1. News
  2. Cloud & infrastructure
  3. Google Cloud · Survey
Google Cloud · Survey2 min read

Old software is slowing down company AI plans, survey finds

A Google Cloud survey of 1,250 US decision-makers says older, locally installed software is delaying or killing AI projects at nearly half of companies.

By the Dashline deskSource: Google Cloud blog
Old software is slowing down company AI plans, survey finds
Photo: Unnerving duck, CC BY-SA 4.0

Nearly half of the US business and IT leaders in a new survey say older software is getting in the way of their AI plans. Google Cloud and Hypothesis Group surveyed 1,250 decision-makers. Of them, 49% said relying on software installed on their own machines has delayed, shrunk, paused or canceled AI or other technology projects.

What the survey found

The study calls software "legacy" if it runs locally on computers and servers instead of in the cloud, meaning on remote servers run by a provider. Respondents named routine patching and maintenance as a main bottleneck. Patches are updates that fix security flaws. The report says AI is finding flaws in old code faster than security teams can fix them. That leaves less time for new AI work.

The survey also asked about choice. 80% agree that a provider's existing customer base should not decide which AI a company uses. And 78% believe the companies that led the last era of business technology may not be best placed to lead the AI era.

What to keep in mind

The findings come from Google Cloud, which sells cloud and AI services. The post was written by one of its executives, Jeanette Manfra. It does not say how respondents were chosen or how results differed by industry. The survey covers larger enterprises, so the numbers may not match a small business. Still, the question it raises applies widely: how much time goes to keeping old software running?

The takeaway

If your business relies on locally installed software, track how much time and money goes to upkeep, since that is what the survey says holds back newer tools.

More articles

You should really read these.

Next step

Let’s talk