- GenAI Improves Software Engineering Productivity by 70%, says Study
- There’s a Critical Need for Robust Security Testing Solutions as Businesses Harness the Power of AI
- NextTech 2024: Explore the Future Insights of AI Advancements, Next-gen Tech, and Human-centric Approaches
- Industrial Metaverse Spend Increases as Immersive Engineering Use Cases take Hold
- AI Can Unlock New Ways to Communicate, But Focus on Getting Basics Right
- Eight Global Tech Companies Commit to Ethical AI in Collaboration with UNESCO
The chief financial officer of a large Indian hospital group had a problem. To optimize spending and enable the organization to provide the highest levels of patient care, she had to analyze hundreds of indicators: patient information, such as age and insurance; clinical data, such as treatment and laboratory tests; operational data, such as food quality and the responsiveness of medical personnel; and financial data, such as costs and revenues. The hospital’s management software was limited in its ability to handle this complex analysis. To address this challenge, the CFO installed a new decision-support system powered by machine learning. After action was taken on the resulting insights, patient satisfaction scores increased by 12%, and the hospital group enjoyed significant savings by eliminating activities that didn’t add value.
Like the hospital CFO, finance leaders in a wide range of industries are trying to integrate data and analytics into finance office processes to improve operations. However, not all CFOs have cracked the code for capitalizing on the promise of these tools, in particular new machine learning applications. In a 2022 study by Gartner, 80% of CFOs said they believe that finance needs to significantly accelerate implementation of machine learning and artificial intelligence in order to effectively support and protect the business. Why are so many CFOs slow to take full advantage of these technologies, and what can be done to bring them up to speed?
In my work as an academic researcher and consultant, I’ve studied hundreds of CFOs and senior finance leaders from more than 20 industries at companies with revenues ranging from $250 million to more than $20 billion. In conducting research that included interviews and a survey, I sought to understand how CFOs use digital technologies in their finance offices, what they want to achieve through digitization, and the main challenges they face in developing these capabilities. I found that the CFOs who had successfully deployed advanced analytics were able to deliver more real-time insights, reduce human error and bias, and speed up processes and decision-making.
Digital finance leaders also differentiated themselves by spending a greater portion of their time on value creation activities — such as performance management, mergers and acquisitions, pricing, and strategic planning — and value protection activities, including fraud detection and risk management. They spent less time on transaction processing activities, such as invoicing, payment processing, and accounting. Most had already gained efficiencies by automating clerical tasks and other manual processes.
Digital finance leaders spend more time on value-creation activities — such as M&As and pricing — and value-protection activities, like risk management.
I identified seven advanced capabilities that underlie these leaders’ accomplishments and serve as building blocks for transforming finance into a data-driven function that creates value from advanced analytics. I share detailed descriptions of each below, along with an assessment tool that can be used to evaluate the current level of data analytics maturity at any stage of the finance office’s digital journey.
Seven Essential Capabilities for Digital Finance Leaders
1. Use advanced analytics to meet strategic needs. Many organizations still expect finance to keep to its traditional operational role of maintaining ledgers, processing transactions, and delivering required reports. More advanced finance leaders, however, think in terms of digital strategy and look for opportunities to apply technology to advance the strategic aims of their organization.
The CFO of a European pharmaceutical company was asked for clearer, faster, and richer insights that could inform the company’s pricing strategy. In response, the company built a new machine learning model that generated dynamic forecasts of drug prices that almost instantly reflected changes in the cost of raw materials, the cost of production, regulatory guidelines, the cost of inventory, the supply of drugs across the globe, and other market and product-related variables.
Armed with this information, business leaders could quickly adjust pricing based on a real-time feed of market data and drug characteristics. This gave the finance department a more active role in developing and implementing pricing strategies and enhancing the value of the company.
Another CFO, at an international retailer, recognized that advanced analytics could be applied to bolster the company’s cybersecurity profile. The CFO worked with the CIO and the company’s cloud provider to implement a machine learning capability that would identify sensitive financial data, intellectual property, employee information, and other data residing in the cloud and warn the company when the data was being accessed or moved in an unusual fashion.
Like these two CFOs, most of the digital finance leaders I spoke with said they consider technology and advanced analytics to be strategic priorities for protecting company assets and achieving measurably better business outcomes.
2. Unpack and explain machine learning outputs. Some people see machine learning models as a black box because it’s unclear how they arrive at the predictions they produce. Digital finance leaders understand these limitations and push for model explainability — techniques and tools that can explain how models reach their results.
The CFO of a global car manufacturer found that explainability and interpretability were among the main challenges of using AI and machine learning to detect financial and credit risks. In such a context, it’s critical to understand how a model is making predictions in order to confirm that it’s not employing proxies for any demographic characteristics that would be illegal or unethical to consider when weighing a credit application.
One CFO and his team wanted to know why a machine learning model used to determine whether an applicant should receive a car loan was denying certain applicants loans but not others, and why it gave some individuals a 70% chance of defaulting but gave the average applicant only a 20% chance. They applied Shapley values, a method derived from game theory, to show the relative impact of each feature (or variable) on the eventual output of the machine learning model. This showed the CFO and his team that home ownership, loan-to-income ratio, and credit card debt were the three most important variables for predicting car loan defaults.
