Michael Lengenfelder, Head of FP&A Product Management at Unit4, explores the emerging role of GenAI in FP&A, stressing the importance of pragmatic integration. While GenAI holds promise, Lengenfelder underscores the necessity of combining its capabilities with human expertise for effective financial storytelling. He advocates for careful consideration of data integrity, training and trust-building within organisations to maximise GenAI’s potential.
Despite the hype, the adoption of generative artificial intelligence or GenAI in FP&A is still in its very early stages. It has a use case, and, at Unit4, we are working on a variety of scenarios to understand where GenAI and other Artificial Intelligence (AI) tools may be applicable across our application suite to automate tasks and drive more efficiency for users.
Our CEO, Mike Ettling, recently talked about merging the power of AI with human-centred design to streamline user interactions with enterprise resource planning (ERP) systems and reduce errors, as well as alleviate manual and repetitive tasks. In the long-term, it is possible that harnessing AI will enable our applications to adapt and respond effectively to triggers such as user behaviours, tasks and location. In the world of FP&A, this feeds into our ambition to create a user experience that is not just intuitive and engaging but which is also context-aware, smart and forward-thinking.
There are exciting possibilities for GenAI in FP&A. Imagine if the annotations that line-of-business leaders add to FP&A reports at quarter end could be used by GenAI to generate a summary report of that commentary.
GenAI’s linguistic capabilities are probably its most valuable asset to FP&A teams right now. When compiling assumptions about a budget, a GenAI tool could sit alongside an FP&A application to provide the report creator with insights that inform those assumptions. Pragmatism, though, is invaluable in this situation. Data hallucinations are a well-documented issue with this technology, so a company’s AI policy should stipulate how the FP&A team engages with GenAI, how it interrogates the answers it provides and how it ensures the sources are authentic.
Lay solid foundations for GenAI
It is important to lay the right foundations for adopting GenAI – starting with the organisation’s policy. This should address points such as:
- Setting guardrails to prevent sensitive financial information from being shared with an external app or being exposed to the public internet. It is also important to have role-based authorisation to ensure only the right individuals are sharing information with the tool
- Ratifying how to train the application to avoid data hallucinations, appreciate nuance and deliver accurate analysis. In FP&A settings this has layers of complexity because the GenAI will require access to subplans to appreciate conditions that are unique to a particular organisation. Getting this right will make it possible to understand and explain how decisions are reached
- Delivering technical and business training to show how FP&A professionals can use the technology and see it as an enabler to empower them to take on more rewarding work
Do not underestimate that adopting GenAI is both a technical and a people challenge. Given the fears that AI could make some job roles redundant, it is important to give employees the time to embrace the technology and understand its value to them. This will prevent institutional memory held by experienced employees from leaving the organisation. Their knowledge of the nuances of FP&A reporting should never be underestimated. If these people feel threatened by AI, they may choose to leave.
Ultimately, it is crucial to build trust across the organisation. The FP&A team must be reassured if they are to work alongside the technology, and senior leaders must have confidence that it is delivering accurate decisions.
Context. Context. Context.
This is why FP&A will still require human input to correlate financial results with external events and pressures. Much of that information resides in the heads of line-of-business leaders. This context is crucial and FP&A professionals must extract it to create a financial performance story that the business understands and trusts. The Actuals may state that marketing has significantly underspent in the first quarter, but the factors influencing this situation can only be extracted by interrogating the data and speaking to stakeholders.
Today, this is too difficult for GenAI. Vendors are building specialist versions of these tools, but they are still quite generic, which means it will be difficult to personalise analysis to an individual FP&A story.
Good FP&A reporting is indisputable
The more pragmatic approach is acknowledging that FP&A teams already have a pretty good model for reporting and telling an indisputable story about an organisation. Building a budget and forecast, then reviewing progress by comparing Actuals to Forecast provides a solid foundation for identifying variances. Today, an FP&A team can examine these deviations by testing assumptions in subplans more quickly than it takes to setup, train and validate the results from a GenAI tool.
In the future, if such tools can absorb all the relevant information within an FP&A plan, understand the nuances of an individual organisation, as well as the impact of external forces on that plan, then GenAI will be able to take on greater responsibility.
That role will rapidly expand in the coming years as the technology matures but, today, FP&A professionals are critical to the storytelling around financial performance. In the next 12 months, FP&A professionals should be available to understand the technology and its capabilities, then begin to map out what needs to be put in place to expand its use case. The responsibilities of FP&A professionals will likely change in the future, but thinking now about how GenAI can be integrated into your FP&A storytelling, it will create new opportunities for FP&A professionals.