Postman added an artificial intelligence (AI) agent to its portfolio of tools and platforms for building and governing application programming interfaces (APIs) that can autonomously perform tasks ranging from development and documentation to exploration and setting up integrations with continuous integration/continuous deployment (CI/CD) environments.
Company CEO Abhinav Asthana said the Autonomous API Engineer significantly reduces the total cost of building and maintaining APIs by automating time-consuming tasks that have historically created bottlenecks in software engineering workflows.
In fact, the AI agent developed by Postman will make it significantly simpler to integrate API development and testing within those workflows, said Asthana.
Designed to be triggered from a pull request, Slack, Postman command line interface (CLI) or the Postman app, the Autonomous API Engineer spins up a secure, sandboxed environment. It then executes tasks and returns verified artifacts, including collections, test results, specs, run logs, pull requests, and a temporary cloud workspace—that teams can review and ship.
For example, the Autonomous API Engineer can run API and quality assurance (QA) tests on every pull request and then post results back into existing developer workflows. It can also investigate API issues and accelerate root cause analysis, tracing dependencies across services, actively testing APIs, and returning actionable hypotheses with reproduction steps.
The Autonomous API Engineer leverages an existing Postman Context Graph database that captures how an API was built, changed, and governed over time. The Autonomous API Engineer then uses that context to generate more reliable output than a general-purpose AI agent, said Asthana.
It’s not clear to what degree the building and management of APIs has been integrated into DevOps workflows, but in the age of AI it will be simpler to achieve that goal as the boundaries that exist between various tools and platforms become easier to bridge. In fact, in many cases the AI agent created by Postman will be communicating directly with AI agents created by providers of other DevOps platforms to complete a task, noted Asthana.
Each DevOps team will, of course, need to determine what types of tasks to assign to one AI agent over another, but it’s clear many tasks once performed manually by DevOps engineers will be performed by a small army of AI agents. The issue then becomes finding a way to put the right orchestration framework in place to manage all those AI agents.
In the meantime, DevOps teams should pay careful attention to how any given AI agent performs those tasks, said Asthana. In the absence of access to some type of method that provides context, an AI agent is likely to burn tokens needlessly investigating how to perform a task when that information is readily available in a graph, he added. At a time when organizations are becoming more concerned about the cost of AI, a graph is going to prove to be an essential element of an effort to rein in those costs, said Asthana.
Regardless of approach, the issue now isn’t so much whether to incorporate AI agents into DevOps workflows but rather to what degree to rely on them based on how they were actually constructed.