Data Science in the Real World

How AI is changing Financial Planning and Analysis

Acceptance and Potentials of AI tools

Rafi Wadan
Towards Data Science

--

Photo by Owen Beard from Unsplash

Digitization has a fundamental impact on all FP&A processes. However, the intensity, extent and affected process steps differ. Literature and current surveys indicate that the impact of digitization is estimated to be strongest on the Financial Planning and Analysis (FP&A), especially on processes of operation planning (budgeting), forecasting, reporting and cost-accounting. Consideration of the use of resources per process also shows that reporting, operational planning (budgeting), forecasting and cost accounting are very resource-intensive. Responsible for those duties is FP&A team.

Corporate financial planning and financial analyst professionals utilize both quantitative and qualitative analysis of all operational aspects of a company to evaluate the company’s progress toward achieving its goals and to map out future goals and plans. FP&A Analysts consider economic and business trends, review past company performance and attempt to anticipate obstacles and potential problems, all with a focus toward forecasting a company’s future financial results.

Although financial analysts have to evaluate several complex financial options and scenarios, they must also be capable of making firm decisions, being able to avoid having a vast array of financial choices paralyze them into indecision.

Due to increasing competition in data-driven markets, firms are adopting state-of-the-art information technologies for competitive advantage.

The received insights are then used to make decisions and to adjust organizational processes to generate value. In consequence, the decisions of a company are based on evidence from analytical results and not just by the intuition of their managers.

The next generation of FP&A Analytics systems — powered by AI/ML — will transform the way that future FP&A team’s work. We see two important trends in FP&A, which are Robotic Process Automation (RPA) and Financial Analytics.

RPA refers to software robots that automatically execute recurring and rule-based process steps within the framework of business processes, often across several systems, and thereby imitate human interaction. Thus, the prerequisite for the application of RPA is the existence of structured, repetitive and rule-based processes, which occur increasingly in data management in FP&A.

Several examples of RPA applications can be found in the financial sector. While many solutions exist for business localization, there is an opportunity for RPA to automate the cross-border invoicing process. RPA can transform this, and a host of other similarly manual processes, in the future.

Financial Analytics is understood as the application of statistical analysis models and corresponding algorithms to data, which usually originate from several different sources, to solve company-relevant problems based on data and to support decision-making. Depending on the time perspective, several variants of Financial Analytics can be distinguished. For example, descriptive analytics is more related to the past and answer the question of what has happened.

Predictive Analytics, on the other hand, is future-oriented and tries to show what could happen, i.e. it wants to forecast future events, which is undoubted of greater interest to decision-makers. New analytics tools can handle this problem by integrating disparate data sources and synthesizing the information into actionable insights, equipping financial analysts with the information needed to make decisions instead of bogging them down in process-related tasks. These tools will increase financial analysts’ productivity, speed, and accuracy.

Business Analytics Ladder

With the help of various techniques such as Predictive Analytics or Prescriptive Analytics, FP&A (Fig. 2) manages to abandon retrospective behavior and devote itself to looking ahead. It also enables us to rely on applications of static methods and not on the well-known gut feeling of analysts. The increased amount of data (Big Data) plays into the FP&A’s cards and enables more accurate analysis. The rule of thumb is: the more data (and variables) are available, the more precise the data analysis can be to conclude about the company. After all, even statistical approaches and possibilities through AI will not be of much support to the FP&A team if the right questions are not asked on the part of the users.

Potential of AI in FP&A

In Figure 3, we can see the potentials through AI for different activities of FP&A. Especially in the areas of Analysis, Planning as well as Reporting we have a strong reference and opportunity of AI. For scenario analysis and variance analysis, AI can be used with predictive analytics and value drivers to make more accurate and deeper analysis to generate more insights. The use of AI can also be a great help for planning and reporting. In planning, for example, prescriptive analytics can be used to solve the question of “How can we make it happen”?

While Descriptive Analytics worked with the help of correlations, clusters and trends, Predictive Analytics uses regression analysis and time series models. This means that it is possible to answer the question of when a certain result can be achieved, for example. The highest level of the Business Analytics ladder, Prescriptive Analytics, uses optimisation models. Here, with the help of decision models, an attempt is made to explain to the FP&A what must be implemented in order to achieve a certain financial result.

This supports the evolution of CFOs and FP&A analysts, as its responsibilities shift past traditional finance, accounting, and reporting duties and towards a mandate of strategic, forward-thinking leadership.

How accepted are tools currently?

During the “Congress of Controllers” of the International Controller Association (ICV), we collected data through a survey. This conference took place in 2019 in Munich under the motto “Prepare for your Future — Ideas. Learning. Networks”. It is regarded as the largest FP&A conference in Europe.

To measure the current status in terms of usage and tool support, the participants rated the perceived support and development (“To what extent are FP&A processes automated”). Constructs in this study were in general measured using a five-point Likert scale, ranging from “strongly disagree” (1) to “strongly agree” (5).

Acceptance of current FP&A Tools

According to half of the respondents, FP&A processes are very little or little automated. Compared to the progress made in production facilities or other areas, this is a low value of automation. Here, potential is noted concerned the use of RPA. Furthermore, analysts have answered that the tools provide a medium value. Especially in FP&A, data must be useful to be used and presented in the reporting. Purely ancillary information is not sufficient to support management in decision-making, e.g., make or buy. However, the question is why the current tools are not complete help for analysts. To answer this question, we asked about usability and integration.

What might be the problem?

Approximately 46% of the respondents stated that the tools were rather difficult. Conversely, this means that every second analyst considers the tools difficult to use. This can partly explain the low level of usefulness.

The question regarding the integration of FP&A tools with the ERP systems was relatively widely distributed. The reason for this may be that some of the respondents meant their analysis tools about BI. Since these BI tools were relatively focused on the integration of the interface to SAP, the result turned out to be quite high.

The aim of the statistical-quantitative calculation of cause-effect relationships between operational value drivers and leading (financial) control variables is to identify the operational value drivers with the greatest influence and thereby create transparency about the entrepreneurial relationships. Provided that data is available, the controller can then, for example, map the key value drivers of the operating result in reporting, which makes corporate management much more precise and agile and provides a better basis for decision-making. Analytics tools such as Stargazr are focused on this and have developed their algorithm that converts operational KPIs into financial figures. Respondents indicated that the tools they use have relatively little connection to operational drivers. This may be an indicator that the results are not yet supported.

Reasons why FP&A tools might fail

It can be summarized that the current analysts are not yet satisfied with the tools on the market. The tools that have prevailed on the market so far have not yet struck a chord with the analysts. One of the most relevant indicators in the area of FP&A is the “Explanatory” area. Current tools do not have this yet. Nevertheless, there is a lot of potentials, especially in the areas of analysis, forecasting and reporting.

In our next blog article, we will answer the question of how AI could impact the forecasting and reporting process.

If you liked our findings and want to support stargazr.ai, please share it or follow us on Linkedin or Twitter!

--

--

Responses (1)