Preparing for a quality review, the compliance officer of a firm with six offices asks the IT partner which AI tools are in use. The answer is an email with four names: Microsoft 365 Copilot, ChatGPT, the bookkeeping package and the scan-and-recognise software for purchase invoices. That same afternoon she watches a colleague in the compilation team have a general ledger explained in a tool she does not recognise.
That is a firm's AI register in one picture: a short list of what IT set up, next to a working day in which AI sits in more places than anyone has reported. What belongs in that register for any organisation is covered in an AI register: what it is and what to record. This piece is about what is different in an accounting firm: which tools, which fields, and where the second source comes from.
Which AI runs in an accounting firm?
More than the chatbot. The general assistants are there: Copilot in the Microsoft environment, ChatGPT and Claude for a client email or a note to the accounts, Perplexity for a tax question with sources. But most of the AI in a firm sits in software that has been in use for years. The bookkeeping package suggests journal entries, the scan-and-recognise software reads invoices, the compilation software drafts a first note, and the data analytics tool in the audit practice summarises exceptions.
That second group never had a new domain, a new account or a new purchasing decision: the vendor added the feature in an update and the firm has used it since. It is usually missing from the register because nobody recognises it as AI. Why that is structural is explained in why tool catalogs age the moment you print them.
Then there is a third group: the tools an employee found on their own. A transcription tool for the annual accounts meeting, a PDF summariser for lease contracts in an audit file, a browser extension that rewrites emails. They are recorded nowhere, which is not a violation but shadow AI: self-serve, without a decision, and actually in use.
Which fields matter more in a firm?
The basic fields are the same for any register: vendor, purpose, owner, risk assessment, data use, decision, date and review. In a firm, four fields do more work than elsewhere.
| Field | What you record | Why it matters in a firm |
|---|---|---|
| Client information allowed | Public only, anonymised draft, or file contents too; and what never goes in | This is where the duty of confidentiality lands in practice, per tool |
| Data processing agreement and location | Whether there is an agreement, where processing happens, whether the vendor trains on input | Without an agreement the tool is unsuitable for clients' personal data |
| Practice | Compilation, audit, tax, payroll, advisory | The same tool is a different risk in payroll than in advisory |
| Embedded or standalone | AI feature inside an existing package, or a service with its own domain | Embedded AI had no purchasing moment and otherwise stays out of sight |
The first field is what separates a list of tools from a register of decisions. Confidentiality is one of the principles in the Dutch professional code for accountants (the VGBA), and the accountant is expected to make sure that people working under their responsibility observe it too. The register is where that becomes concrete per tool: a public statute may go into Perplexity, a client name may not; the file may go into Copilot, within the permissions already on it; nothing about a client goes into the free ChatGPT account. How the professional rules apply precisely differs per firm and is for the firm and its compliance officer to judge.
The second field hangs on the account type. A free or personal account has no data processing agreement, and input can be used for training unless the user switches that off. The difference is set out in free versus business AI accounts; whether an agreement has to sit underneath, in do you need a data processing agreement for an AI tool. If a tool processes personal data of clients or staff, the same application also belongs in the record of processing activities under the GDPR, with a cross-reference both ways.
What does a survey show, and what does it miss?
A firm has allowed Copilot for client work and ChatGPT on a business account for text work without file contents. The register has eight rows, filled by a survey among the practice leads. A first look at which AI services are opened from the browser shows that, next to those two, Perplexity and a free ChatGPT account are in use, plus three specialised tools: a transcription tool in the advisory practice, a PDF summariser in the audit practice, and the AI note in the compilation software. Two of them are recorded nowhere.
That is not an incident, because nothing had been decided about those two. It shows what a survey is: a snapshot, taken by asking people who know about a tool. Whoever pays for a tool privately does not report it. Whoever started using one last week was not there at the survey. And whoever does not know that the new note in the compilation software is AI cannot report it.
How does observed use feed the register?
By putting a second source next to the register: which AI tools are actually opened in the firm, how often, and whether a decision exists. Not who, not what was typed, only which tool, at firm level. The difference between that list and the register is the backlog, and every tool in it enters the register on the most honest status there is: no decision yet. The steps after that are in from shadow AI to a current AI register.
BeeSensible provides that second source. The extension reports only the domain of an opened AI tool, at most once a day per device, with no user id and no page content, and sets it against a catalogue of 865 assessed tools with the vendor, country, training policy and a risk score for each. An admin sees in the inbox which tools were newly noticed and still have no decision, and records the decision; BeeSensible never chooses itself (see the AI tools page).
