Only a quarter of AI initiatives have delivered the expected return on investment, according to an IBM survey of 2,000 CEOs.
The study’s findings, published amid Big Blue’s annual Think conference on Tuesday, show that despite the hype around generative AI, enterprises are struggling to get real value from the token-spewing tech. Just over half (52 percent) of CEO respondents say their organization is realizing value from GenAI investments beyond cost reduction.
Despite this, enterprises aren’t ready to give up their dream of using AI to automate their workers out the door. The study shows CEOs expect the growth rate of AI investments to more than double over the next two years, with 61 percent saying they’re already adopting AI agents and preparing to scale them across their organizations.
Rather than the chatbots, image generators, or search engines many of us associate with the tech, AI agents seek to automate entire tasks by stitching together multiple tools, models, and data sources, with or without a human being in the loop.
As it stands, enterprise AI deployments appear quite narrow. The survey found that just 16 percent of the initiatives have scaled across the entire enterprise. Where enterprises are investing in AI, IBM says 65 percent of chief executives surveyed are prioritizing use cases based on their potential return on investment.
Much of this adoption, IBM finds, is being driven by FOMO. Nobody wants to get left behind on the off chance this whole AI thing actually lives up to the hype. And this appears to have driven nearly two-thirds (64 percent) of respondents to adopt technology before they’ve figured whether it’ll actually benefit the organization.
Given the pace at which generative AI is evolving, IBM says half (50 percent) of respondents have found themselves juggling a growing number of disconnected and/or piecemeal technologies. The sheer cost of AI hardware, whether it’s in the cloud or on-prem, also remains a persistent challenge.
More than half of the participants in IBM’s survey admit their organization struggles to balance funding for existing operations and investment in innovation when unexpected change occurs. That tension may help explain why AI investments, for many, still haven’t delivered the hoped-for returns.
Also at Think IBM:
- Unveiled the Watsonx Orchestrate framework, which aims to simplify the process of building AI agents that integrate with popular SaaS platforms and software libraries
- Detailed its API Agent, which can identify existing APIs or, if necessary, generate custom ones for connecting existing data or services to AI agents.
- Announced agent AI integrations and partnerships with Salesforce, AWS, Oracle, Lumen
While the majority of those surveyed (72 percent) believe that harnessing their proprietary data sets will be key to unlocking the true “value of generative AI,” IBM suggests that many organizations are still struggling to do so. As we’ve previously discussed, data often needs to be normalized and filtered before it can be integrated into AI workflows like retrieval-augmented generation or used to fine tune models.
Staff may also need to be retrained or specialized positions created to build and manage these AI systems. More than half of respondents say they’re hiring for entirely new AI roles that didn’t exist a year ago, while the study suggests upwards of a third of the broader workforce may need retraining or reskilling within the next three years.
However long term, executives remain optimistic their GenAI investments will pay off eventually. Big Blue says 85 percent of those surveyed expected it’ll take at least two more years to achieve positive ROI. ®
