A new chapter in software dependability is beginning to unfold, as artificial intelligence, and the more autonomous agents built upon it, push development teams past the familiar cycle of noticing mistakes only after they occur. More and more, programmers are turning to tools that do not simply reveal faults, but think through the pathways by which those faults might come into being.
This development is quietly altering the standards by which code is judged, changing monitoring from a labour of reaction into something closer to an automatic shield. Such is the view of Milin Desai, chief executive of Sentry, the company formally known as Functional Software Inc.
Speaking with Christophe Bertrand during theCUBE’s coverage of AWS re:Invent, in an interview streamed by SiliconANGLE Media, Desai reflected on the ways agentic AI is recasting the developer’s routine. He described how Sentry aims to move beyond mere observability, towards a practice in which the system reasons about potential failures and prevents them before the developer is even aware they exist.
Sentry’s Seer Unlocks AI-Driven Root-Cause Reasoning for Dev Teams
Seer, the new reasoning layer in Sentry’s arsenal, does more than sift through an enterprise’s production errors and traces to uncover their origins. It draws upon the wider environment in which that software lives, Desai explained to theCUBE. Once the underlying cause is identified, the system can call upon an internal coding agent to produce a remedy and, with growing confidence, to halt flawed code long before it has the chance to reach the outside world.
From Error Tracking Tool to AI-First Stability Platform
Sentry began its life as a tool for tracking errors, yet it has grown into a broad monitoring platform that absorbs everything from performance signals to logs and session recordings across web, mobile and backend systems. This wealth of production detail is what Seer uses to offer practical fixes and to warn developers of dangerous alterations before they escape review, an instance of agentic AI embedded directly into the ordinary course of development.
Final Words
The vision of software that writes itself before breakfast is no longer the fever dream of overworked programmers. The industry is also moving toward the time when artificial intelligence will not merely list the disasters, but will actively collude to avert them, with Sentry and its Seer and other reasoning engines. It is yet to be determined whether this evolution will allow the developers to avoid their 3 a.m. panic attacks or just give them new and exciting means of panicking.
What is definite is that the old beat of ship it, break it, fix it is being replaced by a more careful and much less dramatic one. It is no longer whether AI agents will be able to detect bugs before they fly into the wild, but whether the developers will miss the adrenaline rush of firefighting production catastrophes.
