Artificial intelligence is beginning to emerge across the accounting stack. Open the practice manager, the ledger or the document platform and a faux virtual avatar is waiting, and the chatter promises more. A demonstration will show capable features. What it cannot show is how your firm’s capacity will change.
Speaking with principals and practice managers as this arrives, one thing is evident: the real question is still the business they’re trying to build. The best firms set the direction from the outcomes they seek — where professional time should move, and the recurring work holding it back. The technology follows that decision.
The ambition comes first
You want the firm to grow without it becoming harder to run — less of your own time caught in compliance production and review, more of the practice working in advice, where client relationships deepen and the commercial value is stronger. And you’re building an asset: more useful to clients now, more valuable to a future buyer.
Those aims give the technology work its order. Start with the tools and almost any feature can acquire a business case after the fact. Start with your firm’s time, capacity and margin, and the shortlist becomes more disciplined. Some recurring work is a strong candidate. Some needs to be redesigned first. Some should remain with people.
The constraint sitting beside the ambition
The profession is also short of people. This is a structural constraint rather than a brief dip in recruitment.
Jobs and Skills Australia’s Occupation Shortage List has recorded a national shortage of external auditors for five years running and of tax accountants for the past three. Enrolments in accounting degrees have halved since 2018. CA ANZ notes that enrolments have started to stabilise, although demand is still running well ahead of supply. Enrolments in the accounting Professional Year Program, one pathway into the profession rather than the whole graduate pipeline, fell from 7,122 in 2018 to 340 in 2024. The program concluded in May 2026. The Future Skills Organisation, in figures relayed by CA ANZ, estimates a shortfall of around 6,000 accountants by 2030.
Hiring alone is unlikely to close a gap of that size. Firms will still recruit, but recruitment cannot carry the full growth plan when the available pool is shrinking.
Advisory growth therefore has to draw capacity from inside your existing operation as well as from new hires. Recurring preparation, reconciliation and document work are the obvious places to examine. When that work can run reliably under your firm’s controls, professional time can move towards interpretation, client conversations and judgement.
Why most efforts stall
Most accounting firms have tried AI in some form. Far fewer have made it part of a recurring process that runs whether or not its original champion is available.
A typical pilot works for a fortnight on one person’s desk. Client work gets busy, the owner stops tending it and the process fades from use. The tool takes the blame, although the missing elements were operational: a named owner, a defined workflow, controls, support and a way to keep the process current.
Treating AI as a tool decision leaves those elements unresolved. The practice manager, ledger and document store can each add useful features, while the client’s work continues to cross all three. Someone still has to operate the sequence between them. That is where systems, agents and workflows have to be joined and kept running.
How deliberate automation works
The work starts with a roadmap built from your firm’s own processes. It records where time and money are being lost, which workflows carry enough value to change and which should be left alone. A good roadmap is selective because implementation capacity is finite too. That selective roadmap is what our AI Strategy and Roadmap engagement builds.
From there, choose an accountable workflow deliberately: a compliance cycle, a reconciliation or an onboarding sequence where the result matters. Enter through a bounded component, prove the run and its controls, and expand only when the evidence from that component supports the next step.
The activity-statement cycle shows the difference. Every practice knows the rhythm — quarterly BAS, monthly IAS for the larger clients, the same run repeating: confirm the reconciliation is settled, prepare the statement and its documents, chase what is missing, review, lodge. A governed workflow can test readiness before the work starts, prepare the statements and queue exceptions for an accountant, while peer review and lodgement stay with the people who carry the obligation. The recurring preparation no longer has to be rebuilt by hand each cycle — and the month it used to swallow comes back to the practice.
Someone must then operate the workflow. Integrations change, source formats move and new exceptions appear. Monitoring, maintenance and governance are continuing responsibilities, not a final implementation step. Without them, the same drift that ended the desk pilot will eventually end the larger workflow. That continuing operation is what our platform is for: VonBolt Workforce runs and governs these workflows at scale — monitored, metered and reported — with the firm’s controls bound to every run.
Professional accountability remains with your firm throughout. People decide which work runs, review exceptions and approve material outputs. The run record shows what the system read, what it produced and who signed off. This is how real work can run to the firm’s standard without handing professional judgement to software.
The controls to insist on
Accuracy is the first concern. An automated output needs to be traceable to its sources and held for human review at the material points. The run record should show what happened, which figures were produced and who approved them. If the workflow cannot create that record as it runs, it is not ready to carry accountable work.
The same operating standard should cover the wider team. Many firms use offshore staff as part of a sound delivery model. Their location is secondary to the control question: does the work follow one standard, with clear accountability to your firm? Software, onshore staff and offshore staff all need to sit inside that operating discipline.
The choice in front of your firm
AI features will continue to spread through the systems accounting firms already use. Access to those features will not solve the capacity problem on its own. The practical difference will come from whether your firm has chosen useful workflows, built the necessary controls and learned how to operate them.
That work can begin without betting the firm on an end-to-end transformation. Start with one accountable workflow and a bounded entry. Measure what happens to time, exceptions, cost and control. Expand when the result supports it.
If your firm is deciding where to begin, start with the capacity you want to create and the recurring work currently consuming it. If you want help turning that into a plan your firm owns, get in touch.
Damien Zeller is the Principal of VonBolt Technology Group.
