The short answer: A good accounting hire should understand the accounts, use your approved AI tools and check the result against the evidence. Test those abilities together. Faster preparation only helps if it reduces the work needed to produce dependable accounts. Around Certified is the name for our training approach, covering accounting skills, modern tools and preparation for your firm. The specific systems, responsibilities and market knowledge must still be agreed for the role.
If a new team member can produce a polished explanation in seconds but cannot explain an unreconciled balance, you still have a review problem.
AI belongs in the hiring conversation because it changes how some work gets prepared. It also changes what you need to check. The useful question is whether the person can turn your records into reliable work, explain what they have done and recognise when a decision needs a more experienced accountant.
Start with the accounting, then assess the tools
In my published response to Claude’s small-business accounting pitch, the test I came back to was clean, accurate data. Accounting involves judgement: accruals, deferred revenue, opening balances, adjustments and the explanation behind a posting. A tool needs someone who can recognise when its answer does not fit the underlying transaction.
That judgement should shape your recruitment brief. If you need a bookkeeper to prepare reconciliations, assess their ability to trace a balance to the ledger and supporting records. If you need a senior accountant to review work, assess the quality of their questions and their treatment of unresolved issues.
Software familiarity sits beside those requirements. It cannot stand in for them.
Look for three practical abilities
1. Produce and explain sound accounting work
The candidate should be able to show where a figure came from, explain the treatment used and identify missing evidence. A sensible question raised early is more valuable than a confident answer built on an assumption nobody has checked.
2. Use AI within a defined task
Give the tool a clear purpose, approved inputs and an expected output. An example might be preparing a first draft of a commentary from an already checked management-accounts pack. The accountant then checks the figures and tests whether each explanation follows from the evidence.
Generative AI can produce convincing false statements. The US National Institute of Standards and Technology’s Generative AI Profile identifies this risk and recommends checking sources and citations. For an accounting role, that makes verification part of the task.
3. Recognise the limits of their authority
A person can prepare an adjustment without having authority to post it. They can draft a client response without having authority to send it. They can investigate an unusual transaction without being the person who decides its tax treatment.
Write those boundaries into the role. The same boundaries should apply when AI helps with the preparation.
Assess the systems your firm actually uses
Our technology conversation includes Xero, QuickBooks, XBert, Claude Cowork, Notion, n8n, Make, Fireflies and Vinyl. A useful role brief names the systems the person will need, the tasks they will perform and the level of competence required. It does not need a long list of every product in the market.
| Area | What to establish |
|---|---|
| Xero or QuickBooks | Can the person find the source records, prepare the work and explain the resulting balances? |
| XBert | Can they investigate an alert and document what they found, rather than treating an alert as a conclusion? |
| Claude Cowork | Can they define the task, limit the information supplied and check the output? |
| Notion | Can they find and maintain the instructions and decisions needed for their work? |
| n8n or Make | Can they explain the workflow they use, its exception route and who owns changes? |
| Fireflies or Vinyl | Can they check a transcript or meeting summary against what was actually agreed? |
XBert’s documentation describes alerts for possible errors, anomalies and risks in accounting data. The alert is a starting point for investigation. Anthropic describes Cowork as working across files and connected tools. Access, configuration and the instructions you give it matter. Neither product description establishes an individual candidate’s ability.
Use a short work sample in the interview
Give the candidate a small fictional file with a ledger extract, supporting documents and one unresolved discrepancy. This is an assessment illustration, not a client case. Avoid supplying real client information to an unapproved tool.
- Ask them to prepare the reconciliation and explain the discrepancy.
- Allow an approved AI tool to help draft a short explanation.
- Ask which statements they checked, what remains uncertain and what they would escalate.
- Change one assumption and ask them to explain whether their conclusion changes.
You are looking for accounting understanding, traceable work and good judgement. A candidate who spots a gap and asks for evidence may be stronger than one who finishes quickly and misses it.
What Around Certified means for the hire
Around Certified is our name for the training approach: accounting skills, preparation for modern tools and knowledge relevant to the firm. It should lead to a specific discussion about what your chosen person already knows and what needs to be brought up to speed.
For an Irish role, that may include relevant Irish tax knowledge. A UK or US role needs its own requirements. Local rules and filing responsibilities should be agreed explicitly, with technical decisions assigned to the appropriate reviewer.
Before selection, identify the must-have experience. Before work begins, agree the remaining training, approved systems and first responsibilities. Use our guide to choosing the first work to hand over and 90-day onboarding guide to turn the brief into a manageable starting role.
Start with the responsibility you need to hand over
Bring the work you need covered, your software and the decisions the person will be allowed to make. We can discuss the accounting level and preparation the role requires.
Frequently asked questions
Should an accountant know AI before we hire them?
Where AI is part of your workflow, assess their ability to use it and check its output. The depth required depends on the role. Sound accounting knowledge and the ability to recognise errors remain central to the assessment.
Does Around Certified mean an external qualification?
Around Certified is the name for Around People’s training approach. Professional qualifications and role-specific experience should be assessed separately when choosing the person.
Will every candidate be expert in all the tools listed here?
Experience varies by person and role. Agree the systems that matter to your firm, establish the candidate’s current ability and identify the training needed before expanding their responsibilities.
Can AI remove the need for review?
AI can assist with preparation, but the work still needs an appropriate review process. Decide who checks the evidence, resolves exceptions and authorises postings or client communications.
