Gradient Labs successfully applied Artificial intelligence (AI) customer service is more advanced and faster than most fintechs. Key to his success, co-founder and CEO Dimitri Masin It is to provide a better and faster result that will further satisfy customers with the best human support available.
This approach has allowed Gradient Labs to go far beyond applying AI to the most basic customer questions. Masin said the founding team knew from Day One that they had to reach higher because these interactions represent at most 25% of customer operations. Mission accomplished, Gradient Labs achieved a higher customer satisfaction score than most human teams while automating expert support and back-office processes covering 75% of workflows.
Masin and co-founders Danai Antoniou And Neal Lathia He worked at Monzo and served as vice president of Masin Data, Data Science, Financial Crime and Fraud. He said that as regulations and technology evolve, it empowers neobanks to innovate.
However, this innovation can be pursued in different ways. Some founders push the regulatory framework, believing that regulations will catch up or force them to change it themselves.
It’s better to start from the perspective that innovative products can be brought to market that both push boundaries and make regulators happy. The editing isn’t bad; Organizations such as the Financial Conduct Authority care about protecting customers. When you build a truly customer-focused business, you can safely open new paths.
Why is AI the next logical step in customer service?
Masin sees AI as the next phase in the evolution of financial services and customer service. Branch banking has given way to digital-first fintechs, which often create sleek experiences but hit the same wall.
Huge teams of people were still needed to complete a long list of repetitive behaviors like account queries, KYC, AML, dispute resolution, and more.
“And the second part of the problem is what often still causes bad experiences.” You bought the machine.
Going beyond answering these few general questions is what allowed Gradient Labs to reach 75%. Mapping out the answers to Level One questions is relatively easy, Masin said. Investment questions? Allegations? Conflict resolution? This is a different level.
Common disputes will see more than a dozen normal queries that can be asked in different ways. Further steps are required, including consideration of relevant customer transactions and overall history.
But most of these steps are well documented. Customer service personnel follow standard operating procedures; AI agents can also do this. While financial institutions define these processes, Gradient Labs automates them.
While some think they can simply enter data into ChatGPT and sit back and wait for a solution, Masin said it’s not that simple. This 75% is based on successfully handling the “uncertainty phase” where intent is clearly defined.
“This is very important because this is where the customer finds out if you truly understand their question.” You bought the machine.
While older demographic groups are less likely to be satisfied with agent customer service, Masin believes the rest will be relieved as these processes improve. Even now, 50% of companies do not disclose when a customer is talking to AI.
“At nearly every one of these companies, the AI agent has a higher customer satisfaction score than the best human interviewer.” Masin concluded his words. “They get their answers much faster and are more accurate.”
Gradient Labs in action: Helping a European bank go from 10% to 75%
Gradient Labs works with a European bank with 10 million customers and a digital-first customer experience. The bank’s product list includes savings, investments, retirement, current and business accounts, as well as subscription tiers.
Scaling customer support was challenging, but before turning to AI, the bank wanted a platform on which it could build and code its own rules.
The number of support requests across these many products and layers was in the thousands. To achieve accuracy, the bank needs to share its internal knowledge base of more than 1,200 articles outlining many different service aspects. Banking staff added additional internal notes not mentioned in the articles.
Gradient Labs then added thousands of human agents that record customer interactions. This added context and another 700 reference points.
“The result of this work is a representative who can handle the breadth of the bank’s product portfolio from day one, without requiring months of training or a lengthy preparation period.” states one case study. “This set the stage for the bank’s first successful use cases and rapid scaling.”
The bank added all of Gradient Labs’ finance-specific protections to ensure stringent compliance and quality standards. They cover everything from rapid injection detection and vulnerability detection to financial advice detection, sensitive information handling and complaint flagging.
Each conversation is tracked in real time against pre- and post-response guardrails so each response meets compliance standards. The AI agent had to meet the 95% standard; It scored 98%. The bank has expanded the use cases it allows the AI agent to manage.
