A Q&A with Ashu Roy, eGain's CEO

CRM: During your keynote address at the Solve 24 conference you asked about people getting positive results from artificial intelligence. No one raised their hands. Yet to hear some proponents, AI is the tool for better efficiency, profitability, and cost savings. Is it? Or is it more like the over-hyped dotcom companies of 25 years ago?

<Roy: Look at it as a simple three-layer cake. The bottom layer being [graphical processing units], the next layer being models, and third layer being apps. If you look at the business value, it's going to come from apps. It's not going to come from GPUs; it's not going to come from raw models. The progression now is working its way through the GPUs. You can sell only so many GPUs. People are now using it to build custom models, leading to more sales. The next step is that language models are being built. They're being deployed as tools. People are prototyping, practicing, trying things out, and using these models heavily. Apps are a system that takes all these models and tools and connects them back into enterprise systems, into the environments to operationalize them. That takes time, and there will be false starts.>

What happens with new technology is that people expect it to really perform very quickly, show value very quickly. We overestimate the impact of the disruption from new technology in the short term and underestimate the impact in the long term. That's where people are at today: overestimating the short-term impact. But my view is that it will have a lot of value in the long term.

CRM: Your generative AI assistant is now in beta. Other companies offer their own versions of such a product. What makes yours different?

Roy: We acknowledge that this kind of an AI agent has to work with trusted content and knowledge from the enterprise to deliver real value. We do not pretend that simply driving a large language model, which may be a little bit tuned and customized, running over any content in the enterprise from any source is going to deliver great answers. The business impact of deflection and customer satisfaction, that's the part we are focusing on.

CRM: A lot of companies are offering AI as the unique differentiator in their products What is eGain doing that sets it apart from others?

Roy: When we bring content into the eGain platform, we have a library of knowledge transformation bots that can do microtasks that are specialized to create and curate knowledge for the business. We give them very microtasks, like cleaning up the structure of the knowledge content so that it is more readable as a human. This works well for gen AI tools because they are trained on web content. Web content is well-presented and is designed to be consumable.

CRM: What will companies that derive the promised benefits out of AI do differently than those that aren't successful with it?

Roy: AI isn't a differentiator. AI is a commodity. What is differentiated is how you apply it in the context of the business environment and enterprise content and data. That combination is what we think is unique about us.&

Most companies will get some value out of AI, but some companies will get disproportionately larger impact with their investments. [These are] the companies that look at end-to-end processes in their businesses and look to restructure the value chain in each function. For example, the service model for large companies is built on humans and different digital processes. You can apply AI at each of these points and get a small benefit, or you can look at the entire service function, taking a clean-slate approach for a larger total benefit.

With the advent of AI, you have to reimagine the value chain. You have to acknowledge the reality of your products, your customers, your current complexity of the business, then platform the approach to applying AI in these functions. I think we'll get much better value on the platform.

CRM: Next fall you will be hosting Solve 25. How will the knowledge management and AI environments be different then than they are today?

Roy: The knowledge practice is going to move toward more knowledge synthesis and automation. There's a trend that already started toward the ability to create and curate knowledge faster and faster. You're going to get a situation where you want to be able to test your systems, changing it faster because the change is getting easier. You need to be able to test it against questions that are likely to come up for the knowledge system to respond to. That's not very different from the way language models are trained today in the second layer of the three-layer cake, except that in this third layer, it's specific to the business, not to the entire world.

Large language model training for the entire world just takes worldwide web content. In the third layer, you're talking about training and synthesis and question/answers to human feedback for business-specific knowledge. We are going to synthesis and automation of knowledge in conjunction with AI's capabilities so that you can answer questions from whatever point they come into your organization. It will be a lot more autonomous. We may not get there in 12 months, but the direction will be clear.

CRM: How will eGain be different a year from now than it is today?

Roy: We'll have a lot more AI impact stories. At Solve 24, you saw some, but you could tell they were early in their implementations. Next year, there will be more transformational impacts. You'll see more self-service. You will see more integrated stories from contact centers than what you've seen so far, because those take some time to sort out.