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

During eGain's annual Solve conference in Chicago this year, Phil Britt had a chance to sit down with Ashu Roy, the company's chairman and CEO since 1997, to discuss the future of artificial intelligence and where eGain figures into conversations about the technology.

CRM: How has AI technology evolved over the past year on a macro level, and how has it evolved for eGain specifically?

Roy: There is acknowledgement of the fact that as a society we don't need AI to reach a human level of intelligence, which had been a big thing the previous year. This is very good awareness because it opens up the whole space for how you apply AI to drive efficiencies, and in many cases, to drive creativity.

Secondly, the cost of AI inferencing continues to drop rapidly. That has enormous implications for how it gets used, because if it's inexpensive, it gets used in all kinds of places.

Thirdly, AI model builders are now building hybrid solutions; they are enclosing specifically built machines inside their models. Some are memory-based, some are compute-based. The effect is far more intelligence, where as last year, the general sense was to build bigger and better models and eventually that would overwhelm the problem of any complexity.

CRM: You mention that the cost keeps dropping, yet many companies are spending millions of dollars on the chips and the cooling needed to operate systems using these chips, with very little return. In the late 1990s, some telecom companies spent millions building the infrastructure for broadband but went bankrupt before there was any payout, though those that survived are doing very well. Is there going to be a bloodbath of companies investing AI but never seeing a positive ROI?

Roy: In the model space, there could be. If you think of AI as a three-layered cake, the hardware platforms, the GPU models, are the bottom later. Then you have the model builders and those building and training the models. The revenue for that is going down very quickly, but the cost of building is not going down in a big way. There will be commodization at the model so, only a few will be left standing.

CRM: Where is eGain?

Roy: eGain is in the application (top) layer. The application layer is a great place for creating value for businesses and for consumers. The capability of AI models will keep getting better and better, building systems that will solve real problems in an enterprise environment.

CRM: While AI models have improved in the past year to provide more correct answers more often, they still are known to give incorrect answers. Knowing that, why would a business invest in AI?

Roy: Users need to work with the right model and limit the scope of what the model has to look at. The infrastructure builders (e.g., Oracle) are making a massive bet on future commitments and earnings from companies like OpenAI. OpenAI doesn't have a feasible business model as it stands today, but you will see them develop a model that is quite unique, building e-commerce into it. That taps into the online commerce stream and the advertising stream. But as it stands right now, there is no [viable] business model. Some big tech companies are building their own models, but I don’t know how well that will work.

CRM: There are some economic reports indicating underlying weakness in the economy. Though much of the tech sector is doing very well right now, if the underlying economy weakens far enough, business customers, including those who buy from eGain, are bound to reduce their spending. What is your outlook for the overall economy and for eGain?

Roy: If there is a bump in the economy, which is likely given the way things are inflated, especially in the AI area, then things will slow down. But even if there is a blip, the growth in AI will continue beyond the blip because our clients are going to continue to sign up to drive efficiency. We got JP Morgan Chase as a client a few months ago. We're deep inside their systems. Their [and other clients'] drive toward efficiency will continue.

We are very happy with where we're at and what we are doing. We've built some great new products and have some great new customers. There is a lot of interest in driving CX automation with AI. We are only two years in what is a five-year roadmap.

CRM: A year from now you will host Solve '26. How will our discussion be different then?

Roy: At Solve 26 there will be a lot of stories from the use of the capabilities we introduced at Solve 25.