AI in Finance: Faster Numbers, Less Trust. James Solomons on Automation and the Accountant's Edge
AI is making finance work faster, but faster numbers are not the same as trusted numbers. In this episode of The Practical CFO, Lisa Callaghan speaks with chartered accountant James Solomons about where automation genuinely helps finance teams, where it introduces a trust gap, and why judgement, review and accountability matter more as the tools get quicker.
Key takeaways
- AI speeds up the numbers, but speed without trust is a risk, not a win.
- Automation handles the mechanical work; judgement and review still sit with people.
- Clean data and sound process decide whether AI output can be trusted.
- Accountants add the most value interpreting the numbers, not just producing them.
- The firms that win treat AI as a tool to raise quality, not just cut cost.
Faster numbers are not the same as trusted numbers
AI can now produce finance outputs in a fraction of the time: reconciliations, categorisation, first-draft reports. James Solomons’ point is that speed is only half the equation. A number you get faster is not automatically a number you can trust. The value of a finance function has always rested on trust: that the numbers are right, that they mean what they appear to mean, and that someone stands behind them. AI does not remove that requirement. It raises the stakes on it.
Automation handles the mechanical work; judgement stays with people
The work AI is best at is the mechanical, repeatable layer: matching, coding, drafting. That is genuinely useful, and finance teams should use it. But the interpretation (what the numbers say about the business, what decision they point to, where something looks wrong) is still human work. Automating the mechanical layer only pays off if the judgement layer stays strong on top of it.
Clean data and sound process decide whether AI output can be trusted
AI output is only as good as what feeds it. Weak data and loose process produce fast, confident, wrong answers. Solomons’ view is that the fundamentals of clean data, clear controls and a real review step matter more in an AI-assisted workflow, not less, because errors move faster and look more polished.
The accountant’s edge is interpretation, not production
As producing the numbers gets cheaper, the value shifts to interpreting them. The accountants and finance leaders who stay valuable are the ones who explain what the numbers mean, advise on the decision, and take accountability for the answer. AI changes how the work gets done, not the need for someone to stand behind it.
Short clips
Get the fortnightly brief and every new interview from The Practical CFO.
Subscribe to The Practical CFO