It was a real pleasure hosting our recent Executive Breakfast Briefing at The Wolseley City alongside Michael Kendry from Jedox. The topic that had everyone buzzing? AI and its rapidly evolving role in Financial Planning & Analysis (FP&A). We brought together a brilliant bunch of senior finance leaders and FP&A professionals, all grappling with the same big question: how do we actually make AI work for FP&A, and what’s the smart way to evolve our scenario planning?
The discussion was really grounded. We kicked off by asking, "What does 'AI' really mean for FP&A?" And you could see people were thinking hard. Some felt the AI hype has outpaced practical application in finance. Others pointed out that all the marketing noise can make it tough to figure out what’s genuinely achievable right now.
Interestingly, a few attendees mentioned that their current platforms have AI capabilities tucked away – like predictive modelling or flagging anomalies – but they weren’t always obvious or fully utilised. This led to a good point about needing clear communication and better user education on what these tools are doing behind the scenes. Crucially, we also acknowledged that AI is already here and capable of handling significant foundational tasks very effectively. Think about it providing a solid base forecast, freeing up your team to add their expertise and market knowledge.
Now, one thing everyone kept coming back to? Data quality. It's still the fundamental hurdle. Someone put it well: "Investing in AI when your data's a mess is like putting a Ferrari engine in a Reliant Robin." Exactly. The group agreed that often, the focus is on shiny new tools, but the crucial groundwork – data structure, ownership, and governance – gets overlooked. This naturally shifted the conversation towards the vital need for solid data foundations before layering on AI or advanced planning. It's also vital to remember that while machine learning offers powerful capabilities, human oversight and accountability remain paramount. We can't just hand over the reins entirely; real people need to own the numbers and the narrative.
Scenario planning was another hot topic. While many organisations do some form of it, the feeling was it often stops at basic sensitivity analysis rather than exploring truly different possibilities. People were candid about their teams needing to rebuild the skills – and sometimes the mandate – to create robust, divergent scenarios. One finance leader admitted, "We didn’t realise we’d lost the knack until COVID hit. Then we realised we didn’t know how to respond fast enough.” A key point that emerged here was the need for guard rails around AI-driven insights. We need transparency and clear guidelines to ensure AI doesn't become a black box making wild assumptions.
The conversation also broadened beyond just data quality to data availability. Several attendees highlighted that external data sources, like weather patterns for seasonal businesses, can be just as critical as internal data. However, the reliability and forecasting of these external factors aren't always straightforward. This underscores the complexity of building truly robust predictive models.
Another crucial takeaway was the importance of framing the question and the desired outcome clearly when leveraging AI. You can't just throw data at it and expect magic; the quality of the output is heavily dependent on the clarity of the input and the questions you're asking.
Despite these challenges, there was real optimism about the potential. We discussed how, with the right framework and well-defined parameters, AI can indeed answer more sophisticated questions involving a greater number of variables, providing deeper and more nuanced insights for strategic decision-making.
So, what were the key takeaways? Well, it’s clear that while AI offers huge potential, getting the foundations right is critical. That means:
Establishing a robust data strategy, encompassing both quality and availability. No cutting corners here.
- Rebuilding strong scenario planning capabilities, with clear guidelines for AI integration. It's a core competence.
- Focusing on clear, practical AI applications that augment human expertise, rather than replacing it.
- Starting with targeted AI use cases (like providing base forecasts or anomaly detection) with well-defined questions and outcomes to prove value.
- Implementing guard rails and ensuring transparency in AI-driven insights.
One attendee summed it up nicely: "AI is a powerful tool, but it needs a skilled hand guiding it and a solid foundation to build upon."
Now, we also touched on how having the right kind of digital planning platform can really help tackle these foundational challenges that stand in the way of effective AI. Think about it – a platform that helps you centralise your data (both internal and potentially integrating external sources), structure it properly, and connect your planning models across the business. That’s not just about having fancy software; it’s about putting in place the essential building blocks for AI to actually deliver reliable and insightful results.
Platforms like Jedox, for example, are designed to do just that. They help you get your data house in order, ensuring consistency and a single version of the truth across different departments. This clean, well-structured data is what AI thrives on. It’s about having tools that are fit for purpose to create that solid base, making your data AI-ready and enabling those more advanced capabilities down the line, all while allowing for that crucial human oversight and the implementation of necessary guard rails.
So, the message wasn't so much that software doesn't matter, but that the right software plays a crucial role in laying the groundwork for successful AI adoption. It’s about having the tools that help you get the basics right, so you can then leverage the power of AI effectively, responsibly, and with clear accountability.
Thanks again to everyone who joined us. Your insights were invaluable. It reinforced that the future of FP&A is about smart people, solid data (both quality and availability), clearly defined questions, well-governed AI tools, and using the right platforms strategically to unlock the potential of technologies like AI. We’ll be sharing more on this soon, and if you’d like to explore how the right digital foundation can set you up for AI success, we’re always here for a chat.
