Why Your Firm Has AI Tools and Still No Extra Capacity

Most firm owners we speak with are not short of software. They are short of hours. Reconciliations still take a bookkeeper’s full attention. Client queries still land in an inbox and sit there until someone has a spare half hour. Onboarding a new client still means the same forms, the same chasing, the same manual entry into three different systems. Somewhere along the way, most firms also bought a tool that was meant to fix some of this. Many of them are still waiting to feel the difference.

This is not a niche problem. Deloitte’s 2026 State of AI in the Enterprise report found that Australian organisations are behind their global counterparts when it comes to turning AI adoption into actual operating change. Only 12 per cent of Australian leaders said generative AI was already transforming their business or industry, against 25 per cent globally. A third are still using it only to automate an existing step or two, not to change how the work gets done. The tool went in. The way the firm operates stayed the same.

The pattern holds more broadly across finance functions. Recent 2026 industry data shows that while a majority of finance teams now have some form of AI deployed, only a small minority can point to a clear, measurable return on it. The gap is not really about the technology. It is about what the technology was pointed at, and whether anyone measured the result before calling it a win.

Why adoption and results keep coming apart

Three patterns show up again and again in firms that bought AI tools and got little for it.

The tool was bought to look current, not to solve a named problem. Someone saw a competitor mention AI, or a vendor made a persuasive pitch, and a subscription was approved. No one had first written down which task was too slow, how much it was costing, or what “fixed” would look like. Without that, there is nothing to measure against, so nothing gets measured.

The tool sat beside the existing process instead of replacing part of it. A chatbot answers general questions but the team still manually drafts every real client reply. A document tool extracts data but someone still re-keys it into the practice management system by hand. The old process never actually shortened. The new tool became one more thing to check.

Nobody owned the follow-through. Software implementation is usually treated as a one-off event: buy it, switch it on, move on. But a system that touches client data, drafts client communication, or triggers work inside a practice needs review, correction and a named person accountable for it. Firms that skip this step get a tool that quietly drifts out of use within a few months, because nobody was minding it.

None of this is an argument against using AI in a firm. It is an argument for being far more deliberate about where it goes and how its result gets checked.

What we would ask before the next purchase

Before adding any AI tool to a firm, it is worth answering four questions honestly.

What specific task does this replace, and how many hours a week does that task currently cost? If the answer is vague, the project is not ready to start.

Does it run on the systems the practice already uses, or does it require staff to learn and maintain a separate platform? Every extra system is an extra point of failure and an extra thing someone has to remember to check.

Who reviews its output, and how often? Client-facing work, in particular, should never run unsupervised. A person checks it. That is not a limitation of the technology; it is how the firm protects its name.

What would “working” look like in ninety days, in terms a partner could see in the numbers? If nobody can answer that before switching it on, nobody will be able to answer it afterward either.

The larger point

The firms getting genuine capacity back are not the ones that adopted the most tools. They are the ones that picked one real bottleneck, built something that actually took the manual step out of it, put a person in charge of watching it, and only then moved on to the next one. It is a slower way to work. It is also the only way that has produced results we can actually point to and stand behind.

If your firm has already bought the tools and is still waiting to see less admin land on your team’s desk, the honest starting point is not another purchase. It is a plain look at where the routine work actually goes each week, and an honest answer about whether the systems in place are carrying any of it.

Sources

• Deloitte Australia, “Australian organisations lag global peers in realising AI’s transformational potential” (2026), deloitte.com/au
• Accounting Times, “Aussie organisations falling behind on AI adoption: Deloitte” (2026), accountingtimes.com.au
• Forbes, “AI ROI Measurement: New Metrics For 2026 Financial Returns” (2026), forbes.com