In a Nutshell: Financial enterprises looking to enhance engagement, generate leads, and increase conversions often turn to the CogniCor CIRA platform to automate customer communications. And, as consumers put new stresses on call centers, CogniCor’s AI-based solution has become increasingly popular. The company’s inference-driven CIRA knowledge graph brings context sensitivity to customer conversations as well as learning and adapting over time. Its pre-built templates and sophisticated information ingestion capabilities also allow CogniCor to get business continuity channels up and running for clients in a matter of days, not weeks.
Successful communication is a two-way street, and both parties in a conversation need to understand each other before they can accomplish anything. When they can’t, they may as well be speaking different languages.
That’s why technology from CogniCor plays such a vital role in bringing the efficiencies of conversational AI to organizations — especially those in the financial space. Since 2018, global wealth management, insurance, and retail banking firms have turned to CogniCor and its CIRA digital assistant platform to automate interactions with customers and increase the efficiency of their representatives.
CIRA achieves its understanding of the wealth management and fintech domains by entering more than 200,000 enterprise documents into a knowledge graph. The platform then enhances its understanding with large bodies of proprietary client knowledge, demonstrating a high level of conversational awareness at deployment.
And when customer conversations enter the picture, CIRA’s inference-driven learning framework continues to build context, knowledge, and recognize patterns to generate near-human responses to complex questions.
That combination helps financial enterprises reduce call center pressure and route requests more productively while also satisfying customers.
As Founder and CEO Sindhu Joseph pointed out, with consumer work and home patterns changing quickly in response to COVID-19, CIRA’s modularity becomes even more important as firms seek to promptly establish continuity plans and communicate vital information to users.
“In a situation like this, we can be instrumental,” Joseph said. “Within 72 hours, we can set up an additional channel and make sure customers are getting what they need — even without talking to a human agent.”
Templates and Integrations Enable Turnkey Deployments
Digital transformation is a fact of life in the financial space but call centers are always a point of focus. With at-home customers still needing help managing their accounts, firms that can turn a customer call into a positive experience have an advantage.
“Most call centers are stressed these days because some agents can’t come to the call center and attend the calls — so they have to make sure there are enough people to provide the regular service,” Joseph said.
That’s why onboarding with CIRA can be like a breath of fresh air. The platform’s interactive user interface walks developers through the process of building an assistant in an environment that encourages team-based collaboration, parallel development, and component reuse.
Domain knowledge arrays in a set of templates containing pre-loaded intents. Developers then populate the system with institutional knowledge using a unique document ingestion process that automatically creates intents and educates the assistant.
Most of the time, they enable integrations through a connection with a content management system (CMS). And CogniCor works with all the major CMS providers.
“You have these different flavors and varieties,” Joseph said. “We try to pick up the information from there to put in our knowledge graph.”
Some CogniCor clients, especially those in the wealth space, run back-office operations and desktop and web services that don’t have proper API connectivity. Fortunately, CogniCor has a solution for that, too.
“The good thing about our technology is that we can work around many situations,” Joseph said. “Sometimes, you don’t need to connect to all these back-end systems, and you can still roll it out when you’re ready.”
Leveraging Inductive and Deductive Reasoning to Recognize Context
CIRA maintains knowledge through hierarchical trees and version controls. From there, teams design workflows to streamline and automate processes.
The AI and machine learning capabilities underlying CIRA differ from more linear solutions. Its combination of inductive and deductive reasoning enables a CIRA-based assistant to understand the customer better.
Data-driven observational reasoning combines with fact- and logic-based contextual reasoning to create the kinds of complex responses humans take for granted in everyday conversations. And the CIRA knowledge graph enables that.
“Many so-called chatbots are very systematic in terms of rolling out capabilities,” Joseph said. “They may train for one skill — like, ‘What’s the balance on my account?’ — which might be the only one the system knows.”
Systems typically use machine learning algorithms to gradually internalize variations for how users might ask the account balance question. Once they’re trained on a variation, a workflow that connects behind the scenes to the banking system can then pull the account info and show the balance to the user.
“That journey is defined step by step in a workflow system,” Joseph said.
While those systems may get the engagement part of the problem right, they often fall short in terms of contextual knowledge. If a customer asks a question that falls outside the system’s purview, the exchange may be frustrating.
“Business understanding isn’t really built into those systems, and that’s what differentiates CogniCor,” Joseph said. “We create almost a digital twin of the business in the form of the knowledge graph. That means, from the start, we understand every product you sell and what’s important about those products — including their characteristics and features.”
CogniCor: Call Center Efficiencies for Financial Providers
When customers interact with a system enabled by the CIRA knowledge graph, they receive unique responses — including ones the system hasn’t been programmed to provide. That provides organizations with the support flexibility they need — especially when those systems are stressed.
“There’s a lot of flexibility to understand the context of a question and provide a response within that context,” Joseph said. “For example, if I ask about my interest rate, it knows to ask me back whether I’m talking about my savings account or my checking account — just like a human agent would. Every feature ties back to some node in the knowledge graph.”
That makes call centers and help desks more efficient. And when human agents are allowed time to be more productive, they can deliver on specific ROI goals.
For example, one CogniCor customer — the second-largest retail bank in Singapore — uses CIRA as a lead-generation platform.
“Every year, we help increase their revenue by $100 million,” Joseph said.
Another customer in the insurance space takes advantage of the platform’s sales-automation capabilities.
“We sold around 60,000 insurance policies per year without any human in the loop,” Joseph said.
Meanwhile, CogniCor continues to incorporate new capabilities into the CIRA platform. And that helps the company realize its goal of building a machine that converses naturally with humans.
“And companies don’t have to build it from scratch,” Joseph said.