Ahmed Ezzat, ACMA, CGMA, is rather busy.
Heโs the CFO of a venture capital and private-equity company in Saudi Arabia. The company manages a significant portfolio of investments across various sectors with a streamlined team. Ezzatโs role includes participation on multiple startup boards, helping guide their strategic direction. Part of his work involves evaluating numerous investment applications each year, leveraging advanced technology to enhance the assessment process and ensure a comprehensive analysis of potential investments.
โManaging extensive data is challenging, and we continuously refine our processes to capitalize on potential opportunities effectively,โ he explained in a recent interview with FM. โThe investment decision is critical, especially for a tech company in the early stage.โ
In recent years, as part of his companyโs tech-forward approach, Ezzat has been using generative AI to augment his reviews of companies seeking investment โ looking for the strategic strengths and weaknesses in scores of startups.
Itโs part of a larger trend: Finance leaders, executives, and board members are trying to use generative AI and machinelearning tools to augment strategic decision-making, and technology companies are racing to help them.
Some companies have even appointed AI systems as virtual advisers to their boards, with the promise of surfacing hidden trends and steering a companyโs financial future.
But many individual finance leaders, including Ezzat, have started with much smaller experiments. Theyโre finding ways that emerging technologies can expand the depth and breadth of an executiveโs decision-making process โ and theyโre learning to work around the technologyโs shortcomings.
In a series of interviews for FM, Ezzat and other finance leaders shared their early impressions of how AI could support strategy and leadership.
โIf you are not learning yet about ChatGPT and AI tools, how [they] work, and how to integrate [them] with your system, I think itโs almost too late,โ Ezzat suggested. โIf you didnโt start yesterday, start now.โ
Starting small: How finance leaders are applying gen AI
For some finance leaders, generative AI agents have already become a personal tool for planning, analysing, and communicating clearly.
As Ezzat reviews a startupโs proposal, he might ask ChatGPT to review the basics of its business pitch. Then heโll ask the system to apply business frameworks, such as a strengths, weaknesses, opportunities, and threats (SWOT) analysis or Porterโs Five Forces, to the proposed business model.
For others looking to test generative AI tools, remember to review assurances from the tool provider regarding data privacy, security, and ownership before uploading documents or making statements that may contain confidential or sensitive information to a generative AI system.
Often, Ezzat has found the AI makes astute observations. For example, one startup was selling a financial service for consumers.
โI liked this idea,โ Ezzat said, adding that he could imagine using such a service. But the AI quickly surfaced a concern: The business model had low barriers to entry.
โChatGPT came back and said, โThis market is very easy to enter by other competitors,โ โ he said.
Indeed, when this writer gave ChatGPT a description of the same business model and asked it to run an analysis, the AI bot produced the same concerns. It also listed several other risks โ such as cultural resistance to change, interest rate sensitivity, and the risk of overextending credit to consumers โ as well as several sensible strategic recommendations.
None of the AI ideas were all that original. Ezzat, with his experience, surely would have thought of them himself. But the software agent raised worthwhile questions and successfully drew connections to numerous related concepts and ideas.
In turn, Ezzat immediately knew what questions he wanted to ask the startupโs founders. He believes AIโs role for leaders will be โenhancing their ability to process and analyse vast amounts of information and align this information with a companyโs strategic planโ.
Other finance leaders are using AI in similar ways to generate ideas, to raise questions, and to support and expand research to streamline the strategic planning process (see the sidebar, โUsing AI as a Research Assistantโ).
No matter how they use it, finance professionals should keep in mind that using generative AI-sourced information without ensuring its validity may complicate their fiduciary duties. If the softwareโs methods and inner workings are unclear, but they are relying on it to make decisions, it may raise questions about ethics and standards, warned Russell Tucker, ACMA, CGMA, group finance manager for a multi-family office in England.
Could AI have a role on the board?
While finance leaders experiment with generative AI at a personal level, some companies have started to dream much bigger. Could AI itself take on enhanced strategic roles, especially on the board?
A Hong Kong venture capital fund said an AI system had an informal adviser role for its board of directors, and a major company in the United Arab Emirates has done the same.
But some finance experts remain sceptical of the idea that AI will advance into a true leadership role anytime soon.
โI struggle to believe thereโs an all-singing, [all]-dancing AI member that can hit all professions and functions appropriately,โ said James Owen, FCMA, CGMA, global CFO, Profiles Division at Kantar.
Instead, he sees AI playing an information processing role โ with potential strengths such as tracking board action and helping board members prepare for meetings.
โEveryone at the table will have had their pre-briefing from an AI-inspired source,โ Owen predicted.
Generative AI could also become an interface for the board and other leaders to tap into a companyโs deeper data troves and machine-learning capabilities.
