Photo of Priscilla Lim, Bank of America Merrill Lynch and Priyanka Bhojani, Hindustan Unilever Limited.

Unilever’s Indian treasury team have used the latest available technology to solve a very real business problem: how to automate AR reconciliation when so much of the data required is locked in emails and other communications. The AI-based solution has improved the accuracy of reconciliation at the company from just 33% in the old days of manual processes to 70% and, soon, to 99% – all this achieved within just two months of project start.

Hindustan Unilever hails big success in automating AR reconciliation

The challenge

Unilever’s subsidiary in India, Hindustan Unilever Limited (HUL), supplies consumer goods to retailers of every size, from tiny rural kiosks to urban megastores. That diverse business brings with it some challenges for the company’s treasury team, especially when it comes to reconciling accounts receivable (AR).

Given the significant volume of its business, including from large retailers and distributors in the company’s ‘modern trade’ business, incomplete remittance information typically leads to an arduous and costly reconciliation. Additionally, many of these retailers and distributors tend to have their own systems and processes and often pay multiple invoices through one single payment, advising which invoices the payment covers in a separate email payment notification.

Until this year, this complexity meant that reconciling AR for modern trade customers was a very manual process – AR staff had to make a three-way match between the credits shown in the bank account, the open invoices on the company’s SAP system and the payment information in emails sent by the customer.

Multiply that process by some 90,000 payments received every year, covering hundreds of thousands of invoices, and it was clear that any improvement in the way things were done had to be a good thing for HUL.

The ideal solution was to automate as much of the process as possible – but any new system had to be able to cope with the ‘natural language’ information contained in those emails and, at the same time, deliver reconciliation that was more accurate than the 33% achieved by the previous processes.

The solution

The answer came from HUL’s banking partner, Bank of America Merrill Lynch, which had recently launched a new service in India, designed for exactly this task: Intelligent Receivables.

Intelligent Receivables uses AI and other software to help achieve improved straight through reconciliation through four steps:

• Identifies payers and associates their payments to remittances that are received separately.
• Extracts remittance data from emails, email attachments, electronic data interchange (EDI) and payer web portals.
• Matches payments to open receivables using the enriched remittance data.
• Creates a receivables posting file that the client uploads to their ERP system.

Within weeks of starting up, the new system had pushed reconciliation accuracy past 70% and will reach 99% once very large clients’ invoices are brought into the system – a huge improvement in a short space of time.

Best practice and innovation

The team at HUL have used the very latest technology available (Intelligent Receivables was only launched as a product in 2017) to solve the problem of how to improve operational efficiency by moving away from manual processes and hence reap the benefits of greater automation.

Through their willingness to move fast to bring the new system into operation, the team have become an ‘early adopter’ of next-generation AI technology, becoming one of the first companies outside the US to go live with such a system.

Key benefits

• The new system has improved the accuracy of reconciliation from just 33% in the old days of manual processes to 70% and, soon, to 99% – all this achieved within just two months of project start.
• The AR team can now deliver a much more accurate and timely picture of the company’s AR status than was possible in the past – with positive effects on key business metrics such as days sales outstanding (DSO). The solution’s dashboard reporting can assist in cash forecasting and help clients better understand payer behaviour.

Key learning points

• Continuous improvements in processes through digitisation is one of the key essences for treasury teams today. Introduction of IREC was a step in this direction.
• Integration of AI with existing system (SAP) led to better quality output without major change in the existing IT set-up.
• Technology including NLP has helped to leap forward to a point where manual intervention has become minimal.
• This has also provided an added advantage of improvement in accuracy and efficiency.
• This has proved that smart and practical usage of AI can provide relief in doing repetitive and mundane tasks.