Arguing that AI encompasses the creative analysis and production of intelligent data, and that data is passed to the software stack to be processed, stored, checked, formatted, presented, or actioned in a deterministic fashion, HSBC said AI is destined to be subordinate to the overall software platform within a full-blown enterprise application.
“In large enterprise application platforms, we see subordinated AI (inherently non-deterministic; think errors and different answers to the same question) as simply a part of an overall solution and it is likely to be subjugated across applications by embedding intelligent agents within the numerous enterprise platforms. These intelligent agents will likely be relegated, in a targeted fashion, to domain-specific high-value tasks that are difficult to codify and are better handled by a thinking entity,” HSBC analysts said in a report.
“Thus, we believe that the vast majority of enterprise software will not be threatened by AI, but rather AI will be domesticated within the application stack via agents and will create tremendous value in doing so.”
Major software vendors like Salesforce, Oracle, ServiceNow, and Microsoft have been doing the heavy lifting of designing, coding, and beta testing embedded agents.
“We think that most started in 2024 and are now at a point of maturity where these embedded agents are being rolled out to their large global customers. We see this as the primary mechanism from which the world’s largest enterprises will consume AI and, in doing so, drive the consumption of AI inference by orders of magnitude, justifying the tremendous buildout of AI data centres across the globe. In our view, this is the primary consumption track and how AI diffuses across the world’s largest enterprises,” it said.
HSBC analysts also argue that companies using AI to write their own applications is not practical, realistic, or economically sound.”However, we have identified a party that is best suited to use AI to generate software that is better than existing legacy vendors. And, of course, that is the software vendors themselves. They have domain expertise, robust sales channels, trusted relationships, cross-licensing, and legions of coders that are now more productive at creating new code by using new AI apps and tools. Finally, there are many factors that make up a successful software company and coding is just one factor. Displacing an incumbent software vendor is very difficult, even with a better or cheaper offering,” the brokerage said.
The analysts see legacy enterprise software vendors as among the key beneficiaries and diffusion paths for unlocking the value creation potential of AI within the production of the USD 100 trillion global GDP ecosystem.
“Sector valuations are at historical lows, even though we believe the sector is poised to expand massively and we see strong demand momentum lasting for the foreseeable future,” HSBC said.
(Disclaimer: The recommendations, suggestions, views, and opinions given by the experts are their own. These do not represent the views of The Economic Times.)
