Skip to content

AI in finance: The digital transformation

Simply working a little smarter isn't enough — CFOs must focus on embracing a full transformation centered around artificial intelligence in corporate finance.

CFOs and controllers are under more pressure than ever to deliver better business insights, improved controls and compliance, and more efficient transactions processing. They're being pushed to become more global and scalable, and need to be able to maintain operational continuity in the face of catastrophe, including pandemics. These imperatives require a fresh approach to how things are done. 

How far we have come - what does the human intelligence process look like in finance?

To see where finance can go, it’s helpful to see how far it’s come. It’s also insightful to understand how finance professionals think and incorporate technology in their journeys. And, to help envision how Artificial Intelligence (AI) works in corporate finance, it’s useful to understand human intelligence and their similarities to AI in finance.

To understand how AI applies in finance, we must first analyze how a finance or accounting professional thinks when doing their job. For most of their activities, these professionals ingest data or information by reading documents, discussing with colleagues, or searching their computer and/or the internet. With a goal in mind, such as clearing a customer receivable, they then think about what information they have gathered. Thinking involves: memory, experience, logic and adaptation. For example, when clearing a receivable, the accountant will remember how he/she executed that activity before. They will then recall the experience, including any challenges or obstacles they faced throughout the process. 

Next, they will determine what logic or process to follow and will adapt their approach to account for any differences. This adaptation is the most challenging part for technology to mimic. After that thought process, they will then begin clearing the receivable, which likely will involve matching an invoice number and matching or combining dollar amounts, which involves more thinking. The graphic below illustrates how a finance professional uses their human intelligence to do their job.

Human intelligence process in finance

Human intelligence process in finance

How has human intelligence evolved in finance?

As recently as the nineties, the best accountants were known for being great with “numbers”. They could look at a row of numbers and identify combinations that matched a total. They could find outliers simply by eyeballing a dataset and could spot a trend in a sea of digits. But this eyeballing and spotting was not enough. They needed to reconcile accounts and create forecasting documents.

A few years later, spreadsheets were introduced, which not only helped with analyzing data, but also helped to create documents. Being good with numbers was no longer good enough; those who were known as stars in the CFO’s organization had to be good at working within these spreadsheets. These spreadsheet whizzes worked their magic to get all information and data into their spreadsheet models. They would then use their memory, experience, logic and adaptation to fine-tune or repurpose their spreadsheet models.

Their thinking became more efficient and effective as they were constantly building from tangible models with memory and logic built into them. As you can see in the graphic below, human intelligence is still at the heart of all finance and accounting activities.

Human intelligence process in finance - evolving

How has finance automation been enhanced with other technology?

In the past 5+ years, new technology has been introduced to make finance and accounting professionals more efficient, but human intelligence is still required, especially when situations and activities change.

OCR (optical character recognition), screen-scraping software, NLP (natural language processing) and chat technology are helping gather, organize and ingest data and information. In addition, RPA (robotics process automation) is helping to move information and improve how excel models are updated by replacing onerous record-and-play Excel Macros.

As you can see in the following illustration, human intelligence is still central to accounting activities. Humans are needed to determine if a change is needed and to configure those changes into the technology. This is true of all accounting and finance functions, whether it’s transactional, like inputting supplier invoices and managing exceptions, or if it’s performing analysis and determining how market impacts factor into reforecasts.

Finance automation enhanced with other technology

How does AI in finance mimic human intelligence and help perform touchless activities?

Most of us agree that humans will always be needed in finance, but probably less-so and for more complex activities as AI advances. To better understand where AI can be useful, its necessary to understand how AI works and how it’s evolving. AI stores and catalogs information and experiences. Similar to human memory, and to perform activities, AI draws on logic in the form of programming — think if/then statements.

This programming is matched with algorithms or math (statistical analysis) that helps spot relationships, outliers and trends. With additional calculations, AI can make recommendations and assign a probability of success to these recommendations. And when human feedback is added to these recommendations, AI can then adjust its math to improve its recommendations and their probability of success. In some cases, the recommended actions become accurate enough to eliminate the need for humans to continue stepping in.

An example of how AI learns and adapts is a chatbot. When a chatbot is first set up, humans identify key words and phrases that are used to formulate complete questions, which in turn are used to map to a list of standard answers. In a chatbot’s early stages, non-sensical questions and answers are sometimes produced. As human feedback is incorporated, the answers become more accurate and useful. Over several iterations, the underlying technology maps and correlates corrected answers to similar keywords, questions and answers. In effect, it adapts and becomes smarter on its own.

As you can see, AI mirrors human intelligence. Technology is striving to replicate how the human brain works. As such, AI continues to get better as other technologies are incorporated, essentially better math (e.g., machine learning) and better programming (i.e., deep learning powered by neural networks — interconnected AI cells) to help come closer to replicating how the human brain works in an everchanging environment.

AI in finance mimicing human intelligence

Where is AI most applicable in finance?

Currently, AI can be applied across many finance and accounting activities. It’s good at figuring out what data it needs if you feed it the whole data universe, and it’s great at combining and matching numbers. It’s also excellent at finding events and data that are statistical outliers and can spot trends and predict future performance. More importantly, when it comes to showcasing its talent in the CFO’s office, AI is excellent at identifying and mapping relationships. Finally, AI has gotten very good at translating characters and words.

With proper prioritization, funding, expertise and underlaying technology, CFOs can really start to employ a hybrid, highly effective/efficient workforce, one with fewer professionals focused on more strategic activities. The number of people needed to perform transactional and routine accounting activities can be greatly reduced. As you review the list below, you can envision how AI can really impact how the CFO’s organization can be redefined.

AI impacts how the CFO’s organization can be redefined

How do finance ERPs, such as Workday, leverage AI in finance?

Since most accounting and finance activities are very similar from organization to organization, there is no need for all of this math and logic to be rebuilt over and over again. As such, smart cloud-based ERP finance systems like Workday have embedded AI, machine learning OCR, chat and automated process flows into many of their finance processes, thus enabling plug-and-play automation for many basic activities. For example, data matching is embedded into account reconciliation and customer cash application business processes, thereby automating the most time-consuming part of those activities.

At Alight, we build and operate digital workers that attach to your Workday tenant. This service helps you fully leverage your Workday investment by converting historically labor-intensive activities into a fully automated, touchless experience. One example of this is our supplier invoice automation for Workday Financial Management where we handle the free-of-charge implementation. Alight owns and manages the AI-powered technology. Our clients simply pay for service consumption and achieve immediate return on investment.

Tune into our Alight Quick Takes episode where we discuss the power of Workday’s machine learning tools and how your organization can leverage them.

Related Insights


6 mistakes to avoid when deploying cloud finance technology

In this overview, Alight breaks down the six mistakes companies often make when deploying cloud finance technology.

Top 10 automation partner requirements for finance and HR

Alight has outlined the top 10 automation partner requirements your organization should look for to guarantee a successful deployment.