Generative AI

GenAI and Its Potential: A Guide for CFOs

“By far, the greatest danger of Artificial Intelligence is that people conclude too early that they understand it”  – Elizier Yudkowsky

By now, most of us have been bombarded with multiple views on Generative AI (GenAI) and how ChatGPT will do everything: from solving world hunger, to recreating every Da Vinci painting, to winning a chess marathon against Magnus Carlsen before you can say “chess”.

The hype is indeed real, as the numbers don’t lie:

  • 34% of high-tech and 45% of telecom companies already use AI
  • The worldwide AI market revenue is expected to grow from US$327.5 billion in 2021 to US$1.3 trillion by 2032

So, how can CFOs leverage Gen AI to maximise the benefits as early adopters and avoid being laggards who miss out on the biggest party in the business and consumer world?

Start with a pilot-based approach rather than going in for a big-bang approach. Business process management firms have a Lego block-based repository of such potential pilots that can be launched in an agile mode.

Be bold and open, trusting that nothing is off limits. When the predecessor to GenAI (RPA) was becoming popular, most finance organisations used it to automate transactional activities. Even when applied to complex finance functions, RPA was merely used to insert bots to automate manual tasks; it was not used for deriving business insights from available data.

When leveraging Gen AI, financial executives must be bold and consider leveraging it in complex finance functions such as budgeting, planning and forecasting; post period-close activities like management and statutory reporting; P&L and balance sheet preparation, etc.

Adopt it personally instead of making it a mandate. This means that the CFO, finance directors and controllers must all personally use GenAI in their day-to-day functions and not just delegate it to their assistants or teams. This helps in the true adoption of Gen AI across the organisation.


What do CFOs Stand to Gain from GenAI?

Uplift experience: GenAI can be used to address supplier queries in procure-to-pay and customer queries in order-to-cash. Research indicates that around 80% of queries from suppliers involve two basic questions: “Have you received my invoice?” and “When is my invoice getting paid?.” Gen AI models can trawl through ERPs and supplier billing systems to get back with definitive responses to suppliers and customers instead of the more generic responses that human agents or current chatbots come up with. Who doesn’t like a simple, short answer to their query?

Drive self-service: Currently, the CFO function has a large team of analysts in the FP&A, budgeting and forecasting, and management and statutory reporting teams. Despite having access to different cuts of data and high-end budgeting, planning and reporting tools, most of these teams work manually on their quarterly reports. 

As an example, we know that typical questions from a finance stakeholder are: “What do I need to do to ensure that the variance between the budget and the actuals for next quarter stays within 5%?” or “What are the top three factors likely to cause a variance in my LATAM numbers?” These queries require a small army of analysts to go through various systems and dashboards to come up with a response. The result is typically a fifty-slide deck, which is static and doesn’t really encourage conversational finance. It also doesn’t give the finance leadership team the ability to extract specific business insights of interest using a self-service option. 

This is an area which has massive scope for Gen AI. A large portion of a finance stakeholder’s requests can be pivoted to an “Ask-Finance” based interface built using GenAI. Budgets, forecasts and revenue modelling can all become more agile, real-time and intuitive. As John McCarthy said: “Artificial Intelligence is the science of making machines do things that would require intelligence if done by humans.”

Tighten compliance and proactive fraud prevention: GenAI can also look at past patterns to detect any anomalies in real-time and trigger potential fraud alerts while tightening controls and compliance. Finance organisations can move from post-facto audit of samples (~5% sample) to real-time anomaly detection by in-built triggers across 100% of the data population instead of just a small sample.

Further amplify efficiency: CFOs can continue to consolidate processes (both transactional and judgemental) into a global delivery model, if they have not already done so. They can also leverage GenAI to further automate those repetitive manual tasks which RPA and other automation models have not been able to penetrate. In our experience, approximately 40% efficiency can be achieved within three to five years.


Risks to Be Mitigated

The easiest thing to do would be to jump straight into Google’s Bard and/or OpenAI’s ChatGPT, hire GenAI specialists or retrain your existing team on GenAI and drum up excitement so that they can conjure magic in six months. But that’s easier said than done. 

There are risks related to data security and privacy, regulatory compliance and the potential for algorithmic bias to creep into these models. 

This is where adopting a pilot-based approach is handy, as such risks can be tracked and mitigants identified and deployed at a small scale. If watertight enough, such mitigants can be expanded to the larger scope and the organisation.

GenAI is here to stay. CFOs and the finance function are at the cusp of becoming early adopters. GenAI can help drive immense value both from a market size and a business/ stakeholder impact perspective. There are inherent risks, but these can be mitigated by working with reliable service providers who follow a one-team partnership model. 

Don’t miss the Gen AI bus. Hop on.


Recent Posts