Artificial intelligence (AI) now plays an important part in our lives; many of us interact with digital assistants on our smartphones, in our cars or in our homes via smart speakers every day. But AI can also add significant value for corporate treasury; it is a key tool in the shift from manual to automated processes and can help eliminate errors associated with human intervention. It can also support scalable growth, business transformation, the reallocation of employees and resources to new projects due to time and cost savings, and boost the bottom line.
For treasury managers, CFO and finance teams, AI’s flexibility means that it can be applied to multiple tasks, including payments, receivables and customer services – it can even provide valuable insights for management about company performance. Moreover, unlike a treasury team, AI can work 24/7. However, to get the most from AI, companies need to think about how their treasury function will operate in a future shaped by COVID-19, changing payment preferences, and a population that may continue to work remotely.
Meeting healthcare companies’ challenges
In the healthcare market, there is a rapid shift underway to electronic payments and away from cash and check transactions. Telemedicine usage has increased by over 100%, and start-up medical companies are providing medical and medical-related services to more people. These new business models and entrants necessarily expect their treasury functions and processes, such as data reconciliation, to operate differently to established healthcare firms.
To take just one example, Citi has recently used AI to help a global healthcare company automate its manually-intensive accounts receivable (AR) process of matching payments received against expected and open invoices (known as cash application). This application of AI was outside of the USA, but AI can work both in international and domestic operations in a seamless way. Cash application is a common pain point for many multinational companies, often consuming significant physical and financial resources and potentially acting as a drag on company performance.
This healthcare firm needed a solution for its shared service center and wanted to increase its straight through processing (STP) rates. At the time, its cash application processes were manual and time-intensive. Moreover, these arrangements were not scalable, creating barriers to growth for this rapidly expanding company. Working with Citi, it wanted to address these challenges and optimize the following:
• Remittance detail handling: The AR team used a manual cash application process that involved calling customers and analyzing bank account statements (from multiple banks) to match to expected/open invoices. In order to scale, the company needed to automate this process.
• Applying cash: Credit notes issued against customer payment discounts required manual calculation of adjustments, which often caused delays in the cash application process. The healthcare firm wanted to reduce manual entries, which can be prone to human errors.
• Unidentified payments/postings: Instances of unidentifiable postings such as transaction charges, partial or bulk payments often created additional work for the AR team and could be burdensome. The company anticipated expanding its business both organically and through acquisitions, and wanted to reduce operational complexity and easily integrate new entities to this solution.
At the same time, the healthcare company wanted to increase centralization of cash application at its shared service center, shifting it from regional offices (where it was often a part-time role for staff with other core responsibilities) to dedicated specialists.
How artificial intelligence and machine learning can help
AI and machine learning (ML) have the potential to improve operational efficiency, automate processes, and optimize working capital when deployed to address cash application. ML is the way the software and algorithms in AI optimize the solution through continuous improvement. Citi® Smart Match leverages proprietary machine learning technology along with Citi’s cash management infrastructure to fully automate the cash application process as well as addressing the healthcare company’s unidentified payments/ postings and enhancing its remittance handling workflow.
Citi Smart Match uses a combination of AI and ML to accurately match expected payments to received payments and remittance data. The solution:
• Organizes the automated incoming data stream.
• Scans emails and email attachments for key remittance details.
• Learns from payment patterns and behaviors to continually improve matching rates.
• Offers an intuitive user interface – fully integrated into CitiBusiness® Online or CitiDirect online banking – that allows AR operators to handle exceptions and tweak system logic.
Streamlining workflow and delivering tangible benefits
By harnessing the power of AI and ML, Citi Smart Match provided the healthcare firm with key performance indicators on how the solution is performing, as well as metrics and insights related to paying customers, helping to improve the company’s cash flow forecast. By reducing delayed payment postings, the application has boosted the company’s auto-reconciliation rate. When exception items (unmatched or partially matched payments) occur, they are managed within the platform, strengthening the company’s processes and governance.
The qualitative and quantitative benefits from implementing Citi Smart Match include:
• Boosting match rates: The healthcare company previously achieved a 0% match rate through manual cash application. Just five months after implementation, once the AI trained itself to accommodate the company’s AR working practices, automation has increased match rates to 94%.
• Upskilling the AR team: The company has reduced weekly hours spent on reconciliation by 50%, and this is expected to increase in the future. As a result, shared service center employees have been reassigned to more value-added work.
• Improved customer experience: Automation of cash application has led to fewer follow up calls/emails to customers, a better customer experience and improved customer engagement.
• A streamlined process and enhanced working capital: Richer payment and remittance data enables past-due payments to be closed-out, facilitates data-driven insights from user interface, provides a more efficient order-to-cash process and has improved working capital.
• Potential business opportunities: Payments are applied more quickly, freeing up customer credit lines and creating opportunities for increased sales.
• Easy scalability: The solution was initially deployed in select regional markets but is being expanded to additional markets. The scalability of Citi Smart Match, which is a global solution, means it can be easily integrated as the healthcare company expands organically or via acquisitions.
Conclusion
AI and ML have the potential to bring new efficiencies, improved risk management and greater visibility and control to treasury. While automation has long been a goal for many companies’ treasuries, AI and ML add an additional dimension: the ability for technology to effectively think – and learn – for itself.
The capacity for solutions such as Citi Smart Match to continually improve themselves by analyzing data, working practices and exceptions ensures adaptability and provides a reliable means for solving companies’ challenges. But the solution also future proofs treasury to address the challenges of the future, including those associated with expansion, either organically or by acquisition.
AI and ML have rapidly found a role in our everyday lives but have been slower to catch on in treasury. However, this seems certain to change as the benefits become more evident, freeing up treasury professionals for more value-added tasks and helping to create the efficient, flexible treasury of the future.
