[Posted on Behalf of Pushkal Mishra, AVP IT & CISO HDFC ERGO Health Insurance Ltd ]
There has been a wide adoption of A.I. (Artificial Intelligence) in the last few years. You might have seen them in chat rooms or on companies’ websites providing you information about Insurance, Banking products, Customer support or smart speakers or even inside your favorite Barbie doll, now a conversational doll !
As per Gartner-
“By 2020, over 50% of medium to large enterprises will have deployed product chatbots.”
“By 2020, customers will manage 85% of their relationship with the enterprise without interacting with a human.”
Every customer service, sales and marketing platform is a potential use case for chatbots therefore, you see biggies like Facebook, IBM, AWS, Google have created AI Platform that can be leveraged for variety of Bots.
To put this in perspective, there are over 100,000 bots for Facebook Messenger and this number will only increase as more and more businesses start to realize their value, after all Facebook alone has 2.27 billion monthly active users base. Alexa (by Amazon) on the other hand, passed 56750 skills (built-in capabilities are referred to as skills e.g. playing music from multiple providers, setting alarm, answering questions, weather forecasts, and querying Wikipedia etc). Also Watson (by IBM) is doing cool things like helping banks to answer thousands of customer question everyday, smart ordering for online retailer, personalized medical information & advice, analyzing & predicting crime, in depth health & fitness insights etc.
So you see this is exploding all around us and fundamentally changing the way we engage, interact and learn things. But my intent is not to go gaga about chatbot, we understand that it's much better deal than complicated phone menu or reluctant/ill-informed service representative. The point that I want to bring to your attention is how many of us are actually making well thought through chatbots that are capable of meaningful 1 to 1 conversation at scale ? More often than not, I find one of the following issues with bots -
The sluggish response has to do with the time *NLP engine (technology that powers bot) takes to understand the intent and semantics behind the message and then, crafts the response. Poor conversation designing is the reason why bots produce less than perfect-at times not 100% relevant responses or “I don’t Understand you” error.
On the other hand, asking for more information than required can result in privacy issues and unawareness of regulatory compliance like data hosting/processing requirements could attract heavy penalties.
Now speaking from security vulnerability standpoint, just like any other program on web, Chatbots are also prone to Cyber threats. Vulnerabilities like Cross Site Scripting, SQL Injection, Dangerous Http Methods, Clickjacking expose them to variety of exploits like web defacement, data loss & leak, malicious bot injection etc. As the adoption increases, It's not surprising that you would hear more about hacked or rouge chatbots or data breaches because of them.
Comments