Compared with other CFOs, digital finance leaders place a higher priority on understanding the inner workings of their analytical models and how they affect the results the models deliver.
3. Collaborate on data management across functions. The exponential growth in the size and complexity of data poses a significant challenge for finance teams as they seek to identify a single source of truth for data points used to derive actionable insights for the rest of the organization. Digital finance leaders have a clearly defined strategy guiding the collection and management of, and access to, this rising volume of data — and many play a crucial role in defining the overarching data strategy for the organization.
In developing a new analytics solution, the CFO of an Asian steel company first defined a data strategy and worked to align data standards across the enterprise, raising the issue to the top of the corporate agenda. His team’s objective was to view revenues and profits by customer segment, not just at the product level. That meant identifying the core metrics of customer profitability, such as acquisition costs, order and churn rates, and service costs, and finding a consistent way to define those metrics across business and functional domains. They also implemented a data lake to streamline data sources and automated the process of tracking customer metrics. These improvements not only enabled finance to track the metrics it sought but also helped business units to respond more quickly than competitors to changing customer needs and to serve them more efficiently.
Digital finance leaders are much more likely to act as strong advocates for democratizing data across a company’s silos, in stark contrast with more traditional finance leaders who treat information as power and often guard the data they possess.
4. Embed analytics skills. Digital finance leaders take steps to build data and analytics skills within their finance teams rather than relying on the expertise of specialists. This requires that they make sure the majority of finance team members receive relevant training in data science, data management, and programming. At the same time, the use of advanced AI tools has highlighted the need to develop communication and problem-solving skills and to learn to translate data into actionable insights. It’s also important to encourage more experienced finance and accounting staff members to roll up their sleeves and start experimenting with potential data and analytics use cases and to help bring along those who may resist the use of data and new technologies.
To develop a team with a stronger orientation toward analytics, the CFO of an international hotel group stopped hiring candidates with traditional finance or accounting backgrounds and instead recruited data scientists, mathematicians, and data engineers. She provided these new hires with on-the-job training in the fundamentals of budgeting, financial analysis, compliance, risk management, and other accounting principles. Senior practitioners were required to spend at least one day a week helping to develop the accounting and finance skills of the new staff members. Meanwhile, the new hires with more expertise in analytical techniques helped senior practitioners translate finance and business issues into questions that could be answered with analysis and data engineering capabilities.
5. Explore and experiment. Digital finance leaders rarely make linear progress toward the goal of successfully leveraging data and analytics: Many told me that they stumbled initially. It’s critical that CFOs are patient, take an exploratory approach, and tolerate failure when experiments don’t work as planned. It can take quite a while for AI-driven models to produce high-value outcomes, since they require adequate data, fine-tuning, and scaling.
It’s critical that CFOs are patient, take an exploratory approach, and tolerate failure when experiments don’t work as planned.
To cultivate a culture of exploration and learning, the CFO of a construction engineering company rotated finance employees through other business units, innovation teams, and IT. These employees built important relationships, increased their understanding of colleagues’ needs, and identified new opportunities for improving the finance function’s offerings. The CFO then asked team members to use the knowledge they had gained to suggest ideas for using technology and advanced analytics to improve finance services and awarded bonuses and recognition to employees who submitted proposals. That generated excitement and encouraged the more-hesitant team members to bring their ideas forward.
6. Manage cultural change. Finance teams’ often conservative, cautious culture may contribute to a resistance to new tools and new ways of doing things. Tax accountants who pride themselves on being absolute experts in tax law, for example, might reject the notion that a machine could provide better answers about tax optimization and thus ignore an AI tool’s recommendations. Successful digital finance leaders take steps to establish a data-driven culture and explain how employees’ adoption of advanced analytics and new technologies will enable them to contribute most effectively. For example, the CFO of an international IT company seeking to drive digital transformation made it clear to his finance team that they were expected to incorporate advanced analytics into their work and problem-solving. The CFO also created a booklet for the team’s tax accountants demonstrating how combining their expertise with AI’s tax
optimization recommendations could increase the company’s profits.
7. Include everyone on the team in digital efforts. Digital finance leaders give all team members opportunities to participate in the digital transformation and contribute to the team’s success. An inclusive environment is one where everyone feels appreciated and can develop the right digital skills. Senior finance leaders need to make sure that everyone on the team has access to relevant training, technologies, and data and that learning new skills opens up opportunities.
A CFO at a leading telecom company found that almost half of the finance staff lacked the analytical and technology skills needed to achieve the company’s goals; furthermore, many were engaged in activities that were likely to be obsolete in a decade. Instead of eliminating employees, the CFO encouraged them to learn new skills. In the first year of the program, the company was able to fill a majority of new job openings with internal candidates.
Developing the capability to effectively leverage data and analytics in the finance function requires CFOs to take coordinated actions along several dimensions. To help them identify the investments they need to prioritize, I have mapped the insights from my research to a maturity model and corresponding assessment tool. The results should indicate the analytics sophistication of the finance team and provide clarity about the skills and mindset that must still be developed to drive digital transformation forward. CFOs can repeat the assessment regularly to monitor progress and focus on improving weaker capabilities.
Each finance team’s path toward the successful deployment of advanced analytics will vary based on its starting point. But the common themes I’ve identified among digital finance leaders can provide valuable direction along the way.