Customer First: Business Growth with Generative and Conversational AI | AI in Action 2023

I am Ragu Nuala. I'm co-founder and CEO for Yellow ai. We are an enterprise conversational AI company, and we have been enabling thousand plus enterprises across the world over the last six years in automating their customer and employee experiences.

We are seeing a phenomenal change in technology where generative AI has become a core part of what enterprises are discussing as their future strategy. And today I'm going to talk about how generative AI can be customer first and how companies can incorporate a generative AI as a core strategy to drive customer growth, customer satisfaction, and eventually the business growth. And bottom line yellow AI is a global leader in conversational ai.

Uh, we process more than 12 billion plus interactions every single year, and this is enabled by our deployments on both chat and voice. So people call onto telephone lines and our virtual assistants automate those interactions. We have our consumers interacting on messaging channels like SMS WhatsApp website, mobile app, and using a multi LLM architecture drive interactions of these 12 billion plus conversations every single year.

We are enterprise, uh, great security with ISO hipaa, GDPR, and uh, SOC two, and we have some of the leading healthcare banking financial services customers deploy our products to enable their core transactions on conversations that proves our enterprise capability. We are recognized on the Gart Gartner Magic Quadrant. Uh, we are recognized as leaders on conversation AI by Opus Research, and most recently we have been recognized as the fastest growing conversational company, uh, as a part of the Deloitte Fast 500.

And overall, we are 30th fastest growing company in the Bay Area. We are deployed across multiple countries, and yellow AI supports 1 35 plus languages and over 35 plus, uh, channels. Uh, we have industry specific vertical functions within the platform that support vertical industries like healthcare, banking, uh, insurance, retail, automotive, et cetera, across 12 plus, uh, industry verticals.

Uh, uh, we are proud of bringing AI to customer experience over the last seven years out there in the market. With that about the company, uh, we have seen the entire movement and evolution of the space from 2016 till now. And looking beyond this provided us, uh, a in-depth insight into what are the key hurdles in companies providing seamless customer experience.

Uh, it isn't a surprise that a significant of our calls to customer support end up in really long wait times. There are about 400 billion plus customer support calls, uh, that have made every single year. And, you know, there are 15 million agents, uh, you can easily imagine how most of them end up, uh, in hold for a really long time.

We also, the companies also limited by the high agent attrition and workload, kind of, uh, getting into more inconsistent experience for end customers while dealing with customer support. Also, there is inconsistency, uh, based on the channel in which, uh, customers reach out to the companies. Uh, they reach out, uh, on phone, they get a different experience, they reach out on website, they completely hit a different person, and they can get, get a completely different experience.

So it's not like really very consistent and it takes a long time to even resolve questions as the agents have to scroll through multiple applications, get the data required, and actually satisfy and full the fulfill the customer request. We tried solving that, uh, over a period of time with, uh, the evolution of automation right from, you know, 2011 to 2016 and beyond. Uh, there's been a search of, uh, assistance on consumer assistance like Alex, uh, Siri, Google from 2011 on top of our phones.

And people have been used to get automated voice assistance, uh, for their queries, but they have not been super effective. They've been fairly generic based on global knowledge, et cetera. But 2016 is when the really, uh, enterprises and businesses started this adoption of, uh, chatbots, which were very basic at that time.

But, uh, when Facebook Messenger and WhatsApp opened up these platforms for businesses, uh, they were, uh, the businesses were starting to automate some of these interactions using informational, uh, chat bots in 2016, which saw limited success. I think companies were able to provide information, but um, they were not really completing the transactions, et cetera. We, as a company, started in 2016 and, uh, we pioneer the implementation of core ML and language models techniques in deploying chat bots.

And we saw success in the industry where companies were seeing reduced, uh, uh, number of calls in the contact center include satisfaction rates. Um, and from 2016 till 2021, um, there've been models that have been evolving and MO moved from informational chat bots to transactional chat bots as well, which provided significant benefit to the customers in terms of automating their booking, automating their service requests, automating, uh, claims, et cetera. But beyond that, uh, generative AI has brought together significant improvement and significant uplift in the kind of conversations that, uh, that conversational AI companies can deliver to the end customers.

It did not stop at providing information, it did not stop at providing transactions, but also making, uh, decisions on behalf of, uh, the customer support agents or on behalf of the enterprise, uh, reasoning with the end customers and driving empathetic interactions that have catapulted the automation rates from 30 to 40% that we were getting in the pre generative AI era to what we are seeing at our customers, close to 90% plus automation rates for customer support. And we believe the future lies in fully autonomous customer support, where the agent's role will be replaced by AI operators or bot assistance. Essentially, the, the time spent by contact center agents is in fine tuning and training the AI system rather than directly answering the calls.

