Imagine walking into a vast library, with information about anything you might need at your fingertips.
The library is very nice – who doesn’t love floor-to-ceiling bookshelves – but for it to be as useful as possible, you need to be able to quickly find what you’re looking for: the books must be well organized, it needs to be staffed by knowledgeable librarians, and it requires a search tool for looking up information.
So it is with Knowledge Management Systems (KMS) platforms at TD. KMS platforms are a bit like the library lookup system of a large organization — they are pieces of technology that capture, organize, and retrieve knowledge within a given entity.
The Bank has lots of knowledge and information that make up the "books" of its library – think detailed policies and procedures, analyst reports, historic financial data and more. Powered by cutting-edge AI, a KMS platform can quickly sort through all that information and provide the relevant response to a query, whether it's finding the correct answer to a customer service call in a timely manner or synthesizing information requested by an analyst on a particular stock or industry.
The Bank can then deliver that information in a user-friendly format. A virtual assistant or chatbot, for example, can help a colleague (the librarian in this metaphor) efficiently connect a client with the information they need, especially in response to specific policy points. Chatbots, an example of a KMS, work by sorting through vast amounts of information and providing answers to specific prompts, thereby supporting the user.
Developing KMS platforms to seamlessly connect with that information could help transform the customer experience, enabling TD colleagues to focus on tasks like spending more time with customers.
Maks Volkovs, the Chief AI Scientist at TD, oversees AI research and development at Layer 6, the Bank’s AI centre of excellence. And he sees KMS platforms as a promising way to help transform the banking experience at TD.
"It's a very exciting time right now because we're just starting to unlock the value in KMS platforms," he said, adding that with advances in AI technology, like the Bank's recent launch of TD AI Prism (a new predictive AI foundation model), TD now has the potential to reach a new level of productivity.
"Our teams are rapidly iterating on this technology, to implement and drive it in a way that maximizes its value and the experience for our customers and colleagues."
There's a lot of work put in to unlock that potential for customers and colleagues. It has included identifying what parts of the Bank have the best potential for KMS platforms, sorting the Bank's unstructured data so that that computer programs can easily summarize and make it available when needed, and continually working to improve accuracy of results provided by the KMS platforms.
Now TD is putting that work into action.
How some TD colleagues are using AI chatbots
One way is through the TD Securities AI Virtual Assistant, currently being piloted among some TD Securities colleagues. The TD Securities AI Virtual Assistant is a proprietary generative AI-powered chatbot, a type of KMS platform, designed to synthesize vast amounts of information into concise, context-relevant summaries.
“Colleagues are constantly dealing with a firehose of information,” said Dan Bosman, Senior Vice President & Chief Information Officer, TD Securities and Payments.
Where the virtual assistant comes into play is by filtering and organizing all this data to help meet client and business needs.
"It's a great tool to summarize relevant information" Bosman said, describing how the tool can go through hundreds of PDF reports to collect just the relevant reports that provide the required information, accessed in one place.
He went on to say that the tool can help colleagues prepare research for meetings, save time on routine tasks, and answer questions that would previously need a dashboard, data scientists, or in-depth financial analysis.
"It's about focusing on what really matters and ignoring the noise. The objective is to build deeper relationships with our clients and get to the heart of relevant information. That helps us deliver the best outcomes possible for our clients," Bosman said.
Working through the queue
Improving the customer experience is also the focus for Mushtak Najarali, Executive Vice President of North American Customer Operations (NACO).
Within NACO, colleagues field close to 60 million customer phone calls a year, and sometimes finding the right answers to questions can be an involved process. One of the goals of the team is serving customers in a timely manner, and that's where a KMS platform comes in.
"With the help of Layer 6, we piloted a virtual assistant to help colleagues access answers for customers quickly, and the early results were promising," said Najarali.
TD piloted the program with TD EasyLine. Early estimates suggest that the virtual assistant will be able to support 450,000 to 900,000 EasyLine customer inquiries annually, getting relevant information to colleagues so they can answer customer questions more quickly.
"Our goal is to use our AI solutions to help make life simpler, more accessible, and faster for everyone involved," said Najarali.
KMS platforms have shown so much early promise that the Bank has plans to roll out chatbots to additional lines of business by the end of 2025. That will add to the more than 60 AI solutions that have been implemented at the Bank.
Najarali added that as TD scales its KMS platforms with cutting-edge technology, the Bank remains focused on staying true to its roots.
"Our goal is to equip our colleagues with the latest technology, enabling them to provide timely and informed advice,” he said.
"That's in line with our values at TD, including thinking like a customer and acting with speed and impact. Our KMS platforms must be grounded in those principles too."
For Najarali, meeting that standard is the test of a positive banking experience, and it is made better by providing colleagues with organized and structured knowledge at the fingertips, just like finding the right information in a library.