Observe.AI's path to bringing AI in Call Centers

Swapnil Jain
May 14, 2019

What we think about the call center industry

Call Center is the first industry both academia and industry thinks will be the most impacted by AI in the near term. There is a lot of buzz about Chatbots, Conversational IVR, RPA and other applications of AI which can truly disrupt call centers as we know them.  But there is a fundamental difference in how we at Observe.AI think about AI in Call Centers vs the rest of the industry. The industry primarily talks about automation and self service which is about automating the entire customer experience. We strongly believe this is not how AI will manifest in the call centers as its first application. Automation is the end game. There is a gradual path which will lead to automation. That is the path we are on.

Inspiration from latest technologies

A good analogy to this is self driving cars. Tesla wants to build self driving cars. They want to automate the car driving experience which a human is doing today. But Tesla did not come up with self driving cars on day 1. They started with putting cameras, LIDARs and other sensors on car which was driven by a human to collect data and provide insights on the driving behaviors. Once they had enough data about how a person is driving they came up with features and functionalities  to augment that person with AI in their driving. Think auto pilot, cruise control and parking assist. Today Tesla is a great example of  “AI in the loop” driving. It still requires a human to be on the steering wheel. Slowly we will see Tesla launching completely self driving cars but we are a bit far away from that.

This is exactly how we believe AI will manifest in the call center - Analyze → Insights → Augment → Automate

With our Analyze + Insights product we are analyzing millions of calls per month to provide the right insights to call center managers, supervisors and agents to improve customer experience, enforce compliance and improve agent performance. This is only possible with AI. Manual QA team can only review 1% of the calls which is a very small set to extract any meaningful insight. This is our entry point to bring AI in a call center. A legacy process which was manual up until today.

Now that we have analyzed millions of calls we have build a deeper understanding of how agents are handling customer calls. We believe there are opportunities where AI can help an agent to provide a better customer experience while increasing their efficiency. AI + Human can provide a way superior experience to the customer than a human alone. Our Augment product listens to every calls in real time and provide the right guidance, information, next best action, upsell/cross opportunities to the agent. Imagine every agent powered with this technology. They do not need to put the customer on hold to look put any FAQ/Knowledge, they know the best way to resolve the problem, they know opportunities to upsell, etc.  Augment product forms a complete feedback loop from the recommendation → agent action → customer sentiment. This loop helps improve the recommendations with every single call our system analyzed.

Automation as an output

The biggest problem with today's automation system is they lack the understanding of what to automate and how. The current system requires a call center manager to go and type all ways customer can ask a certain question. The call center manager can barely listen to a few calls and can come up with a few ways which are obviously not exhaustive. This leads to poor customer experience. We have seen this story play-out with chatbots who have struggled to even remotely deliver on the hype. You can actually experience this by calling United Airlines call center at XXXXXXXXX.

We think of automation as an output for the Insights + Augment. The massive amount of calls we have analyzed helps us build conversation maps for the 100% of the call center conversation. The maps then feed into our automation system which powers the automation for the calls. A bot that has learned from millions of conversations on all the thousands of ways a customer says something. This forms the beginning of our automation journey. True automation where technology is learning to adopt how customers speak and behave and not forcing the technology limitation on the customers.

This is the path we are on!

About the author: Swapnil Jain is the co-founder and CEO of Observe.AI. He has a strong technical background and a passion for solving problems.

Turn every agent into your best brand representative.
Get Demo