A new survey I worked on, sponsored by AWS and the MIT Chief Data Officer Symposium, has been released, and it’s one of the first I have seen that addresses how organizations are dealing with generative AI. There are some interesting findings, the most important of which is that while those who are leading data and AI efforts at their companies think that generative AI is likely to transform their organizations, they haven’t done much yet to prepare for the change.
Let’s get into the details. First, there is some diffusion of responsibility relative to generative AI. The survey was of 334 global chief data officers (CDOs), who usually lead the development of analytics and AI as well. But in this survey, only 42% of CDOs said they have responsibility for strategy and execution with respect to generative AI. This is despite 49% saying that they are responsible for building AI capabilities, and 61% for advanced analytics capabilities. It’s not clear who else has the oversight for generative AI, but AI is highly related to analytics and highly dependent on data, so I hope it moves into the CDO bailiwick eventually.
Overall, the surveyed CDOs are excited about the potential of generative AI. 80% believe it is likely to transform their organizations. 45% agreed that they foresee or are already seeing wide adoption of it within their organizations. That should motivate a lot of activity to get ready for it.
Experimentation, Not Production
Despite this enthusiasm, it is clearly early days for AI in most organizations. 16% of CDOs said their organizations ban it altogether, which to my mind is a bad idea—it’s easy enough to protect intellectual property if that’s your concern. 47% said their organizations were allowing experimentation at the individual employee level, either with or without clear usage guidelines. Only 19% said they had department or business unit-level experimentation with gen AI use cases, and 11% had organization-wide experiments. A mere 6% said they had one or more gen AI use cases in production deployment. We clearly need more of those.
In terms of what organizations are experimenting with, 40% said their focus was on overall personal productivity. That’s fine, but the technology clearly offers more. One potential benefit it offers is customer operations benefits, such as with gen AI chatbots. 44% are prioritizing this type of use case—the highest percentage. 37% are prioritizing use cases for software engineering, such as code generation. 32% are pursuing marketing and sales use cases, and 14% are prioritizing R&D applications.
Data Is Key, But It’s Not Ready
CDOs believe data is key to preparing for generative AI, but they haven’t done much yet with it. 93% agree that “data strategy is crucial to getting business value from generative AI.” However, 57% said that they had made no changes to data yet to prepare for generative AI. Only 38% agreed that “My team and I have the right data foundation to pivot to generative AI,” and only 11% agreed strongly with the statement. 71% agreed that “generative AI is interesting, but we are more focused on other data initiatives to achieve more tangible value.” Tangible value is great, but perhaps this low priority is why many CDOs haven’t been given responsibility for generative AI. At least they are planning to spend more on the technology: 62% said that their teams are planning on investing more in generative AI.
Those who are preparing seem to be taking relatively small steps. 25% are undertaking “data integration or cleaning of datasets for generative AI use cases;” 18% are surveying data that might support gen AI use cases, and 17% are curating proprietary documents or textual data to train domain-specific generative AI.
Upping Their Game
I believe that CDOs need to increase their focus on generative AI. It is the technology that has most excited both executives and employees in organizations around the world. CDOs can use this to build their own reputations; one that we interviewed for the survey report, for example, said that she feels like the “belle of the ball” since generative AI came out.
If organizations are to succeed with generative AI, they need to increase the focus on data preparation for it, which is a primary prerequisite for success. Leaders in the adoption of generative AI, such as Morgan Stanley, have been working aggressively on these types of initiatives for a couple of years.
They also need to become experts on the technology and how to take advantage of it. Among other things, this means using gen AI in their own work. In our survey, only 53% said that they were currently using generative AI themselves. Admittedly, it’s almost a full-time job to keep track of all the changes in this technology, but you can’t become an expert just by reading about it.
Generative AI is the most dramatic advancement of our age. It’s time that companies were more active in their preparation for and adoption of it. Chief data officers can lead the way, but it’s not clear that they are doing so yet. Granted that these busy executives already have a lot on their plates (see survey report for details), but this is already or soon to be the main course.
Read the full article here
