How to automate AML and KYC client onboarding in a small accountancy practice

Most of AML and KYC onboarding is process, not judgement: chasing documents, running checks, scoring risk, building the file. Here's what can be handed to an AI employee, what stays with the partner, and what it changes.

For a small accountancy practice, client onboarding has quietly become one of the biggest admin burdens in the business. Most of it is AML and KYC: collecting proof of identity, running checks, scoring risk, and building a file a partner can sign off.

Almost none of that is judgement. The judgement - the decision to accept a client, and at what risk rating - takes minutes. The hours go on everything that leads up to it.

That split is the whole opportunity. Here's what can be automated, what should stay with the partner, and what it actually changes.

What takes the time today

In most owner-run practices the onboarding admin falls on one person, often a senior one, repeating the same steps for every new client:

Chasing proof of identity and address, usually more than once
Re-keying client details into the practice system
Running the AML and sanctions search and saving the result
Scoring the client against the practice's risk policy
Assembling the file so a partner can review and sign off

Every step before the sign-off is process. Only the final decision needs a qualified human.

What an AI employee handles

The approach that works is not another piece of software for the team to learn. It is a managed AI employee that runs inside the tools the practice already uses, given a defined role:

Send and chase document requests automatically, so nobody is nagging clients by hand
Read the documents, extract the details, and populate the system
Run the AML and sanctions check through the practice's existing provider
Apply the practice's own risk rules and produce a scored summary
Hand the partner a complete file with a recommendation, for a yes or no

The regulated decision stays with the partner. The chasing, keying, checking and filing do not.

How much time it saves

Industry analysis of KYC and onboarding automation points to up to 80% of onboarding time removed, and AML review time cut by around 60%, when the repetitive work is handled this way. Financial firms automating data entry, document handling and checks have reported KYC cost reductions of up to 70%.

For a small practice the more useful way to read those numbers is simpler: the senior person currently doing onboarding gets most of their week back, and the practice can take on more clients without another hire.

Is it compliant?

Yes, when it is built so the agent prepares and the human decides. The AML decision and the client risk rating remain a documented human sign-off. The agent's job is to assemble a complete, consistent, audit-trailed file so that decision is faster and better evidenced, not to make the regulated call itself.

What it costs, in rough terms

Less than the alternative, which is a £25,000 to £30,000 onboarding or compliance hire. A managed AI employee is a one-off setup fee plus a monthly management fee, and it is run and improved for you rather than handed over as software for your team to operate.

Who this suits

Owner-run accountancy practices where the principal still signs off onboarding personally, AML is a real and growing burden, and the admin is starting to limit how many clients the practice can comfortably take on.

The AR Dept. builds and deploys AI employees for owner-managed businesses, including accountancy practices, law firms, and recruitment agencies. If you'd like to understand whether your onboarding is a fit, get in touch.