A virtual assistant for lenders' employees
David Karandish, the CEO and co-founder of the artificial intelligence chatbot startup Jane.ai, has rebranded the company and taken it in a different direction.
Jane.ai on Wednesday was renamed Capacity. It still offers virtual assistant technology, but the tech is now intended primarily to solve a different problem: helping employees, and not just customers, find information.
“We are moving from being primarily known as a chatbot that answers people's questions, although we're still doing that, to a platform for helping to automate the key aspects of the workplace,” said Karandish, who was also the founder of Answers.com.
At least one financial institution, West Community Credit Union, is already using the technology for both purposes. (More on that later.)
Capacity, based in St. Louis, is also announcing $13.2 million in Series B funding from a network of private and angel investors in the Midwest. Karandish described it as an anti-Silicon Valley approach to fundraising that is “prioritizing local and regional investors who share our vision and Midwestern pragmatism versus the most common names on the coasts.”
- Karandish will speak at BankAI 2019.
The Capacity AI engine trawls through documents, applications and knowledge bases to gather answers to questions. It has connectors to applications like Salesforce, Microsoft Office 365, and Ellie Mae Encompass. Its data mining technology draws data out of documents, spreadsheets and web pages. It also has an expert finder that helps identify people who are knowledgeable on specific topics.
It can also become part of existing workflows.
“If you've got a big, messy offline process, like for example, processing a mortgage loan, we can represent that process through a digital workflow, and then over time, parts of that workflow, instead of being assigned to individual people, get assigned to the Capacity bot itself,” Karandish explained. “So we wanted to rename the company to reflect not only who we are, but what we're doing for our customers: helping to free up their capacity, to help them do their best work.”
Credit union puts tech to the test
West Community Credit Union, of O’Fallon, Mo., was looking to integrate artificial intelligence into several areas of the company, according to Jason Peach, the credit union's CEO.
He and his team met with Karandish and decided to deploy the Jane chatbot on the credit union’s website, to help visitors quickly get what they need.
“When people come into our retail branches, there are always people there to greet them, work with them and help them find the information they need,” Peach said. “We realized more people visit our website than actually come into our physical branches. And when people land at our website, there was nobody there to, for lack of a better term, greet them.”
The chatbot accepts questions written in everyday English, interprets them and supplies answers. It obtains its answers by ingesting the contents of the website and the documents to which it links. As customers ask new questions and staff type in the answers, the chatbot gathers new knowledge. It can now answer 3,000 questions.
The credit union had another motive, too: it wanted to better understand why people were visiting its website and what they were interested in.
“We always monitor what people do on our website and how the traffic flows [using Google Analytics], but just because somebody goes to a particular page doesn't necessarily mean you can track down what their specific question or need is,” said Koren Greubel, vice president of marketing at West Community. “The analytics we're getting from Jane are educating us on why people are going to our website, which helps us make better pathways and linkages and use more understandable verbiage.”
For instance, she found out that many customers were unfamiliar with the term “mobile endorsement.” They did understand, “sign the back of your check for a mobile deposit.”
“We've taken the technical speak out of it and put it into more of a consumer-friendly vocabulary,” Greubel said.
The software produces weekly reports on all the questions people ask, segmented by topic, and the chatbot’s answers, so Greubel and her team can evaluate how well it did.
Call center agents were asked to use Jane and help answer questions posed to the chatbot.
Then unexpectedly, staff throughout the organization, including in the retail branches, began turning to the chatbot when they could not remember something about the credit union’s products or services, rather than use the intranet it had created for that purpose. Peach plans to integrate Jane into the company intranet.
Later, the credit union may integrate Jane with its newly implemented Encompass mortgage platform.
An unanticipated benefit of the Capacity technology is that it can take the place of time-consuming data integration projects, Peach said. Connecting the Capacity technology to the different data sources allows them to all be accessed from one point.
“A lot of financial institutions spend a lot of money just trying to get their data organized and accessible,” Peach said. “It occurred to us that maybe this could allow us to make the data accessible without having to do all the technology and infrastructure changes. This was an ‘aha’ moment for us.”
Poised for expansion
Sherry Comes, managing director in conversational artificial intelligence at Deloitte Consulting, has implemented similar systems, which she calls “expert Q&A systems” or “knowledge gardens” that can access structured and unstructured data.
“It’s a way to get your arms around multiple stores of structured and unstructured data,” Comes said. Such systems rank the answers they generate by confidence level.
“You can even have a human in the loop if your answer doesn’t have a high enough level of confidence,” she said. “You can have the human tell the correct answer for that question and then the next time the machine knows that. So it's continuously learning.”
Lately, Comes has seen a lot of financial institutions using this kind of technology for debt collection, which is a stressful task, she said.
“Think of it as a collections assistant — somebody who's going to help you pay or inquire about a bill. There's a lot going on. Chief financial officers are all over this.”
She also has seen heavily regulated companies like banks and insurance companies using chatbots like Jane to help with compliance.
“Every year comes new regs and the knowledge burden changes,” Comes said. Plugging the latest changes into a chatbot lets a company push them out to customers and employees, providing consistency. It also helps with compliance auditing, she said, recording and tracking how regulatory changes are communicated to customers and employees.
In contact centers, such technology helps lower attrition, Comes said.
“CEOs will tell me attrition is going down,” she said. “When you ask the agents, they'll say, ‘I no longer get yelled at like I don’t know the answer. I do know the answer.' ”
The future of such systems, in Comes’ view, is voice interaction.
“Voice is the only interface that everybody knows how to use,” she said. “Everybody's going to want a voice solution — at home, at work, in the car.”