Multicloud networking and security startup Prosimo today added artificial intelligence operations (AIOps)-enhanced networking and zero-trust security functionality to its multicloud platform.

Founded in 2019, the company emerged from stealth mode this spring with a multicloud platform designed to address the application performance and security challenges faced by enterprises as they moved to adopt multicloud architectures. The company’s Application eXperience Infrastructure (AXI) is a cloud-native platform that runs in all the major public clouds, including Amazon Web Services (AWS), Google Cloud Platform (GCP), and Microsoft Azure. Once connected to customers’ workloads, AXI automatically routes application and user traffic across the public cloud provider’s private backbones.

Today’s update also introduces several new features to the AXI App Transit service, which now tap machine learning to more intelligently optimize traffic routing and optimization.

According to Mani Ganesan, head of product at Prosimo, these capabilities were added after it became clear customers were struggling with multicloud network configuration and performance. He described an early customer with more than 400,000 employees around the globe, which struggled with application performance, especially when users in the Asia-Pacific region attempted to access workloads running in the U.S.

“The performance was a key factor,” he said. “In fact, that was what was slowing down their cloud migration.”

To address this, Prosimo turned to machine learning to forecast optimal routes on a per-application basis.

On top of the IP traffic layer, “we have the ML layer, which is kind of crunching all the data in terms of going and discovering what the customers are doing,” Ganesan said.

This model allows Prosimo to automatically identify optimal traffic paths for each application, what kind of optimizations — for example, image caching — are applicable for it, and what kind of segmentation policy needs to be enforced to maintain the enterprise’s security posture.

These capabilities are now available as Fastlane and Autonomous Transit in AXI.

Fastlane uses machine learning to dynamically optimize the performance of latency sensitive applications by spinning up additional Prosimo points of presence (PoPs) at the edge of the network. Meanwhile, Autonomous Transit tackles app-to-app communications and provides recommendations for improving performance or reducing latency.

The company is also using machine learning for what it calls its Dynamic Compliance service. Because Prosimo sits in the middle of the traffic flow, the service can apply zero-trust network access security policy to lock down access to applications or workloads based on factors including identity and location, Ganesan explained.

For example, Dynamic Compliance can identify if a user accessed an application from New York, but then minutes later attempted to log in from Hong Kong. If anomalous activity is detected, the user’s reputation score is adjusted accordingly, and if the score reaches a threshold defined by the customer, access to certain workloads can be restricted or cut off entirely, Ganesan explained.

Prosimo’s ambitions for automating multicloud networking and security don’t stop there.

The company also wants to close the loop on AIOps capabilities by taking the user entirely out of the equation and enabling the system to automatically optimize and remediate traffic paths and mitigate security threats. The platform for now is limited to notifying administrators of optimization opportunities or applying predefined security policy, and notifying the security operations center for follow up.