AI readiness assessment for accounting firms

It is the last week of the BAS lodgement rush. A graduate has a client's bank export open, a messy set of transaction descriptions, and a free chatbot in the next tab that will tidy the lot in seconds. Nobody told them they could. Nobody told them they could not. Plenty of practices sit in that gap right now.

An AI readiness assessment for accounting firms is a way to see that gap clearly. Calibre Forum is a short survey every person answers on their phone, from partners to the bookkeeper who started last month. You get one branded report showing where AI is already in the work, which tools are involved, and which controls your insurer and your clients are likely to ask about.

Where AI shows up in accounting firms and bookkeeping practices

Bank feeds, coding and reconciliations

Worth finding out whether anyone pastes transaction lists or bank statement extracts into a general chatbot to suggest account codes, or to explain odd entries. Separately, which bookkeepers have switched on the AI features inside their ledger software, and does anyone check what those features send away or store? The answer is rarely yes or no. It is usually a handful of people doing it quietly, each with a slightly different method.

Client emails and advice letters

Ask who drafts replies to client queries about deductions, super or instalments with an AI assistant, and whether the client's name, ABN or figures go in with the question. The same goes for engagement letters, year-end summaries and the polite chaser for missing records. A tool register built from what people say they use shows the pattern without anyone having to be caught out.

Tax research and technical memos

Staff may use AI to get a first read on a ruling, a trust distribution question or a fringe benefits issue. The things to learn: who relies on the answer, who checks it against the primary source, and whether the working paper shows where the conclusion came from. Junior staff often lean on it most, which makes the question about supervision and training as much as tooling.

Document capture and file notes

Receipt capture, invoice reading and meeting transcription tools often arrive as browser extensions or phone apps rather than approved software. Find out which ones are running, on whose accounts, and whether recordings of client meetings are being kept anywhere the practice cannot see. These tools tend to arrive one subscription at a time, so the firm's real list is longer than the approved list.

What puts accounting firms and bookkeeping practices at risk

Tax file numbers and client financials in personal accounts

Client financial statements, TFNs and bank details are among the most sensitive records a practice holds. If staff paste them into a personal chatbot account, the data sits under that person's terms, not the firm's. The report shows how common that is among your team, and where it is happening. It also matters for payroll files, trust and estate work, and anything where a client would be upset to learn a stranger's system held their records.

Professional obligations and review trails

Registered tax agents carry responsibility for the work that goes out under their name. An AI-drafted position that nobody reviewed is still the practice's position. Partners should know where AI touches advice, and whether the review step that already exists is actually being done. The useful question for partners is not whether to ban or allow, but whether the review that exists on paper survives a deadline week.

Insurer and client questions at renewal

Professional indemnity and cyber renewals may ask what rules apply to AI tools and client data. Larger clients and lenders may put similar questions in engagement terms. A scored controls checklist gives you something concrete to answer with, instead of a guess. Knowing the honest position before the questionnaire arrives is a calmer place to start than answering in a hurry.