AI in Finance: Stop Talking, Start Experimenting

Published 19 September 2026

Over the past two years, AI has become a fixture at finance conferences and in leadership discussions. The technology has advanced fast — new platforms keep emerging and existing tools keep getting more powerful. Yet I wonder how much of this has translated into real change in our everyday work. My sense is that adoption remains low, despite all the conversation about its potential. Why?

The gap between knowing and doing

One of the biggest barriers is a lack of practical knowledge and skill. We know AI can do remarkable things, but knowing what it can do is very different from knowing how to apply it to your own work. Adoption also varies widely: some teams are actively experimenting and finding new ways of working, while others have barely started. That is a missed opportunity — without hands-on experience across a wider group of people, it is hard to move beyond isolated experiments and build AI into everyday work.

Spend less time talking, more time experimenting

Start small. Pilot AI on a practical, well-defined task rather than attempting a broad transformation.
Measure the value. Weigh the cost of the solution against the benefits — time saved, better quality, reduced effort, or other measurable outcomes.
Scale what works. Where a pilot shows clear value, expand it and build it into existing workflows.
Learn and share. Capture what works, what doesn't, and the lessons, so teams build on each other's experience.
Experiment responsibly. Put sensible controls around data, confidentiality, accuracy and human oversight — without creating barriers that discourage people from trying.

Where it's already working in finance

I have already seen plenty of practical applications emerging across finance:

Data analysis — interrogating large datasets to surface trends, anomalies and relationships, faster and in more depth.
Smart narratives — drafting commentary and analysis of financial results.
Reconciliations — matching data between systems and accounts, identifying differences and supporting journal preparation.
Discovery — searching large volumes of information to quickly find what's relevant, along with patterns and issues.
AFS review — supporting the review of annual financial statements, including consistency and disclosure checks, and preparing cash flow workings.
Working papers — preparing and updating working papers and supporting documentation.

Use AI as a super-assistant

AI output is not perfect. But used responsibly, it lets finance teams do more with less, work faster and extend their capacity. Appropriate review and validation remain essential to keep the accuracy, consistency and quality of any analysis or reporting.

Stop talking. Start experimenting.

AI will keep evolving. The tools will get more capable and new opportunities will appear — but we won't discover them by talking about them. The real challenge for finance leaders is to create the conditions for experimentation: giving teams the access, the room and the freedom to try new tools, test new ways of working and learn from the results, without putting organisational data or decision-making at risk.

If you are exploring how AI could be applied practically within your finance environment, let's talk — or see how our Power BI consultants can help.

Related guides

← Back to Insights & Checklists