And this is the future which we are accelerating towards, where we see about $80 billion plus of contact center cost reduction while pushing up the, uh, CSAT scores and customer customer satisfaction scores significantly higher. And yellow AI is, uh, playing a core part, uh, in this evolution, uh, with over a thousand plus enterprises adopting our platform and with over 12 billion plus interactions that we are automating every single year. Uh, so the biggest change that is coming from the core chat bots to generative AI based, uh, conversational assistance in is in addressing the four biggest problems, uh, that we as companies, uh, solve with the traditional intent-based chat bots.

One is, um, the accuracy and reliability. Those were completely being dependent on the coded workflows in the chat bot, um, uh, development platforms. And with generative ai, this eliminates the need to limit to the, uh, workflows that are defined in the platform and can extend to a very wide knowledge base.

Uh, so that is solving the problem of accuracy and reliability. Uh, the second one is the integration and, uh, scalability. Uh, the traditional platforms required specific hardcoded integration to the core workflow systems and generative AI right now enabling through plugins the ability to create workflows on the go.

So the, the generative AI based, uh, virtual assistants can dynamically create workflows and solve for the situations and queries that are not pre-configured. The setup times have come from few months to actually a day or minutes where with generative AI companies are now able to create chat and voice assistance by just pointing to the content and creating them in their second. Um, uh, a good example of that is our offering eulogy that, uh, we have launched six months back with, with tens and thousands of companies, uh, that are able to create this chat assistance with a single click of a button by pointing to the source content.

The biggest problem also has been around mechanical responses, where the traditional conversational AI infrastructure, uh, required companies to quantify the kind of responses that they're providing to the end users. And with generative ai, these, uh, generative AI virtual assistance just need to be trained on the historical content, and they actually generate the responses that are in tune with the customer and the kind of prompt that they are given, rather than someone mechanically, um, coding these responses. This has left to, uh, led to a significant shift in, uh, the customer experience.

Uh, we are more and more able to manage, uh, complex conversations that require empathy, that require, uh, reasoning capabilities that require on the fly mathematical abilities, and, uh, that, that require, uh, intense, uh, human-like decision making. And we have, uh, we have enabled chatbots that provide human-like empathy. Uh, to give an example, someone looking, uh, to cancel a flight ticket, uh, because they were getting at a, a better rate in a different website.

They need to be talked about in a different way by, you know, offering a better price or offering a better discount and retaining the customer versus someone coming and canceling because they had a health issue. You have to really show empathy out there and, you know, make it easy for them to cancel the booking. So ability to understand these emotions and provide human-like interactions, uh, is, is something that yellow AI enabled with, uh, genetic AI and LLMs as the core.

The time to value, uh, has been really fast where it took months to now days in, in, in, you kind of launch a chat bot or a virtual assistant and you start automating your customer support calls from day one, and you can go on and integrate and, you know, uh, get it to more depth, but you start to see results, uh, within days instead of months. And, uh, enabling agents to be super productive where agents are right now training the system. So a single agent can probably, uh, manage, uh, you know, a customer base of few thousands versus probably managing a customer base of few hundreds, uh, historically.

So there is a clear amplification of the productivity of the agents with, uh, generative AI powered, uh, virtual assistance, uh, through LO ai. What we have seen, uh, with our customers to take an example, rather than, um, a staffing company that has come to us with an expectation of driving 30 to 50% automation for the broad set of payroll queries for their, uh, workers. And, uh, we were pleasantly surprised that with generative ai, they were able to get to 90% of automation with almost zero touch.

Um, uh, and, and that too within a matter of 30 days, it just took two weeks to get, get it up and running, and within a month, they were able to reach 90% plus automation. This is purely possible with the multi LLM generative AI model, uh, within the ALO AI ecosystem. You still need to have the humans in the loop to take care of, uh, queries that these are systems are not able to answer.

And the human loop provides a great mechanism not only to answer the customer queries, but also provide, uh, training back into the system so that the automation and the AI keeps improving based on the responses generated by, uh, the agents. This is also helping our customers craft end-to-end customer journeys and not just, uh, focus on support. Every support conversation or every interaction with the customer is an opportunity to, uh, talk more about the brand, uh, cross-sell, uh, upsell, um, and make a note of their preferences and update their customer models.

So, um, companies have started to look at customer interaction, not just from a customer support point of view, but use AI across the entire customer lifecycle in crafting beautiful and delightful customer journeys. We are here to work with some of the largest enterprises, and we are on our mission to make the 400 billion plus customer support calls happening across the world more delightful and impactful through AI and automation. And looking forward to, uh, hearing from you.

I'm here at, uh, rag at Yellow ai and I'll be super happy to receive any questions and responses and eager to work with you all. Thank you.