What accountancy practices should fix before they automate

Most accountancy practices do not have an automation problem. They have an operational readiness problem.

The appetite is not the issue. The profession is willing, and in many firms the younger staff are already using AI weekly. Yet most practices have not moved beyond experimentation. Technology plays a part, but the reasons are mostly about operational readiness: process, data, governance, skills and confidence.

So the question a managing partner should be asking is not “what software should we buy?” It is “is the practice actually ready to automate?” A practice that automates before it is ready does not save time. It industrialises its existing mess.

This article sets out what to fix first, and a simple way to check whether you are ready before you spend anything.

The evidence

The willingness is well evidenced. In a survey of 2,718 Chartered Accountants across 48 countries, 85% said they are willing to use AI, given the opportunity, and 83% of those aged 18-24 already use AI tools weekly (Chartered Accountants Worldwide and Ipsos, April 2025).

The gap between willingness and action is just as clear. Sage's Practice of Now 2024-2025, a vendor-sponsored survey of 1,000 accountants and bookkeepers, found 89% believe automating processes would free up time, yet only 37% are doing it.

The most-cited barriers point to readiness rather than the technology itself. Chartered Accountants most commonly cite data security (33%), lack of training (28%) and company policies that prevent access (14%). Among C-suite executives, 53% say they do not feel prepared for AI's impact on their role over the next five years.

There is consistent evidence, beyond accountancy, that process standardisation is a precondition for successful automation. Peer-reviewed research identifies it as a critical success factor for robotic process automation (Ge et al., Scientific Reports, 2025), and a systematic literature review reaches the same conclusion for digital transformation more broadly (Goel, Bandara and Gable, 2023). ICAEW guidance stresses governance and professional scepticism when using AI, which reinforces the same point: the groundwork of process, data and governance comes before the tool.

The Failsafe interpretation

Read together, the evidence describes readiness, not technology. Our reading is that buying software first is the wrong first move.

Failsafe works in one order, always: Business, then Operations, then Technology, then AI. We built our method this way because technology is rarely the first problem. Before a practice automates anything, it should understand where time is really lost, how the work actually flows, who owns each process, whether the data can be trusted, and what its existing systems can already do.

One of our principles is that you never automate a broken process. Automating a broken process only makes it fail faster, and at scale. A repetitive task is not automatically a good automation candidate. Sometimes the right answer is a better process. Sometimes it is connecting two systems you already own. Sometimes it is automation, sometimes AI, and sometimes it is nothing new at all, because the tools already in the building cover the gap once the process around them is fixed.

This is not caution for its own sake, and it is not anti-technology. It is the opposite. The practices that will get the most from automation and AI are the ones that did the operational work first.

The Failsafe Pre-Automation Readiness Assessment

We've turned that principle into a practical six-question test that a practice can apply before automating a process. Answer the six questions in order. If you cannot answer one with confidence, that is where the work is, and automation waits until you can.

The order matters. It is the Business, Operations, Technology, AI sequence made practical: the first four questions are about the business and its operations; only then do the last two turn to technology and to where a person must stay accountable. The questions are industry-neutral; the examples change from accountancy to real estate to logistics, but the six questions do not.

Where the Business Performance Review fits

These six questions are one lens. A Business Performance Review applies them within a much broader examination of the practice: its objectives, its people, how work actually flows, the systems it runs, the quality of its data, and the controls and governance around it. The readiness questions sharpen a single automation decision; the wider review sets the objectives, weighs the trade-offs across the whole practice, and produces a prioritised picture of what to fix, in what order, before any spend on change.

Explore the Business Performance Review

Next steps

If you are weighing up automation or AI in your practice, start here, not with a shortlist of tools:

• Map one process before you automate it.
• Check that it is standardised and owned.
• Confirm the data behind it can be trusted.
• Ask what your current systems already do.

If you can confidently answer all six questions for the processes that matter most, you are ready to choose technology well. If you cannot, that is exactly what a Business Performance Review is designed to uncover.

References