Anecdotally, it’s being said, at least anecdotally, that the AI hype has been cooling as of late. Gartner, the consultancy that devised its oft-quoted hype cycle, agrees.
“GenAI has passed the Peak of Inflated Expectations,” wrote Afraz Jaffri, analyst with Gartner, in a bulletin published in November. “By the end of 2024, value will be largely derived from projects based on familiar AI techniques, either stand-alone or in combination with GenAI, that have standardized processes to aid implementation.” In other words, time to get past the shiny-new-thing stage to roll up one’s sleeves and make it all work.
Industry leaders agree that it’s time for the hard and purposeful work to begin. “While there is a lot of hype around AI, we have seen impressive outcomes in such a short span of time,” said Sri Elaporu, global head of the AWS Generative AI Innovation Center. “Companies are already seeing real benefits of generative AI.”
“There is certainly a lot of hype and potentially unrealistic expectations around AI, especially generative AI, in the short run,” said Steve Chase, vice chair of AI and digital innovation at KPMG. “This can put unrealistic pressure on teams to deliver immediate results and cause doubts about continuing investments when those lofty expectations are not met right away. However, many companies are already seeing tangible benefits from AI. The mid to long-term potential of AI is likely understated.”
Elaporu pointed to some large customers his organization is working with, and how they are gaining ground with targeted genAI projects. “For instance, Rocket Mortgage – America’s largest retail mortgage lender – leveraged genAI built using Amazon Bedrock to improve the path to homeownership and clients have seen a 10% increase in resolutions during their first call to Rocket,” he said. Another client. Bayer, is also using genAI “to accelerate and re-envision the development of food across Bayer’s 120 million acres of growing fields.”
KPMG is working with its clients “to implement AI solutions across various functions, such as marketing, sales, customer service, field service, procurement, and software development,” Chase said. “In highly regulated industries we are helping clients use AI to process large quantities of written content to produce structured and repeatable summarized documents — think rate cases for utilities or required product labeling in life sciences.”
At its core, both operational and generative AI require human input. “At KPMG, we’ve observed that the primary challenges companies face in genAI adoption are engaging employees, helping them understand the benefits of the technology, and keeping pace with the rapid advancements in the field,” Chase said.
“There’s often a gap between leaders’ enthusiasm for the disruption that genAI brings and the concerns of employees who are nervous about what the technology might mean for them, especially as it evolves quickly,” he added.
Enterprises express concern about AI “around data privacy, integration with existing systems, model selection, and the skill gaps,” Elaporu observed. “However, the first step—knowing how and where to start—is often the most significant barrier to entry.” He stated that moving from proof of concept to a productized service is “the most difficult step to generative AI adoption.”
Chase advocates a human-centric approach to AI to deliver working value. “This involves focusing on robust change management, training, and communication efforts,” he says. “Engaging with employees requires connecting with them on an emotional level, using storytelling techniques to share examples of how the technology can be used and celebrating successes.”
Reskilling is also critical, “as many companies lack the necessary skills to fully leverage AI technology,” Chase continued. “We have also found that adoption occurs faster when AI capabilities are embedded into the tools or workflows that employees are already using, rather than requiring them to navigate to another system.”
