How are CFOs identifying the most suitable areas within finance for automation and AI implementation?
It’s important to have clear objectives in mind first. I think the best use cases are those that have a strategic focus from the start. For example, asking “how will evolving a given repeatable workflow or analysis with AI create a compelling advantage for my business and ultimately return incremental shareholder value?” An obvious starting point is reducing costs, and it’s not simply people costs. All too often there is a quick reaction to suggest that GenAI will create further automation that will simply replace all finance management functions, and that’s simply not true. Rather, adopting AI intelligently into an existing strategic workflow can yield incremental shareholder value.
Within our focus in Corporate Treasury, we are seeing AI drive more insightful and confident cash flow forecast analysis, which in turn means CFOs can invest excess cash for longer durations and pick up incremental yield. Or conversely, for regions or organizations who rely on short-term borrowings to fund working capital, a more accurate short-term forecast powered by AI allows CFOs to borrow from lower cost debt instruments as opposed to convenient and expensive overnight debt facilities. These benefits have real and significant P&L interest income and / or interest expenses benefits that can be well into the millions of dollars, especially in the higher interest rate market we operate today.
What are the biggest challenges CFOs face when implementing automation and AI solutions?
Accuracy, confidence and cost. With the benefits of automation, there runs a risk of decision-making control. This is compounded when you consider AI is based on datasets, and the risk of quality data. Thus, if the datasets being incorporated into the AI and ML output are not complete, or contain errors, then of course the output may be flawed. GenAI and co-pilot recommendations are not infallible, thus ensuring the CFO has validation approval control is very important, particularly in the early adoption phases. Another challenge is cost and talent. Large data sets require more hosting and computing throughput, and this comes at a cost both in the form of technology and also the correct talent to build, interpret and tweak the models.
Are there specific AI applications, like chatbots, that CFOs are finding particularly valuable?
Yes. Like many aspects of professional and personal technical experience, individuals expect answers to very specific questions quickly and accurately. Chatbots, including integration to commonly used messaging services like Microsoft Teams and Slack, have been adopted by CFO teams to quickly ask specific questions and return accurate answers quickly and easily. For example, “What was my LATAM free cash flow in Q2 2023?” or “What is my bank balance in Indonesia?” are now questions that can be answered in seconds leveraging APIs and APIs into leading FinTechs from messaging services or bespoke built chatbots.
Looking ahead, what are some emerging AI applications that you are excited about for the future of finance?
I am particularly excited about how GenAI will augment many of the daily repeatable workflows in the office of CFO. Whether this is preparing journal entries, investigating bank reconciliation discrepancies or preparing insightful analysis of future FX hedging strategies. I think the future is very exciting with the many applications of AI within the office of the CFO to perform routine functions more quickly, accurately and ultimately at a lower cost.