Over the last few months, I’ve had countless conversations with my peers about one topic: the rise of AI coding assistants.
I know this isn’t exactly breaking news, and I’m sure you’ve had these conversations as well. But there’s a reason the common coffee chat today is 10 percent small talk and 90 percent about the AI-first world that we live in. Tools like GitHub Copilot, Cursor, and Devin are fundamentally changing how we write software. We’re shipping more code and we're doing it faster than ever before. But this new velocity has created a familiar problem in a new context. As our codebases expand at an exponential rate, our understanding of them is beginning to shrink.
The core mission of Cortex was born from my experiences with microservices at Uber, a challenge my co-founders Ganesh and Nikhil also experienced firsthand. We knew there had to be a better way to bring order to these software ecosystems. The question for leaders today is similar, just in a new context: how do we manage an ecosystem where humans and AI are co-creating at a scale we’ve never seen?
We believe the answer lies in evolving our thinking about the role of the Internal Developer Portal (IDP). For too long, the IDP has been viewed as just a system of record. To meet the demands of an AI-first world, it must become an active, intelligent platform that doesn’t just map your ecosystem, but helps you drive meaningful change within it.
A framework for the intelligent IDP
Our journey building Cortex has been defined by three key shifts in this thinking, which ultimately created a framework for evolving the IDP from a simple catalog into the copilot engineering leaders need.
Set the standard for excellence. The first step was realizing that a catalog alone isn’t enough. In 2021, we introduced the concept of Scorecards to the industry. The goal was to move beyond simply seeing what you have and start defining what “good” looks like. An IDP shouldn’t just tell you who owns a service; it should help you ensure that service is production-ready, secure, and well-maintained.
Measure the impact of change. Once you can drive standards, the next logical question is, “Is it working?” In 2023, we were the first IDP to integrate Engineering Intelligence to close this feedback loop. A leader might have a hypothesis that improving documentation will drive down MTTR. With an intelligent IDP, you can create a Scorecard to track runbook completion, launch an initiative to drive adoption, and then directly measure the impact on your core reliability metrics. This transforms the IDP from a system of enforcement to a platform for continuous, data-driven improvement.
Build intelligence into the platform. The scale and pace of AI-driven development demand a new level of intelligence within the IDP itself. This is the current frontier. Manually cataloging AI-generated services or predicting ownership based on static
CODEOWNERSfiles is no longer scalable. This is why we’ve invested in capabilities like AI-powered ownership prediction and a Model Context Protocol (MCP) that allows engineers to conversationally query their entire engineering ecosystem. The IDP must use AI to manage the complexity that AI creates.
This framework has led us to our most important philosophical conclusion yet: you shouldn’t pay for metrics, you should pay for change. Data without action is overhead. That’s why we believe so strongly in this vision of an active, intelligent IDP that we’ll be making our Engineering Intelligence capabilities free for all Cortex customers.
To truly succeed in this new era, we need a new mental model. We’re calling it the AI Excellence Framework, built on the pillars of AI Readiness, AI Impact, and AI Maturity. It’s a way to ensure you’re adopting AI on a stable foundation, proving its value with real metrics, and scaling its use with standardized best practices. An intelligent IDP is the platform that makes it possible.
The playbook for this new era of engineering is being written right now. In our keynote at IDPCON, Ganesh and I shared more of our perspective on this topic. If you're interested in learning more about how we're thinking about the future of IDPs, talk to our team about Engineering Excellence.
