Spending over 6 months creating the next year’s budget is a thing of the past. With newer technology, not only is a huge amount of time saved, but budgeting and forecasting has become more in depth and analytical. Predictive models are now more sophisticated as well. Artificial intelligence (AI) and machine learning (ML) play a big role in this, but how do they come together and how will they continue to influence FP&A?
The London FP&A Circle put together a digital panel and created presentations and discussions about these trends and how companies envision the changes. There are many different terms for combining machine intelligence with human intelligence, and in this webinar the word “articulated intelligence” was used. This is because it covers both the known driver based models that people use very well, and the hidden drivers and patterns that machines cover.
Throughout the seminar there were a few different study cases and polls in which there is a lot to learn from:
Case Studies
Case #1: Managing Director overseeing digital FP&A transformation at Konica Minolta, a multinational manufacturing company of business and industrial imaging products.
The managing director, Igor, joined the company over 10 years ago and has seen huge changes in the FP&A process over the years. What started with reporting cycles taking a full week or more to complete and shifting through hundreds of budgeting and planning spreadsheets, turned into a completely different process.
Although Igor and the team maximized the usage of spreadsheets, they are still complex, prone to manual error, not team based, and overall time consuming and inefficient. In terms of budgeting, the limits are even greater, as bringing in operational data sets is limited in scope and it is simply harder to integrate.
After setting out on a digital FP&A process, the company implemented data models and SQL statements. Data visualization tools and predictive algorithms were also adopted.
As a result, the mundane and frustrating tasks mostly ended. Budgeting and forecasting reports were updated in real time and integrated across different markets. Drilling deeper into the data now gave them more long term outlooks. And most importantly, the entire team became more engaged in decision making and skilled processes due to the time freed from the manual FP&A input of only a few years ago.
Case #2: VP of performance control moving towards AI in FP&A at Siemens Healthineers, a German medical device company
The VP of performance control, Saurabh Jain, says that their company is in the middle stage of transformation: “We are evolving from the trigger-based data automation center to a gradual improvement of robotics (macro on steroids), cognitive computing (leveraging of machine intelligence) to dynamic self-learning AI to fully automate the tasks.”
Finance and budgeting has always combined data from multiple transaction and source systems, but automation and advanced ML allows it to be faster and more efficient. This allows the team to bring in more contextual and complete data, such as from data sets outside of the company that weren’t necessarily available before.
Jain, described 3 cases in which the company used ML and AI to improve:
- Using ML to ingest over 5000 data elements and perform calculations to identify performance drivers and predict revenues
- Run supervised learning algorithms across historic data to predict patient behaviors
- Perform process mining to identify optimisation opportunities within the order-to-cash cycles, improve the cash position, and cash flow projections
What Jain learned from watching the evolution and change over the years:
- Over the last ten years, there is now less time spent on running queries, cleaning, checking and mapping data.
- In the past, FP&A has lacked business understanding. They need to know the influences and drivers, in order to be able to apply the insights
- More time is being spent on data governance and data processes
Case #3 Head of FP&A, BI, and Cost Out initiatives applying machine learning to revenue forecasting at Amazon.
Amazon famously has a very efficient budgeting and planning process, however there are always ways to improve. The company has a huge number of vendor agreements which are continually disposed of and replaced by new agreements. Disposals create more than 10,000 hours of rework in finance departments!
By applying ML and collecting enough data, Amazon started to detect risk agreements prior to billing. By using an out of the box ML model to crunch data relating to past agreements, 26% of agreements were detected as a high probability of being disposed of.
While 26% may not seem like a lot, it still saves a significant amount of time, and this is just the beginning with limited ML input. Since ML improves as it is fed more data, this percentage will only increase. Amazon expects the percentage of detected agreements to increase to 60%.
Here is one of the key takeaways from the implementation from Amazon: “Getting started with ML does not cost much. There are a lot of applications readily available. However, you need to invest time to learn. Where once SQL was considered rare in finance, it is now commonplace. Similarly, ML will become a standard requirement in future FP&A roles.”
Polls
Poll #1 How would you describe your current FP&A process?
A whopping 60% responded that they “have some predictive elements (e.g. some drivers)” while 22% responded “static, not predictive”. Surprisingly, only 5% of participants responded that their current FP&A process is digital, meaning fully using automation, predictive analytics, and AI/ML.
Poll #2 To what extent would you say your organization uses AI/ ML in Finance and FP&A?
Once again the response was quite surprising as close to half (49%) responded that they are using zero AI/ML in finance or FP&A! A more encouraging 36% “plan to use AI/Ml in the future”, while 14% use some AI/ML in finance and FP&A. Most shockingly was that 0% answered that they are using it completely.
Poll #3 What do you believe are the biggest challenges on the FP&A analytical journey?
As fitting with the challenges to today’s workforce, the biggest response to challenges is talent and skill set (47%). Next was culture (21%) followed by data (20%) and finally technology (13%).
Poll #4 What is the key factor to further advance the digital journey in finance and FP&A?
Interestingly enough, for the 2nd poll in a row, technology wasn’t the biggest challenge in using or advancing digital FP&A. Leadership support (43%) led the poll followed by skills and competencies in finance (32%). Technology was only named in 14% of the answers and business partners buy-in at 11%.
Conclusion
Both the case studies and the polls have shown us what stands out amongst companies all around the globe: Almost every company is aware of the benefits and strong need to move to FP&A automation and ML, yet very few are taking immediate and complete action. Many organizations are in the process or intermediate stages, but very few have reached the optimal level of efficiency.
While many people think that using ML and AI in financial processes requires a big monetary investment and a complete 360 turnaround of the way things are done today, this is not necessarily true. SaaS technologies provide easy to use platforms for automating data and forecasting. With short implementation times and affordable prices, companies see immediate returns and improvements across all FP&A areas.
