The conversation around AI today reminds me of the dot.com conversations we were having in the late ’90s, or the direct-to-consumer conversations we were having in the 2010s. And we’re seeing similar results, too: a proliferation of projects that nobody in the organization truly has their arms around.
In some ways, that’s great—after all, innovation is a numbers game. Statistically, the more you try the more likely you are to find something that will be valuable. But containing costs can be a huge challenge.
According to Gartner, by next year 35% of companies will have appointed a chief AI officer, a big indication that serious resources are going into AI. Another study, this one from Boston Consulting Group, projects that spending on GenAI will grow 60% over the next three years, accounting for more than 7.5% of the average corporate IT budget.
Spending on AI systems looks a lot like spending on earlier-generation Software as a Service (SaaS) products. One recent report concluded that the average organization has 269 SaaS applications. For large organizations, there can be two or three times this many applications in place. Of these, the same report finds, only 17% are managed by a central operation, such as the IT department. Further, a typical company adds six new SaaS applications every 30 days (and we’d be willing to bet they are not retiring an equal number). For a company with 10,000-plus employees, the average IT spend on SaaS products is $264 million!
Shadow IT and app proliferation
Often called “shadow IT,” app proliferation creates noticeable operational problems. The budget for these things is not transparent; typically they are lumped together in an overall spending line and nobody knows whether the software is working, whether anybody is using it, whether it’s a duplicate (with a duplicate payment), or whether it’s producing desired benefits.
The spending is bad enough, but the impact on the employee experience is even worse—a point made by Tiffani Bova. A 2021 study (conducted even before the current AI buzz) found the following:
- 69% of respondents reported that finding the information they need to do their jobs is difficult.
- 54% said that the applications used to access information made the work harder, not easier.
- 49% were worried that information would get lost.
- 43% reported spending too much time switching between different online tools.
Strategic use of AI
Doing a lot of experimentation with AI is great, in my view. But experiments and tinkering, while often useful, aren’t going to yield a corporate-wide competitive advantage unless their results are discovered and scaled. So this is probably a good time to consider whether your AI efforts are set up for strategic success. Consider the questions below (devised in collaboration with my colleague Kes Sampanthar):
