanced for the last three years has learned very little, and will start asking for the working papers. An examiner who can pull any month at random and see the same disciplined comparison performed on time, with the exceptions listed and explained, has learned everything they came for and will usually finish early.

The Three-Way Reconciliation Is the Whole Exam

Strip away the paperwork and every trust examination reduces to a single comparison performed repeatedly over time. The bank's balance, adjusted for outstanding items, must equal your trust cash book or control account balance, and both must equal the sum of every individual client ledger. Three numbers, one answer. Regulators across common-law jurisdictions describe this differently and demand it on different cycles, but the arithmetic is identical whether you are in Texas, Ontario, Manchester or Brisbane.

The cycle is where jurisdictions genuinely diverge, and you must confirm your own rather than assume. The SRA Accounts Rules in England and Wales set an outer limit on how often client account reconciliations must be performed and signed off, and firms there are also subject to an accountant's report regime with defined exemptions. The Law Society of Ontario requires a monthly trust comparison prepared within a set number of days after month end. Australian jurisdictions layer an annual external examination on top of ongoing reconciliation duties administered by the relevant state or territory legal services body. In the United States there is no single rule at all, because trust obligations sit with each state bar, and requirements for reconciliation frequency, record retention periods, and interest handling on pooled accounts vary materially from one state to the next. Read your own rule text and calendar the actual deadline rather than working from what a colleague in another state told you.

The Records Request Letter, Decoded

Regulators phrase their requests differently but ask for a consistent core. Expect bank statements and cancelled item images for the full review period, deposit records showing the source of each receipt, the receipts and disbursements journal for the trust account as a whole, individual client ledgers for every matter that held funds during the period, your completed reconciliations with supporting schedules, and the underlying authority for withdrawals. That last category is the one firms underestimate. The examiner does not merely want to see that money left the account. They want to see the document that entitled you to remove it.

For an earned fee transfer that means the invoice, dated on or before the transfer, itemising the work. For a settlement disbursement it means the settlement statement and the client's written direction. For a filing fee it means the receipt or the court's demand. For a refund it means the calculation showing what remained unearned. Firms that bill from a timekeeping system and move money from a banking portal, with nothing connecting the two, end up manually pairing hundreds of transfers to hundreds of invoices during the review window. That pairing exercise is where errors surface, and the errors surfaced under pressure are the ones that become findings.

Reconstruction Is the Real Risk

Here is what nobody warns small firms about: the act of rebuilding a trust record retroactively is itself dangerous. When you reconstruct, you are inferring. You decide that the unlabelled deposit in March was probably the Hendricks retainer because the amount is close and the timing fits. You allocate a bank charge to the client whose balance can absorb it. You write a ledger entry today describing an event from eighteen months ago based on what the amount suggests must have happened. Every one of those inferences may be correct, and every one of them is unsupported.

An examiner who spots a pattern of contemporaneously dated entries that were plainly created in a single sitting will treat the entire record as unreliable, and they are right to. The value of a trust ledger comes from it being written at the time, by the person who knew, from the source document in front of them. That property cannot be added later. This is why the honest firm with disorganised records often has a worse examination than the disciplined firm with a small genuine error, and it is the single strongest argument for keeping trust records inside the system that also holds the matter, the invoice and the document rather than in a general ledger that knows nothing about either.

The Documentation Gaps That Become Findings

There is a short list of gaps that account for a large share of written findings, and none of them involve missing money. The first is a client ledger that does not exist for funds that provably passed through the account, usually because a receipt was recorded against the firm's general trust balance without matter attribution. The second is a transfer to the operating account with no invoice, or with an invoice dated after the transfer. The third is a bank charge, wire fee or returned item deducted from trust and never reimbursed from operating funds, which quietly makes every client in the pool short by a few cents. The fourth is a deposit whose source is unidentified, which is worse than it sounds because unattributed money in a client account is not obviously yours to hold.

The fifth and most consequential is a negative individual client ledger, even for a single day. A trust account with a positive overall balance can still contain a matter that went into deficit because a disbursement was made against funds that had not yet cleared or against another client's money. Regulators treat that as using one client's funds for another, regardless of intent and regardless of whether the pooled account ever went overdrawn. Firms miss it because their bank shows one number and their bookkeeping shows one number, and neither of them is per-matter. This is precisely why trust accounting in Casely blocks any disbursement that exceeds a matter's actual trust balance at the database transaction level rather than raising a warning that a busy person can dismiss. The ledgers are isolated per matter, so the condition the examiner is looking for cannot come into existence in the first place.

Timing: The Transfer That Precedes the Invoice

Sequence is a compliance fact, not an administrative detail. In most jurisdictions the entitlement to move money from client account to office account arises when the fee is earned and the client has been billed or notified in the manner the rule specifies, and the record must show that the entitlement existed before the money moved. A transfer executed on the fifteenth against an invoice generated on the twentieth is a defect on its face, even if the work was genuinely done on the first.

Firms create this problem structurally when billing and banking live apart. Somebody notices the operating account is thin, moves a round number across from trust, and tells the bookkeeper to bill against it later. It feels like a cash flow decision. On the record it reads as a withdrawal without authority. The fix is not more diligence, it is making the invoice the trigger. When one-click invoicing turns every unbilled hour on a matter into a single itemised draft, and the transfer is raised against that specific invoice, the sequence is correct because there is no other way to do it. LEDES 1998B export matters for the same reason on institutional work, since the billing detail an insurer or corporate client receives should reconcile to the same entries backing your trust movement.

