The big picture: ChatGPT is among numerous large language models that will transform industries, corporate functions, the global economy and other realms. These models also pose something of an innovation litmus test for finance leaders.
- Many CFOs remain wary of GenAI applications in, and implications for, the finance group and throughout the rest of the organization.
A key point: Leading finance groups already are putting GenAI tools through their paces to improve cash flow management, FP&A, liquidity risk management, fraud detection, workflow efficiency, scenario planning and more.
- GenAI adds an entirely new layer—a narrative that explains the why behind financial analyses—to these existing capabilities.
The bottom line: There are tangible benefits associated with being a smart, early adopter when it comes to the transformational opportunities GenAI offers finance functions.
Go deeper.
ChatGPT is among numerous large language models (LLMs) that will transform industries, corporate functions, the global economy and other realms. These models also pose something of an innovation litmus test for finance leaders.
As we’ve all learned over recent months, generative artificial intelligence (GenAI) technologies create text, images and video based on queries and requests, leveraging datasets and algorithms. With ChatGPT leading the way (for the moment), they now feature conversational, user-friendly interfaces that make them more accessible and understandable than their predecessors from years past. (My colleague, Christine Livingston, provides an informative rundown of GenAI here).
Many CFOs remain wary of GenAI applications in, and implications for, the finance group and throughout the rest of the organization. While the investigation by the Federal Trade Commission (FTC) of ChatGPT maker OpenAI confirms the need for robust AI governance guardrails, CFOs should nevertheless look into how LLMs can benefit the finance group and the company as a whole, in both the near term and long term.
Make no mistake—despite the need for caution and governance, the clock is ticking. “Clear leaders and laggards are already emerging,” according to The Economist, which reports that only about 70 companies in the S&P 500 have shown no signs of investing in AI yet. Research conducted by Protiviti shows that nearly half of organizations worldwide currently are using artificial intelligence and machine learning, with another 27 percent planning to implement these technologies within the next three years.
Those investments are projected to pay off handsomely. Research from Goldman Sachs forecasts that improvements related to GenAI could deliver a 7% boost in global GDP (nearly $7 trillion) while increasing productivity growth by 1.5% over the next decade or so. Three-quarters of the value that GenAI is poised to create inside companies will occur within the customer service, sales and marketing, software engineering, and R&D groups, according to McKinsey.
Leading finance groups already are putting GenAI tools through their paces to improve cash flow management, FP&A, liquidity risk management, fraud detection, workflow efficiency, scenario planning and more. While recent strides in advanced analytics and machine learning have sharpened financial forecasts and equipped finance groups with greater predictive proficiency, GenAI adds an entirely new layer—a narrative that explains the why behind financial analyses—to these existing capabilities.
For some time, finance organizations have leveraged automation to produce deeper analyses. Yet those outputs did not come with automated narrative explanations. That meant that finance professionals had to grab their shovels and start digging up the story underneath the numbers. This is where GenAI can change the game. It articulates the rationale and implications of financial analyses by producing meaningful narratives about various predictions and their implications. These summaries can be generated in different formats: executive summaries, bullet points, slide decks, a four-paragraph memo or even iambic pentameter (if the finance team is so inclined).
While the options for outputs may be a bit staggering when comprehended for the first time, it is important to keep in mind that GenAI’s more basic productivity enhancements—writing emails, producing meeting notes, creating outlines—take a back seat to the deeper insights the technology adds to existing forecasts, analytics and scenario planning. For example, BloombergGPT—an LLM trained on the financial documents, terminology, trends and data that Bloomberg has collected for more than 40 years—will create entirely new workflows, economic analyses and financial benchmarks for its customers, the company says. Many FP&A platforms are unveiling GenAI features that produce automated narratives supporting their analytical insights.
Further, GenAI and technology companies are developing new partnerships designed to help organizations bring LLMs “inside their four walls” with GenAI capabilities that offer stouter data security and privacy controls. This is an especially compelling development because the approach will help organizations avoid data security missteps, such as uploading proprietary data and/or sensitive code to publicly available LLMs.
There is great promise in the finance organization’s use of GenAI. But as everyone knows by now, there also are risks. Harnessing GenAI’s benefits in a secure, reliable and unbiased manner requires a number of foundational actions, including the following:
- Creating a cross-functional team: Regardless of whether finance, sales and marketing, R&D, or another group is evaluating a GenAI solution, it helps to examine the technology through different lenses. Domain experts are crucial, of course, i.e., finance experts should spearhead GenAI solutions in the finance group. But a technology lens is also important: The team needs a technologist who understands how LLMs operate and how the models have been trained (so as not to assume either quality or accuracy as an outcome, given that the LLMs are only as good or as accurate as they have been trained to be). Cross-functional teams also require expertise in data security, data privacy, risk management, and legal and compliance.
- Recognizing and mitigating the risks: GenAI applications can produce “hallucinations” resulting from factual mistakes in source materials. Remember, these technology tools do not distinguish fact from fiction—rather, they produce the most likely response in a well-presented manner based on available data. Any content generated with AI must be evaluated to ensure it is factual. Thus, supervision and review apply to GenAI content just as they would to content generated by a finance function employee. Likewise, model hallucinations should be investigated to determine reasons why they are occurring. Other risks include data privacy, veracity and authenticity, ownership/intellectual property, and compliance with applicable laws and regulations. These types of risks must be identified and mitigated with human oversight, quality control within the organization, and coordination with the compliance and general counsel functions.
- Assessing the benefits: GenAI’s ability to articulate the why behind financial analyses offers many applications that finance leaders should explore. Summaries of FP&A outputs can be generated for procurement teams, who can then make faster, more precise adjustments to fluctuations in customer demand, raw material prices and other variables. AI-generated summaries of customer payment patterns might signal the need for crucial cash flow management refinements. Fraud-prevention narratives can uncover previously neglected AML compliance risks. Capital allocation strategies, investment planning and scenario planning also can benefit from GenAI.
- Addressing talent needs: As organizations and finance groups customize GenAI solutions to meet their unique requirements, new skills will be needed. In many cases, AI expertise can be upskilled. Hiring teams of data scientists with PhDs may not be necessary; industry knowledge and domain expertise will be vital, however.
“Hesitant” best describes the vibe among the finance leaders who attended a session devoted to AI during a recent CFO Leadership Council conference. While that mindset may reflect the CFO’s traditional response to cutting-edge technology, it requires a quick upgrade to becoming “AI curious.” There are tangible benefits associated with being a smart, early adopter, especially when it comes to the transformational opportunities GenAI offers. At the very least, finance organizations and companies just beginning can start with simple applications to create familiarity.
This article originally appeared on Forbes CFO Network.
Add comment