Salesforce looks to generative AI to tackle banking's "$83 billion problem" with customer disputes

According to research, last year $83 billion of transactions were disputed by customers with their banks. The Sift Digital Trust & Safety Index report also found that the cost of dealing with such disputes has increased by 16% year-on-year, so that’s a big deal for financial institutions to manage. 

Last week at its World Tour New York City event, Salesforce introduced a number of AI-enabled offerings to tackle this issue, including as per the official announcement: 

• AI-Powered Transaction Dispute Management that helps bank service agents streamline the entire dispute management cycle – from the time a dispute is submitted until its resolution – while enabling more transparent and consistent communications with customers, card networks, merchants, and issuer banks.
• Einstein Copilot Banking Actions, a conversational AI assistant for CRM that enables bank service agents to ask questions and receive trusted, relevant responses grounded in metadata, and automate tasks directly in the flow of work.
• Service Processes Library, which spans an expansive list of pre-built service process templates in Financial Services Cloud for common banking use cases, including  transaction disputes, fee reversals, and card replacements. 

The new features will be included at no added cost for all Financial Services Cloud (FSC) users. AI-Powered Transaction Dispute Management is generally available, while the Services Processes Library is available with 19 processes today, including Transaction Dispute Management. Einstein Copilot Banking Actions will follow in Fall 2024.

Getting worse

According to Cory Haynes, VP, Financial Services Marketing at Salesforce, the problem of customer disputes is only getting worse in the Vaccine Economy: 

After COVID, everybody went digital - 4.4 billion customers are using digital payments in digital transactions. That's causing a lot of problems - 'I didn't make this transaction, I didn't purchase it'. There's fraud. And people's lives are being put on hold because they're waiting for money to come in their account. 

You're trying to figure this out. You can't beef up your contact center. You can't triple the people in there. If you went to your CFO and said, 'Hey, I need to add three times the amount of people in my contact center', they would throw you out. But what if you have AI automation and data to help you with that? 

Chargebacks and disputes are part of the daily routine for Gaurav Mittal, EVP Ethoca, Cyber and Intelligence Solutions at Mastercard. He said: 

Fundamentally, chargebacks and disputes are one of the two reasons consumers use cards in the first place. The reassurance [is] that their money is safe and if it's used in ways that are not intended by them that they can get those get that money back. The problem is material and large. It costs $37 to $40 to handle the dispute and there's over 400 million of these disputes in a given year. So, the ability to bring tools to our issuing partners, to our agents that serve on the frontlines of solving this problem for consumers every day, is really an important one. 

Trust is, of course, a crucial commodity in banking circles - if customers lose faith in the institution, the consequences can be horrendous. So transparency is hugely important, said Mittal: 

We're excited about democratizing not only [transaction] information, but also then giving the agents the tools to solve disputes, not over weeks or months, but sometimes in days and often in minutes. Trust is really important in financial services and particularly regulated industries, and the ability to provide consumers visibility into where their money was spent, what they bought oftentimes, is a great tool. We've seen consumers use it to resolve confusion about the purchases, but also at the time of doing their tax returns or going back often to re-purchase something. 

There’s also potential bottom line benefit for financial institutions here as well, he added: 

We find that consumers who are provided this transparency by their banks will often improve the usage of those cards at those banks, driving up top of wallet behavior, creating value not only for the consumers, but also for the issuing customers and the retail banks that serve them.

Data, data, data 

Data is key here, said Haynes, as well as to having a holistic view of the customer journey as a whole: 

You want to think about the entire journey of your customer from start to finish when they first open a checking account until they retire. This is important and why we invested in our Personal Financial Engagement tool. We launched it about a year ago, but we've really enhanced it with three main things. 

Number one, we spent time making sure that our data foundations were there. Now that we can pull in data from Data Cloud, we've also made sure that FSC is also an object that goes throughout the cloud making sure the data comes in. We also have partnered with third party data to make sure that we can take in core information from your banking transactions, taking custodial information from your wealth, and also working with financial planning tools to pull that in as well. 

We wanted to make sure that we have visualization of data. There's nothing worse than having data but you can't see it. You can't understand it. So when you're having a conversation with your customer, you can let them know where they are at what point in time in their financial journey. 

And lastly, is making sure that we have insights into actions, not just showing you the data but actually acting upon it, giving triggers. Because Data Cloud can pull in real-time information, we can let you know when that customer is actually spending too much and maybe they have a call or conversation or maybe to change their goals and plans. You don't have to wait every quarter to have a conversation; you're engaging with your customers as a great coach, as an advisor as it happens.

My take

One of the things we know is that we're not just doing AI for AI's sake. It's not just to be cool.

One of the banes of my life is dealing with finances and when I run into problems, my frustration levels rise quickly. Improving the time taken to address and resolves contentious banking issues or transactions is something that would be hugely welcome. If AI’s going to help that happen, then that’s a genuinely useful and practical application of the tech to an issue that can impact on every one of us.