You can already ask Amazon Echo with digital assistant Alexa for a weather report or to order a pizza delivery. But in the not-so-distant future, an IT professional will also be able to use Alexa (or Google Assistant, for that matter) to provision servers or storage in the cloud.
There are already test cases online. One video from a local Amazon Web Services (AMZN) meet-up in Michigan last March demonstrates how IT professionals could use Alexa to control AWS deployments, re-deploying a given workload from one AWS region to another, or requesting a company’s current AWS bill.
More recently, storage player Tintri posted a video showing how an IT pro can use Alexa to provision storage—as in telling the storage system how much disk space to allocate to a specific job.
Get Data Sheet, Fortune’s technology newsletter.
Yet another video from IT consultant Janakiram MSV shows him issuing voice commands to Google Assistant to set up a Kubernetes cluster. Kubernetes is a method backed by Google (GOOG) and its allies as their choice method to orchestrate and manage clusters of computing power in a cloud or data center.
“There is an emerging concept called VoiceOps—DevOps handled through voice,” Janakiram said via email. The idea is to enable operations professionals and administrators to use voice commands to perform common data center operations.
DevOps is an industry term combining development and operations functions—a more modern way to ensure software developers and operations counterparts work together early in the process so that software can be easily deployed. This was not the case in the past, when the teams were separate and developers weren’t concerned if features and functions built in put undue stress on the underlying infrastructure—a responsibility previously reserved for the operations department.
But as companies saw the need to update their software faster and more efficiently, they started to mandate that development and operations teams collaborate.
For more on Alexa, watch:
VoiceOps, sometimes also called “ChatOps,” may not be commercially available yet. But these demonstrations illustrate what is possible.