Dormant Balances, Stale Items and Unclaimed Funds

Every long-running trust account accumulates residue. A closed matter with eleven dollars left. A cheque issued four years ago that was never presented. A refund to a client whose forwarding address stopped working. Individually trivial, collectively a recurring examination theme, because holding money you have no current instruction to hold is a live obligation rather than a rounding issue. Most regulators expect firms to make documented efforts to return residual funds and, failing that, to deal with them under the applicable unclaimed property or escheat process. The specifics differ sharply. Some jurisdictions direct residual client funds to a bar foundation or law society fund, others to a state or provincial unclaimed property administrator, with different dormancy periods and notice requirements. Confirm yours before you move anything.

The operational failure is that nobody looks. A dormant balance report is a five second query if your ledgers are per matter and your matters carry a status, and it is an afternoon of spreadsheet work otherwise, which means it happens once a year at best. Attaching the review to something the firm already does, like the point where a matter stage tracker moves a file to closed, converts it from a project into a step. The residual balance question becomes part of closing a matter rather than an annual archaeology exercise, and the effort to return the money is documented at the moment it is made, which is the only time the documentation is worth anything.

Corrections, Voids and the Deleted Entry Problem

Every set of books contains mistakes, and examiners know it. A wrong amount, a receipt posted to the wrong matter, a duplicated entry. None of these are findings on their own. What turns them into findings is how they were fixed. If the correction was made by editing the original entry, the record now asserts something that was never true at the time, and there is no way for anyone, including you, to see what actually happened. If the ledger was reprinted after the edit, the printed record and the earlier printed record disagree, and you cannot explain the difference.

The defensible pattern is that nothing is ever deleted. The erroneous entry stays, marked void, with the reversing entry and a note recording who made the correction and why. The ledger becomes slightly longer and completely honest. Casely enforces this by design, voiding corrections and keeping them visible rather than removing them, which means the reviewer sees the mistake and the fix side by side. That is a far better outcome than a clean ledger the examiner has no reason to trust, and it removes the temptation, which is real in a small firm under pressure, to quietly tidy history before handing it over. Similarly, keeping a comment field on every document that records what changed and why extends the same discipline from the ledger to the supporting file.

Who Touched the Account, and Who Could Have

A part of the review that surprises firms is access. Examiners ask who has signing authority on the trust account, who can initiate transfers, who reconciles, and whether those are the same people. Segregation of duties is the point. A firm where one person receives funds, records them, disburses them and performs the reconciliation has no control at all, only a trusted individual, and regulators have watched enough matters go wrong to be sceptical of that arrangement even when the individual is beyond reproach.

The same logic applies inside your practice management system. If everyone can see and touch every matter's financial records, then your access record says nothing. Granular access matters here for a compliance reason rather than a privacy one, and it needs to be real. Ethical walls enforced at the server and data access layer, rather than hidden in the interface, mean a restricted user genuinely cannot reach a walled matter through search, the calendar, or a forwarded link. When you can show an examiner that access was constrained by the system rather than by convention, the conversation about who could have altered a ledger ends quickly. Conflict checking that searches the full contact and matter history, across every role a party played including closed matters, supports the same evidentiary posture when the question turns to why funds were held for a particular party in the first place.

The Two Weeks Before, Done Properly

If the notice has already arrived, work backwards from the exam rather than forwards from the earliest record. Start with the most recent reconciliation and confirm it is complete, supported and signed. Then produce a list of every matter with a trust balance and eyeball it for anything negative, anything unusually round, anything belonging to a closed matter, and anything you cannot immediately explain. Those exceptions are your work list, and there will be fewer of them than you fear.

Then handle the transfers. Every movement from trust to operating during the review period needs its authorising invoice attached and dated correctly. Where the sequence is wrong and cannot be fixed, do not hide it. Prepare a short written explanation of what happened and what you have changed to prevent recurrence, and hand it over before you are asked. Examiners respond well to a firm that identified its own defect and closed it, and badly to one that let them find it. The worst possible posture is a firm that discovered the problem during preparation, said nothing, and hoped the sample would miss it.

Building a Firm That Is Never Not Ready

The permanent version of all this is unglamorous. Reconcile on the cycle your regulator sets and treat the deadline as a real one. Never let money move out of trust without a dated document authorising it. Never let a receipt land without a matter attached. Never edit history, only add to it. Review residual balances when a matter closes, not when someone remembers. Keep the people who move money separate from the people who check it. Six habits, none of them difficult, all of them impossible to retrofit.

What makes those habits sustainable in a firm of five or fifteen people is not discipline, because discipline degrades under deadline pressure and always has. It is that the system refuses the shortcut. A disbursement that exceeds a matter's actual trust balance should not be possible, not discouraged. An invoice should be the thing that produces a transfer, not a document you write afterwards to justify one. A correction should leave a visible scar rather than a clean surface. When those constraints live in the software, the audit-ready record is simply the byproduct of ordinary work, and the notice on a Tuesday becomes a scheduling matter rather than a crisis. If you are building that foundation now, start with how the money is held and controlled in trust accounting software for law firms, then make sure the billing side produces the authorising documents in the right order with legal billing software.

None of this is legal advice about your obligations, and the rules genuinely differ. Reconciliation frequency, retention periods, interest handling on pooled accounts, residual fund procedures and external examination requirements all vary by state, province, territory and country. Confirm the text of your own regulator's rule and calendar the specific deadlines it creates. What does not vary is the underlying test, which is whether your records can prove continuous control over money that was never yours. Build for that, and the rest is paperwork.