According to the vendor hype, business customers are all in when it comes to generative AI. However, like with any newer technology, huge corporations prefer to tread carefully. CIOs took notice this year as companies frantically revealed new generative AI-powered technologies.
Some businesses have been attempting to cut down on spending or at least keep even, rather than searching for new ways to spend money. The significant exception is when technology allows businesses to function more effectively and accomplish more with less.
Generative AI has the capacity to do so, but it comes with its own set of expenses, whether it’s a greater fee for these capabilities in a SaaS product or the cost of hitting a huge language model API if you’re creating your own program internally. 
In any case, it is critical for those adopting the technology to evaluate whether or not they are receiving a return on their investment. According to a July Morgan Stanley study of major firm CIOs, many were going cautiously, with 56% saying that generative AI was influencing their investment choices, but just 4% having begun substantial initiatives. In reality, the majority of them were still in the assessment or proof-of-concept stage. This is a fast-moving topic, but it corresponds to what we’re hearing from CIOs as well.
However, similar to the consumerization of IT a decade ago, CIOs are under pressure to provide the kinds of experiences that users are experiencing when they play with ChatGPT online, according to Jon Turow, a partner at Madrona Ventures.
“I believe it is undeniable that enterprise employees, who are the CIO or CTO’s internal customers, have all tried ChatGPT and know what amazing looks like. “They recognize where it’s early, where it’s inspirational, and where they perceive brilliance, for want of a better phrase. As a result, CIOs are under pressure to meet that standard,” Turow told Eltrys.
It has created a conflict between the urge to impress internal customers, particularly when some of that pressure may come from the CEO, and a CIO’s natural instinct to proceed slowly, even with something as potentially disruptive as generative AI. According to Jim Rowan, partner at Deloitte, who is working with clients on how to construct generative AI across enterprises in an orderly manner, this will need some structure and organization around how it gets applied over time.
“A lot of how we work with businesses is thinking about what infrastructure they need to be successful.” By infrastructure, Rowan does not necessarily mean technology, but rather “who are the people, what are the processes and governance, and giving them the capabilities to set that up?” A large part of it is discussing use cases and how to apply technology to a specific issue.
This is consistent with how CIOs we talked with approach implementation in their firms. Monica Caldas, CIO at Liberty Mutual, began with a few thousand-person proof of concept and is looking for methods to scale it up for her 45,000-person firm.
“We know generative AI will continue to play a critical role in virtually every part of our company, so we’re investing in many use cases to further develop and refine them in service of supporting our employees and giving them better internal capabilities,” she went on to say.
Mike Haney, CIO of Battelle, a science and technology consulting business, has also been investigating generative AI application cases this year. “So we’ve been doing this whole push for AI over the last maybe six or nine months, and we’re at the point right now where we’re building specific use cases for each different team and function within the firm.” He warns that it is still early and that they are constantly studying how it may assist, but so far the results have been positive in terms of providing more efficient ways to accomplish things.
Kathy Kay, executive VP and CIO of financial services firm Principal Financial Group, says her business began with a study group. “So we allowed any employees who had an interest or passion to join, so there are about 100 people.” “It’s a mix of engineers and business people, and we’re curating probably 25 use cases that they’ve gone through, three of which will go into production [soon],” she said.
Sharon Mandell, CIO at Juniper Networks, says her company is participating in an initial pilot with Microsoft around Copilot for Office 365, and she has heard anecdotal feedback ranging from people who love it to those who are less impressed, but she says measuring increased productivity remains difficult, even with Microsoft beginning to provide dashboards that at least show the level of adoption and usage.
“The difficult part is that you don’t have data on people’s productivity levels.” So, whatever, you’re utilizing anecdotal data until you become very skilled at understanding these dashboards from Microsoft that show you how people are using it,” she said.
When businesses learn about the potential power of generative AI, it’s natural for them to want to learn more about it and put it to use to help their organizations run more efficiently. However, executives are correct to be cautious, recognizing that this is still early days and that they must learn through experimentation if this is truly transformative technology.
