While generative artificial intelligence is a hot topic of conversation these days, we should not forget a long and successful history of using non-generative AI, sometimes called legacy AI, especially for numerical and structured data. . Applications such as forecasting customer demand or revenues or detecting patterns such as fraud or money laundering are prime examples for CFOs and accountants.
These tools and use cases improve their capabilities every year and provide tangible business value.
Uses legacy AI.
These non-creative AI systems can also provide significant support in meeting compliance and regulatory requirements and generating analytical reports for these purposes. Detecting which receipts and payments belong together, especially in cases of partial discrepancies, is almost universally used today and relies on AI.
Many more sophisticated management dashboards and systems, including both accounting and enterprise resource planning software, ultimately rely on such AI systems, for example inventory management and planning. Complex processes such as just-in-time or just-in-order, cannot function without a legacy AI backbone.
Limitations of Generative AI
Turning to the oft-discussed topic of generative AI, we admit that many of the claims are hype. For example, any tool has a desired scope of use for which it is helpful and provides value. Outside of this scope, it is not helpful and can cause harm. Large language models are intended to manipulate language, not numbers, and so are generally not successful in dealing with numbers where we expect absolute accuracy.
A case in point is the analysis of a company’s annual report. If we do this using LLMs, we may get answers that are “inflated” by information outside of the report, or we may get numbers that are not in the report. Such usages are not fair and misleading. So what can we use them for?
Multimodal Uses of Generative AI
A step change ahead of generative AI is its multimodal facility – the ability to work with text and images simultaneously. Imagine taking a mobile phone snapshot of your latest restaurant bill and having it automatically submitted to your company’s travel expense form. What a time and hassle saver! It is very accurate and thus prevents human error as well. The same goes for invoices, receipts and other paper forms.
If a legacy AI model detects an error—such as a fraudulent or partially paid invoice—it’s Creative AI that can turn that discovery into a human-readable message that explains What is happening and what to do about it. We’ve talked about explainable AI for years, and it’s LLMs that can offer explanation even if the content of that explanation requires other systems to weigh.
Natural language dashboards
We’ve all been in board meetings where one person asks an analytical question that doesn’t have the right numbers. Oh terrible. An analyst would have to be kept busy for a few days, the charts would be sent, and the result would not be actionable for a long time. Those days are gone! Generative AI can translate a query from English to the database, SQL language, and get the resulting table of numbers. This table is then translated into the codified language of the dashboards and displayed to the human user as a graphical image.
It all happens in the blink of an eye. Most importantly, the result is not misled by LLM but comes directly from the database—the answer can be trusted. This allows more questions to be asked directly in the board meeting, ultimately achieving actionable results in a shorter amount of time. I was in a meeting where eight point questions were asked and answered in less than 10 minutes, leading to new insights and board decisions. It was an eye opener.
Support services
Responding to inquiries from employees, customers and suppliers is a major pressure on any accounting department. Generative AI can help by trying the most common questions and automatically providing correct and intelligent answers. From helping with dreaded expense reports to filing invoices, AI can largely automate the day-to-day accounting process, including matching it to the right expense account and getting approval.
Security is important, especially when money is involved. Generative AI provides a new level of sophistication to detect various types of attacks such as phishing and hacking.
Some of the uses where AI, creative or not, can help in the accounting realm are listed here. In addition to managing the company’s finances, the CFO also has to make many decisions for the rest of the company. AI can help analyze scenarios, find referral data, and contextualize the conditions and offerings of competitors or other vendors. This can help visualize and compare the benefits of multiple options so that the CFO can better decide which one to choose.
Finally, generative AI delivers real business value to the CFO organization after all the hype is cut through. Most impressive is the creation of dashboards based on human language queries. If you do nothing else, give it a good look.