Infrastructure-as-code startup Pulumi Corp. is tapping into the power of generative artificial intelligence, launching a new feature called Pulumi Insights that uses OpenAI LP’s large language model GPT-4 to automate cloud infrastructure management.

The AI smarts are embedded in a new command-line tool known as PulumiGPT, which uses generative AI to enhance the productivity of cloud engineers.

Pulumi is an emerging player in the IaC market, selling tools that help companies to automate the management of their cloud computing environments. It provides a way for teams to use code to provision and manage cloud infrastructure, instead of adjusting hundreds of settings manually.

The launch of Pulumi Insights is meant to enhance the benefits of IaC. It takes advantage of Pulumi’s cloud supergraph, which is a connected graph of infrastructure that stretches across dozens of popular cloud platforms, to gain a deep, semantic understanding of the infrastructure customers are using.

As a result, Pulumi says, Insights delivers a more comprehensive search experience, where cloud engineers can ask questions of their infrastructure through structured search queries or natural language prompts. Teams can use it to locate a single resource among many hundreds of clouds and environments, or even perform more sophisticated queries, asking it to track down all untagged resources.

The real benefit of Pulumi Insights, however, is the advanced automation it brings, with its new embedded AI smarts helping teams to better understand cloud infrastructure usage patterns and deliver recommendations on how to optimize these resources. Teams can use PulumiGPT to write new infrastructure as code for any architecture on any cloud, in any programming language.

Alternatively, engineers can use PulumiGPT to automatically manage their cloud infrastructure in response to natural language prompts. In a nutshell, it uses AI to write the code that’s used to manage the cloud infrastructure.

Insights also enhances the analytics capabilities of Pulumi, with new dashboards, a REST application programming interface and data export features. With these features, teams programmatically create queries to add greater automation around search results, while better integrating with data analytics platforms such as Snowflake.

Holger Mueller of Constellation Research Inc. told SiliconANGLE that generative AI is rapidly disrupting many aspects of software, with DevOps and ITops both being key areas of progress. “Generative AI has the ability to create good quality code, and it can change the dynamics of IaC too,” Mueller said. “Pulumi is making it possible for platform engineers to use generated code for script delivery, and it will enable much greater velocity for software teams.”

“This is yet another step-function improvement to engineering productivity, and now it’s easy to leverage the supergraph and unique AI insights to have a big business impact,” founder and Chief Executive Joe Duffy said. “Pulumi Insights is a huge one-way door for infrastructure as code, and we’re only just getting started.”

Pulumi Insights and PulumiGPT are available now for customers.

Image: Pulumi