The following is a guest post from John Parkinson, senior director of emerging technologies at FreeClimb. Opinions are the author’s own.
I get asked a lot about whether artificial intelligence can help a busy CFO with process automation and productivity for routine tasks, such as budgeting and financial reporting. As CFOs navigate the complexities of modern finance, they are increasingly being pitched with the potential benefits of AI, and like most senior business leaders, they can’t avoid the “generative AI solves everything” hype.
With the ability to process vast amounts of data in real-time, AI can do things that no amount of people can accomplish, but how will this help CFOs make more informed decisions, streamline processes and drive business growth? There is potential here, but the path forward is not going to be without its challenges. Different types of AI, such as extractive and generative, can be leveraged to provide valuable insights, improve decision-making and support business operations and growth.
Extractive AI: Unlocking accurate and reliable insights from existing data
Extractive AI uses machine learning models that are trained to analyze existing data sets to identify patterns, trends and correlations. This type of AI is particularly useful for CFOs who need to extract insights from large datasets, such as financial reports, market research or customer behavior. Extractive AI is “deterministic,” ask it the same question multiple times and always get the same answer, provided the underlying data has not changed. It doesn’t “hallucinate” in the way that generative AI sometimes will.
Which means it can deliver multiple benefits, including:
- Improved forecasting: By analyzing large volumes of historical data quickly and accurately, extractive AI can deliver timely insights on operational performance and help CFOs make more accurate predictions about future financial returns.
- Enhanced risk management: Extractive AI can identify potential risks and opportunities by analyzing large internal and external datasets in real-time or near real-time.
- Streamlined reporting: Automated reporting tools powered by extractive AI can reduce the time spent on manual data analysis, freeing up resources for higher-value tasks.
- Knowledge capture: Extractive AI can collect and organize the internal process knowledge that only resides in the people who do the work, easing succession planning and supporting efficient working practices.
However, there are also challenges to consider:
- Data quality issues: Extractive AI is only as good as the data it’s trained on. Poor-quality or incomplete datasets can lead to inaccurate insights and inconsistent reports.
- Limited creativity: Extractive AI is designed to analyze existing patterns and trends, which may not always result in the kind of innovative or outside-the-box thinking required for the development of business strategies.
- Cost: Training extractive AI models on large volumes of data can be expensive, and periodic retraining may be required if the areas of application change rapidly and frequently. Real-time inference can also be expensive. Extractive AI does not always produce better results than conventional methods, and it may be better to use non-AI tools for analysis.
- Security and privacy: Using external services for extractive AI may expose confidential business information to service providers. While many extractive AI tools can be used in-house, the associated infrastructure costs and operational overhead may make this less attractive.
Generative AI: Potentially unlocking new insights through “creativity”
Generative AI refers to machine learning models that create new data or content based on patterns learned from existing datasets. This type of AI has the potential to enhance the finance function by generating novel insights and ideas, rather than simply analyzing existing ones. Here again, the ability to examine and refine very large volumes of data helps generative AI tools to offer “never-seen-before” ideas.
The benefits of using generative AI can be significant:
- Innovative thinking: Generative AI can help CFOs think outside the box by generating new ideas and scenarios that may not have been considered otherwise.
- Improved scenario planning: By creating multiple possible outcomes for a given situation, generative AI can help CFOs develop more comprehensive contingency plans and design processes to better track and report on unanticipated changes in business conditions.
- Enhanced strategic decision-making: Generative AI can provide CFOs with a range of potential strategies to achieve their goals, rather than simply analyzing existing data.
However, there are also serious challenges to consider:
- Lack of domain expertise: Generative AI models may not always understand the many critical nuances and complexities of financial processes or the specifics of regulatory environments. Publicly available data for training may not sufficiently represent the internal requirements of a specific business or market.
- Unreliable outputs: The quality of generative AI output can be difficult to evaluate, as it’s often based on complex proprietary algorithms and assumptions that users may not have access to. Models may “hallucinate” — produce outputs that look to be substantiated by evidence but are made up and supported by fake data.
- Cost: Generative AI models are expensive to train and may be expensive to use. In areas where the business environment changes frequently, repetitive retraining may be needed to allow the models to function.
Consider the options for a hybrid AI approach
As CFOs navigate the opportunities and challenges presented by AI, a hybrid approach that combines both extractive and generative capabilities may hold the key to success. By leveraging the strengths of each type of AI, CFOs can:
- Efficiently analyze and make productive use of existing data: Use extractive AI to analyze large datasets and identify patterns, trends, and correlations that can be used in the day-to-day management of corporate finances.
- Generate new insights: Carefully apply generative AI to create and examine novel ideas and scenarios that can inform strategic decision-making.
AI has the potential to revolutionize finance by providing CFOs with powerful tools for analysis, forecasting, risk management and scenario planning. While there are significant challenges to consider, a hybrid approach that combines extractive and generative capabilities may hold the key to unlocking new insights and driving improved business performance, enhancing the role of the finance function in driving business success.
Key takeaways
- Extractive AI: Can be useful for analyzing existing data sets to identify patterns, trends, and correlations.
- Generative AI: Can generate novel ideas and scenarios, but requires domain expertise and careful evaluation of outputs by human experts.
- Adopting a hybrid approach: Combining extractive and generative capabilities may hold the key to unlocking new insights and driving improvements in finance and other areas of business.
As CFOs continue to navigate the complexities of modern finance, a deep understanding of both extractive and generative AI will be essential for leveraging these powerful tools effectively. By embracing this hybrid approach, CFOs can unlock new opportunities for growth, innovation and strategic decision-making in an increasingly complex financial landscape.