As the technology becomes more reliable, instead of depending on dashboards that highlight particular data series โ or waiting for quarterly updates from staff โ CFOs and board members could use natural language interfaces to simply ask an AI agent for specific analyses of large amounts of company finances, forecasts, and other data.
However, at least for now, the technology may not be as painless or powerful as itโs made out to be. For example, Microsoft has ambitious plans for its Copilot platform to tap into a companyโs datasets to generate insights and content based on plain-language inquiries.
But Owen cautions that in its early days, his companyโs implementation of Copilot has required troubleshooting and oversight. The technology struggles if files and folders are not uniformly organised, he said. it canโt always โmake the linkageโ between differently organised files and data, Owen explained. โItโs teaching us to be a lot more organised and a lot more sophisticated.โ
Itโs a symptom of a larger issue with implementing AI systems, and not just generative AI. โYou need structured data for machine learning to be plugged into. And I feel thereโs a lot of organisations still in the early part of that journey,โ Owen added.
Preparing for AI-empowered leadership
Business technology companies are steadily incorporating generative AI into software most companies rely on, from cloud accounting platforms to enterprise resource management systems.
At the same time, deeper and more effective AI capabilities are being baked into consumer devices and software.
The ultimate limits of this technology have yet to be determined. It may never reach the heights that some are promising today โ a truly autonomous system that can execute complex tasks from start to finish.
However, it seems clear that generative AI already offers potent benefits for many of the tasks most important to executives: analysing, summarising, and communicating.
โYou need to understand, not necessarily how does it work, but what can it do? What canโt it do?โ said Donny C. Shimamoto, CPA, CGMA, managing director of the CPA firm IntrapriseTechKnowlogies.
Those answers should become clearer as the initial hype around generative AI dies down, Tucker said. โWe can figure out where the limitations are and then implement the technology to reduce all the things that we do that are time wasting,โ he said.
For Ezzat, thereโs a twofold task ahead: Individual leaders need to improve their AI literacy, while organisations need to understand how generative and other types of AI can support their strategic decision-making.
โWe need to identify specific areas where AI can close the gaps in decision-making and enhance strategic forecasts,โ Ezzat said.
Within a matter of just a few years, these experts predicted, AI skills will be a key part of a leaderโs skillset โ if theyโre not already. โWe โฆ need to take on AI and implement it in a way that weโre all empowered by it and not replaced by it,โ Owen said.
Using AI as a research assistant
Many finance leaders are using consumer-grade AI products like ChatGPT as research assistants โ skimming through materials to surface key ideas. As a research tool, AI can help:
1. Quickly digest large datasets
โI think itโs really good in terms of digesting information,โ said Russell Tucker, ACMA, CGMA, group finance manager for a multi-family office in England. โFor me itโs processing articles, PDFs, books โ getting through the data, in order to speed up that process for me, because obviously, day to day, weโre really busy.โ
2. Streamline your searches
An AI query can sometimes prove more efficient than a web search. Traditional search engines lead to webpages that could be out of date, are loaded down with ads and paywalls, and may or may not answer your question.
At its best, generative AI can directly answer a question by tailoring its knowledge base to constraints provided by the user. No surfing required.
3. Clearly communicate your findings
Generative AI can also help with explaining research and ideas to others. Tucker uses it to simplify and summarise his writing.
โIโm a really technical guy. Sometimes, Iโll use it to try and bring things into a laymanโs term and help present data in a different way,โ he said.
4. Transcribe conversations
AI transcription services, such as Otter and Rev, donโt just produce written records of meetings. They also identify key insights and decisions made during meetings, which is helpful for executives who spend much of their day on Zoom or Teams.
Remember: Be cautious
Of course, these tools also have a reputation for making false statements. So, while their developers have worked to reduce โhallucinationsโ, fact-checking AI answers remains mandatory.
An overreliance on AI can lead to not just misstatements but, potentially, strategic mistakes. While it can be helpful for basic questions and initial research inquiries, itโs unlikely at present to come up with truly novel ideas, and it may struggle with situations that are not addressed in its knowledge base, said Donny C. Shimamoto, CPA, CGMA, managing director of the CPA firm IntrapriseTechKnowlogies.
Andrew Kenney is a freelance writer based in the US. To comment on this article or to suggest an idea for another article, contact Oliver Rowe at Oliver.Rowe@aicpa-cima.com.
LEARNING RESOURCES
In this course, youโll learn how to use ChatGPT to streamline your workload, save time, and boost your overall efficiency.
COURSE
AICPA & CIMA RESOURCES
Articles
โChatGPT: Use Cases and Limits to Its Reliabilityโ, FM magazine, 2 May 2024
โ4 Considerations for Finance Teams About Gen AIโ, FM magazine, 19 April 2024
โGenerative AI: Whatโs the Potential?โ, FM magazine, 12 February 2024