At a time when generative artificial intelligence (AI) solutions on lines of ChatGPT, Google Bard, etc, are gaining momentum across industries, Dell Technologies’ CTO John Roese believes that the implementation of these models by enterprises within their systems will remain slow.
This is because the default behaviour of the companies suggest that they want to focus on key strategic projects with regard to generative AI and that will be lesser in number. Therefore, before deploying an enterprise-grade generative AI system, the companies would first want to understand the potential of these solutions beyond a chatbot kind of model at present, and whether it will generate meaningful return on investment (ROI) for them.
“The issue is this technology is resource intensive. It is costly to implement. It is a strategic investment and no company can implement 360 strategic projects at the same time,” Roese said at a media interaction on Friday.
Further, owing to the complexity of these projects, the implementation of generative AI has been in trial mode for a large section of companies globally.
“We are going to see a slowdown (in enterprise generative AI implementation) not because we don’t have a lot of ideas, but because these are large projects,” Roese said. “These are significant strategic programmes and so we are clearly going to see the actual number of projects being worked get smaller,” he added.
Even as the usecases for enterprise generative AI will continue to expand, the default behaviour of most companies will be to work on handful of important projects, according to Roese.
With regard to early usecases of generative AI, currently enterprises are using a chatbot interface on top of an enterprise information base. These chatbot solutions are being used for customer support and for having a better communication interface on top of important data that was previously hard to get access.
However, beyond chatbots, the companies are seeing generative AI usecases for automating the lifecycle of the selling process, among others, but that is mostly in the experimentation stage for a large number of companies.
“The early use cases tend to be putting a better communication interface on top of data that’s hard to access, but there’s much more than that in the generative AI discussion. Our assessment is that customers are going to prioritise and pick one or two solutions and then make them real in the next year,” Roese added.
In India too, several IT companies are working to build specific generative AI usecases for their clients. However, the take-up of the solution has been slower and mostly in the experimentation stage.
“What we are finding is that 2023 was mostly a year of experimentation. For larger companies, it was prioritisation as they picked which area they’re going to focus on and bring that into production in 2024,” Roese said.
According to Roese, none of the companies so far globally has been able to implement generative AI solutions in full.
