Kasisto – A Conversational AI Platform for Banking
The term “conversational AI” has a much better ring to it than the term “chatbot“. One implies an interaction that’s useful and informative while the other implies a forced interaction which leads to a frustrating experience. We were recently reflecting on how easy “conversational AI” can be implemented in the world of finance because things are usually straightforward when it comes to how people interact with money. What we should expect to see soon is that the majority of client interactions will be funneled towards chat because it’s far easier to put a “human in the loop” when needed. An interaction might happen like this:
- HUMAN – <Clicks link to chat with customer service>
- AI – Hi, this is John. How can I help you?
- HUMAN – Hi John, I was looking at the $60.30 charge from last week on my credit card at 4-Hands Thai Massage in Bangkok and had a question about which forex rate you’re using. The charge seems high and maybe it’s the forex rate. – (Notice that as the human is typing, the AI is parsing what is being typed for tone and context.)
- AI – One second, let me take a look.
- HUMAN – Ok, thank you
- AI – Looks like you were charged 1,953 baht or $60.30 USD which means the exchange rate used was 32.39 baht to 1 U.S. dollar
- HUMAN – Ah, maybe I did ask for extras. But I just looked up the forex rate on that day using my Bloomberg terminal and you morons are charging me something exorbitant. – (At this point in time, AI knows that the human is upset and that the human is questioning the forex rates for a particular transaction. Immediately the query can be summarized and sent to some who specializes in forex – seamlessly – so that a human takes over immediately.)
That’s the whole point of “conversational AI”. It’s able to handle a customer query using plain English and hand the problem over to a human when it needs to. One company that’s building conversational AI is Kasisto. (Note that Kasisto always lets the customer know when they are chatting with a chatbot and when they are chatting with a human.)
Kasisto – Conversational AI for Banking
Founded in 2013, New Yawk startup Kasisto has taken in $28.5 million in funding from a slew of investors including Wells Fargo, Mastercard, DBS Bank, BBVA, and Harvard Business School – along with loads of venture capitalists as well – with the intent of “enabling companies to engage and transact with their customers through intelligent conversations, anytime, anywhere.”
Update 09/03/19: Kasisto has raised $15 million in additional Series B funding to expand adoption of its services to all segments of the financial services industry. This brings the company’s total funding to $43.5 million to date.
Update 02/05/20: Kasisto has raised $7 million to meet the increasing demand for their products in existing and new markets as well as deepen partnerships with existing customers. This brings the company’s total funding to $50.5 million to date.
The conversational AI is tied in with the bank’s back-office systems so that it can resolve queries 82% of the time without needing a human to be involved:
Just imagine how much manpower these banks are saving, not to mention that the chatbot will always do a better job than humans while all the while getting better over time as it’s constantly being trained with new data.
Kasisto’s Platform, Kai
The technology behind Kasisto’s platform, KAI, comes from a firm called SRI International which happens to be the same outfit that created Apple’s Siri. In its earliest form, KAI consisted of MyKAI, a consumer banking bot built with KAI Banking. The idea behind MyKAI was that users could link their financial accounts to a really cool free chatbot that would tell them insightful things – like you’re spending too much money on Thai “massages”. Fast forward to today and we see a complete banking platform with all the bells and whistles:
As you can see, the Kai platform is omnichannel. Information is not only being analyzed in text-form, but also from voice conversations. All that big data being gathered is now being used to sell clients things they need, identify fraud, and gauge people’s intent based on their habits. It’s only a matter of time before these banks cannibalize all those “new credit score” startups that desperately try and get you to link your checking account so they can extend you credit to buy things you probably don’t need.
Big Data and Personalized Customer Service
Yes, AI algorithms are becoming an indispensable part of running a banking operation in today’s day and age. DBS Bank claims that they use one-fifth the resources that a traditional bank does when it comes to operations, thanks to tools like Kai which can answer 1,178 unique questions specific to DBS digibank products. Kai has also split their platform into two verticals; consumer banking and business banking. J.P. Morgan recently announced that KAI Business Banking will run the AI engine for their treasury customers – a business that moves $5 trillion for corporations every day. The platform now operates in 6 countries, speaks 8 languages, and rumor has it that soon they’ll be able to understand why women spend so much money on shoes. The amount of big data that’s being generated every day on the platform has a snowball effect on how much more capable it becomes over time.
Then there’s this notion of personalized interactions. We’ve talked before about how today’s marketing seems to be moving towards personalized echo chambers where you hear what you’re comfortable with. Kai offers a conversation manager which lets you tailor conversations – in their words:
Your Platinum customers won’t get college account conversations, your Alexa conversations are designed for voice interactions, and mobile conversations include graphics and emojis.
While emojis and memes are tools used by tools who are unable to express themselves using the languages that humans have used for decades to communicate with, at least it’s comforting to know that banks are able to “connect” with those crazy millennials now – and sell them more stuff.
Then there are all the dashboards that can now be created to see how you are interacting with your customers. If 82% of your customer interactions are completely automated, just imagine how much you can learn from all that big data. Being able to understand a customer’s spending habits tells you everything you need to know about them.
We’re not sure what it’s like today, but it used to be that big banks like Morgan Stanley almost always chose to build vs buy, spending massive amounts of money on technology investments. With platforms like Kai, the algorithms have such a head start having been trained for so long now that it makes almost no sense for a bank to build. Seems like the best way to build a platform like this would be to offer the tool for free, then collect all that big data – generate from 20,000 different financial institutions – and use to train algorithms which you would then remove from the public domain and sell to banks as “conversational AI”. While Kasisto is hardly the only player in this space, it seems like they’re the smartest so far.
Tech stocks are volatile investments during the best of times. Here at Nanalyze, we complement our tech holdings with a dividend growth strategy that performs extremely well during recessions. Find out which 30 dividend growth stocks we're holding in our report - Quantigence - A Dividend Growth Investing Strategy - freely available to Nanalyze Premium Subscribers.