High-Impact AI Use Cases for Treasury
The huge hype around artificial intelligence in corporate finance makes it hard to identify the best path forward. Here are 3 options to consider.
Digital transformation has left few areas of business entirely untouched. Almost every corporate function, in all types of companies, has experienced technology-led productivity gains in the past decade. And the pace of change in business might only accelerate from here!
Indeed, as companies increasingly develop smart ways to leverage the vast amounts of data they generate through their day-to-day operations—and determine how best to automate business processes to improve efficiency and secure a competitive advantage—digital transformation is likely to continue gaining momentum.
The finance function is certainly not immune to this trend. Treasury and finance workloads, which are often data-heavy and repetitive, are highly suitable for improvement via modern analytics and advanced automation. Most finance functions have been collecting data and automating processes for years, with good results in most cases. But the idea of automation in finance tends to bring to mind simplistic chores that can be easily delegated to machines, while humans keep performing any work that requires judgment or thought.
Now, driven in large part by artificial intelligence (AI) and machine learning (ML), finance can move beyond automation of simple, repetitive tasks and toward an environment in which people and machines collaborate to transform finance and business capabilities. Gartner calls this “autonomous finance.”
As with every emerging technology trend, AI is surrounded by a lot of hype, which casts a long shadow over the market. This makes it difficult for treasury and finance teams to identify the right innovations for their organization. Blindly accepting lofty claims about the benefits of AI, without understanding the costs, can lead to underwhelming projects that don’t deliver on their promise.
To help clients on their journey toward autonomous finance, Gartner experts analyze the many use cases for AI in finance and rate them, broadly speaking, in terms of both their potential business value and their feasibility. In the realm of treasury and risk, three use cases stand out as offering the most value, while also being feasible to implement in a typical enterprise finance function:
1. Anomaly and error detection. Accounting, finance, and treasury teams face the ever-present risk that recorded transactions or balances might be erroneous, or might violate accounting principles or policies. Such problems can be nearly impossible to root out, largely because rapidly growing complexity and data volumes make it very difficult for staff to manually find errors. New and continuously evolving rules, regulations, and accounting policies exacerbate the problem, increasing the chances of embarrassing accounting mistakes.
If programmed correctly, machines are inherently more capable of identifying unusual or concerning activity in very large datasets compared with the typical finance staff member. Simple automation tools can accelerate the error-discovery process by using predefined business rules to scan a company’s records for known types of problems. This dramatically increases the likelihood that a problem will be caught, due to the sheer quantity of data that a software tool can review in a given period of time. But even bigger benefits are achievable by leveraging AI and ML.
Anomaly-detection software uses a series of ML models to identify transactions or balances that might be incorrect or noncompliant. These tools scan massive amounts of data, using AI algorithms to pick up correlations between seemingly unrelated data points. They look not just for known issues, but also for ‘unknown’ patterns that have not been specifically defined in advance by humans. And they can compare activity across functional boundaries like shipping, receiving, attendance records, accounts payable, and accounts receivable.
When they detect an anomaly or error, the systems can provide an alert, either in real time or via periodic batch processing, which enables designated users to take appropriate investigative or corrective actions. A comprehensive anomaly-detection solution will also include real-time analysis during data entry to prevent errors from entering the workflow and requiring costly downstream corrections.
Anomaly-detection systems help improve data quality and reliability. They also identify unusual activities that may indicate theft or fraud, potentially flagging irregularities that the finance and accounting team never thought to search for. Automating anomaly and error identification in a way that leverages AI can enable a treasury or finance team to spend less time fixing problems and responding to audit findings, and more time supporting business objectives.
2. Cash flow forecasting. Increasing visibility into a company’s cash positions and correctly projecting future cash flows enables better business decision-making. Capital investments can be made with greater confidence, and short-term financing options can be optimized. Improving the speed and accuracy of treasury leaders’ insights into cash flows is particularly important in today’s volatile economic landscape.
AI can help provide those insights by combining data on sales, production, expenses, and collections, then forecasting cash flows based on that consolidated information. This is essentially an extension of cash collections, in which the AI tools use patterns in past periods’ data to predict the net cash impact to the organization of future internal and external events.
For treasury and finance teams building their operational plan for 2023 against a backdrop of unprecedented uncertainty in global markets, the benefits of better forecasts should be obvious. Those that continue to rely on outdated or overly optimistic predictions risk getting hit with unexpected setbacks during periods of economic turmoil, like the one we’re in right now. This is an area that treasury leaders should currently be assessing for AI support.
3. Compliance and risk monitoring. Similar to anomaly-detection tools, compliance-oriented AI systems can review transactions or groups of transactions looking for problems related to regulatory or internal-policy compliance. ML algorithms can highlight areas in which corporate practices may be introducing an unacceptable level of compliance risk.
Rapidly proliferating regulations, in addition to the growth in data volume and complexity—which is exponential in many cases—make it almost imperative for machines to be involved in monitoring for compliance violations and risk. Using automation supported by machine learning, a finance group can validate that the company has followed compliance rules, while using far fewer resources than human monitoring would require.
An AI approach has the added benefit of minimizing the uncomfortable and discouraging disruption of compliance audits. Moreover, ML capabilities can help treasury and finance teams measure changes in the likelihood that the company will face specific risk scenarios, make business threats easier to quantify, and indicate whether counter actions or response measures are necessary.
Find the Right Solution for Your Organization
There are certainly other ways in which treasury and finance teams can harness the power of AI and ML. These three use cases sit at the intersection of high potential for benefits and reasonable feasibility for a typical enterprise treasury function, but applicability may vary across organizations and industries.
It’s worth noting that the most valuable use case exploits a company’s unique strengths and enables that organization to further differentiate itself. Specific businesses may well find better potential AI use cases for their unique context than these three. Whatever path your organization takes on the AI journey, it will be important to carefully select use cases that align to your organizational needs—ideally, as part of a finance technology roadmap that spells out short- and long-term goals.
Start with small steps and lower-risk iterations, and build up from there. An incremental approach not only helps avoid big mistakes, but also gives staff enough time to evolve in step with the technology. As time goes on and the solutions’ potential is realized, iterative cycles of improvement will start to cover a broader range of processes and responsibility.
See also:
- AI and Machine Learning in the Order-to-Cash Cycle
- Best Practices in Cash Forecasting
- AI Tips off Regulators to Possible EU Data Privacy Faults
- How Intelligent Finance Decodes Data
- Treasury Races Toward AI