Let’s face it. Time is the scarcest resource for today’s finance leaders. Work is getting more complex, expectations are rising, and yet the number of hours in the day hasn’t changed.
Pressure is mounting on finance teams to do more with less. Smart automation and AI can help. Not to replace your team, but to unlock the time and insight you need to lead more effectively.
But despite all the noise, many finance leaders are still left wondering where the real impact is. Everyone’s talking about AI, yet your team’s day-to-day reality hasn’t changed much. So what do you actually use AI for and what do you keep human?
Weighing up the risk
A post I read recently from The Secret CFO summed it up well. AI is right 90%+ of the time, but it gets things wrong the other 10% and it does it with total confidence. Now ask yourself: would you accept even 2% confidently incorrect work from your best finance person? Absolutely not.
That’s why this blog isn’t about handing over critical decisions or sensitive data to AI. It’s about something far more practical. Using AI to improve how your team works day-to-day. Processes and productivity don’t have to be perfect. They just need to get better. That’s where the real opportunity lies.
Start with the right problems
One of the most effective ways to make progress with AI is by focusing on the right kind of problem.
There’s a useful distinction between two types of challenges:
- “What problems” are high-risk, high-reward. These are where we delegate our work to AI. These include tasks like reimagining forecasting, fraud detection, expense management and compliance. AI has the potential to truly revolutionise these areas, but in many cases, the technology, data infrastructure, or organisational readiness simply isn’t there yet.
- “How problems” are lower risk and highly practical. These are opportunities to upskill ourselves with AI – to make ourselves better at doing the work, vs delegating the work. These focus on making current processes faster, cleaner, and more efficient, delivering value immediately without requiring a major transformation.
Most of the real, high-impact AI opportunities in finance in 2025 fall into the “how” category. Whether it’s automating tasks, reducing rework and errors, or simplifying routine communication, this is where AI can help right now. Identify the small, strategic improvements that unlock capacity and build momentum.
Here are some ‘how’ problems you can solve
Once you have identified a few “how” problems in your team, such as repetitive, manual or inefficient processes that slow you down, the next step is to test where AI can help you work smarter.
Here are a few real examples I have explored with finance teams recently, along with the prompts we used:
1. Root cause analysis of journal errors
In one test, I asked ChatGPT to help a finance team that was correcting the same journal errors month after month.
Prompt used: “Act as a market-leading transformation consultant advising a finance manager. Our teams are correcting the same mistakes in journals each month, rather than fixing them at the source. What could we do to improve this? What free or low-cost tools could help?”
ChatGPT responded with a structured root cause analysis framework, including error lifecycle mapping, process redesign, and governance recommendations like regular error reviews. It enabled the team to move from firefighting to identifying and eliminating the root causes of recurring issues.
2. Analysing high-cost spend
For teams managing high-cost categories like software, external suppliers, or specialist equipment, it can be useful to find out where costs can be saved.
Prompt used: “We need to deliver major cost savings while protecting business performance. Please give us a framework to deliver this work in the next two weeks, and potential ideas to hit the target.”
ChatGPT recommended building dashboards, using vendor scorecards, and creating structured feedback loops between finance and procurement. This helped move the process from reactive cost tracking to proactive cost control, enabling earlier interventions and better spend decisions.
3. Reducing time spent on data analysis
One team wanted to reduce the time spent preparing monthly outputs, much of which was tied up in manual data tasks.
Prompt used: “How could a team reduce the time it takes to dig into data and produce outputs, focusing on root cause analysis and identifying key areas that need attention? What low-cost or no-cost tools could be implemented?”
ChatGPT suggested quick wins like using drop-downs, standardised templates, and validation rules to cut down on errors and rework. These changes helped the team streamline their workflow and free up more time for value-added analysis.
Focus your time where it counts
One of the frameworks I often use with clients is the 80/20 rule. 80 percent of your results come from 20 percent of your effort. The same principle applies to how your team should spend their time.
If most of your day is absorbed by low-impact, repetitive work, you’ve got less bandwidth to focus on the important task. But when you can speed up, simplify or completely offload those routine processes, you free up time for more meaningful, higher-value work.
The goal isn’t perfection. It’s leverage. AI can handle the routine, rules-based work that eats into your team’s time, so they can concentrate on the areas where judgment, insight and leadership are essential, these are the ‘how’ tasks.
Start small and strategic
Small, targeted steps can free up hours, reduce friction, and build serious momentum without a big transformation plan. The teams doing this well are already seeing results. 84 percent of self-described AI experts save more than four hours a week compared to just 24 percent of novices. Over a year, that’s an extra month of capacity per team member.
This isn’t about replacing humans. It’s about giving them the tools to do their best work.
You don’t need perfect systems or fully integrated tools to start seeing the benefits. Some of the biggest wins come from looking closely at how work gets done. The manual fixes, duplicated effort, or clunky handoffs that quietly drain time every week.
Think about vendor queries. If your inbox is full of similar supplier questions, AI tools can help draft standard replies or triage requests based on topic or urgency. That frees up your time while still keeping response times sharp.
Know when to stay human
Of course, not everything should be automated. There are areas where your experience, intuition and leadership are irreplaceable, and always will be, as well as the ‘what’ tasks that AI can’t handle yet.
AI might provide useful inputs, but it can’t weigh up competing priorities, understand political nuance, or navigate the trade-offs of a strategic decision. That’s where human judgment matters most.
Similarly, when it comes to building trust with stakeholders, whether it’s your CEO, the board, or your team, relationships still rule. Influence is built on human connection, empathy and credibility, not automation.
Ethical oversight is another area that should always stay human. Just because something can be automated doesn’t mean it should be. Finance leaders often serve as the ethical compass of the organisation, and that’s not a responsibility you can outsource to an algorithm.
And when it comes to leading people, supporting a colleague through change, helping someone bounce back from a setback, or shaping team culture, there’s no substitute for showing up as a leader.
Progress over perfection
You don’t need a 12-month automation strategy or a new tech stack to get started. Just pick one process that consistently slows your team down, test an idea, and build from there.
Start small. Trial one improvement. Track the impact. Then repeat.
Each win creates momentum. Each step frees up capacity. And over time, your team becomes faster, sharper and more focused on what truly matters.
Ready to take back your time?
If you could free up four hours this week, where would you spend it?
Would you use it to think more strategically? Lead your team through change? Tackle that project you’ve been putting off?
Wherever your answer lies, AI can help you get there. Not by replacing your work, but by clearing the way for it.
And if you’re not sure where to begin, get in touch. I’d be happy to help you find the first step.
