HSBC Trials Bot-to-Bot Interactions With Other Symphony Users

HSBC Securities Services (HSS) is seeking to improve customer experience with chatbots. Through the messaging platform Symphony, the firm has been able to facilitate a chatbot-to-chatbot interaction that is handling customer problems independently of humans.

After first using Symphony internally, HSS rolled out the new chatbot externally last year. Stephen Bayly, global head of securities services technology at HSBC, says he has seen a dramatic difference in the number of phone-related queries since the firm implemented the chatbots. “In the first month of activity, we had a reduction of 27% in the number of phone calls and manually answered queries we were receiving. And that’s just continued,” he says.

Users type questions into the chat, the bot then automatically interprets the meaning using natural language processing (NLP), and then sends them to the appropriate HSBC system via an API, with users getting a response back almost instantly.

This routine problem-solving has sparked a chatbot-to-chatbot interaction. Bayly says that some of HSS’ clients have noticed both chatbot systems interacting independently. “Some clients who are also Symphony users are trialing with us chatbot-to-chatbot interaction with no human intervention on either side,” he says.

Bayly says that this interaction eradicates the need for human intervention: “Automating these interactions enables us to provide clients with a faster and more efficient service, while freeing up our people to focus on more value-added activities.”

In some cases, a bot-to-bot exchange can be easier and produce a faster solution because the language the bots speak is structured the same way on both sides. “If it’s a human on one side and a bot on the other,” he says, “then the conversation is open to more variability, with different syntax and typos that the bot would have to navigate.”

Bayly says that as HSS enhances the service, the bots will be able to answer more and more queries “fairly routinely” through a chat. The goal, he says, is that the bot will eventually address all possible query content that does not need human intervention—HSS estimates this to be around 70% of all calls. Bayly admits that the bots may have limitations, as well as a diminishing return as the firm gets closer to that 70% threshold.

Symphony has amassed over 470,000 users, many of whom develop their own applications and bots that connect to the platform to automate workflows and enable a range of proprietary data and content to be accessed from a single interface. 

The software comes with a set of APIs that simplify the process of creating something like a bot, for developers at a bank or on the buy side, allowing them to fast track the creation of a product.

HSS can roll out its chatbot software more easily to other firms that use Symphony because they already have the same framework, says Goutam Nadella, executive vice president for client solutions at Symphony. “These bots are made on Symphony’s platform, then they interact with Symphony using those APIs, and so they can send and receive messages like users and they can react to events like users,” Nadella says.

As well as solving clients’ issues faster, chatbots can potentially eliminate the chance of a mistake or misinformation. When the HSS chatbot doesn’t know the answer to a question, the user may try phrasing it in a couple of different ways, but, says Bayly, the chatbot doesn’t “just have a punt” if it doesn’t know the right answer: the bot admits it cannot solve the problem and the issue gets passed on to a human who can.