It can be difficult to decide where to start with AI when all the advice points in the same direction. Something that saves one person hours can leave someone else wondering what all the fuss is about and how AI could possibly work in their business.
Those experiences can actually tell us something useful. AI isn’t a one size fits all. In fact, AI’s value depends on the business, the task and the person using it. The question for a founder is where it could make a worthwhile difference in their own business.
Start with the work you already do
In his conversation with Kate on the Scale Up Smarter with Artemis Clarke podcast, Charles Radclyffe made a practical point about AI projects: starting with an exciting idea before identifying a useful job for it can lead you in the wrong direction.
His approach is to look at the work already happening and consider where AI could help. That gives you something specific to assess.
For a finance team, a useful starting question might be why the monthly reporting pack takes so long to prepare, or why the same information has to be entered more than once. One of our Fractional Finance Directors, Anna, talks about how AI can be useful for finance teams in the episode: Has your business outgrown its finance function?
Look at the business as well as the tool
Before choosing an AI tool, consider what the business needs to improve. Is the priority getting information sooner, reducing repetitive work or giving the finance team more time to investigate performance?
Then look at what’s causing the problem. If people are chasing missing information or following an unclear process, those issues need attention as part of any change.
For example, if preparing a cash forecast takes too long, it helps to identify where that time actually goes. Collecting bank balances, chasing updates and deciding when customers are likely to pay are very different jobs. Each needs a different approach.
Choose a task you can assess
A broad aim such as ‘using more AI’ is difficult to evaluate. A defined task gives you a way to decide whether the change is helping.
Choose a manageable piece of work and describe what a better result would look like. That could be a report that is quicker to prepare, fewer corrections or information that reaches the decision maker sooner.
Include the time needed to prepare inputs, check outputs and correct mistakes. Compare the total effort and cost with the current process, as well as the quality of the result.
This also helps distinguish where AI is useful from where a simpler process, a connection between existing systems or a clear automation rule would do the job perfectly well.
Involve the person doing the work
The person who carries out a task will often know where the awkward steps and exceptions sit. Involve them when choosing what to try as well as when assessing the result.
They also need time to learn the tool, understand its limits and work out how it fits into their day. Someone checking a draft explanation needs enough knowledge to spot an incorrect assumption or a number that does not make sense.
Agree who reviews the output and what happens when it is wrong. In finance, saving time only helps if the resulting information is useful enough to support the decisions being made.
Build an answer for your own business
Different opinions on AI are a reason to ask more specific questions. Which task was tested? What information was available? Who used the output? How was the benefit assessed?
Those details help you decide whether someone else’s experience is relevant to yours. Start with one useful task, review what happens and use that evidence to decide the next step.
A practical place to begin
The Finance Engineers helps growing businesses understand what they need from finance, identify what is getting in the way and decide where process improvements, technology or additional skills could help.
If you are wondering where AI could be useful in your finance function, visit The Finance Engineers page and tell us what you would like to improve. We can help you work out a practical starting point.
