The AI Knowledge Agent: Making Internal Developer Portals Smarter

Joining a new engineering team can be exciting, but it can also be overwhelming. You spend the first few days figuring out what each service does, where documentation lives, and who owns what. Even experienced engineers lose time switching between tools, looking for context, and waiting on answers.

Internal Developer Portals (IDPs) were designed to fix that problem. They centralize information about services, documentation, environments, and workflows so developers can work from one place.

The challenge is that most portals stop there. They display data but don’t help developers act on it. That’s where the Harness IDP Knowledge Agent comes in.

The Knowledge Agent lives inside the Harness IDP and serves as an AI-powered teammate. It helps developers find the information they need, understand how systems work, and take action faster. You can ask questions like, “Who owns the authentication service?” or “How can I improve my score?” and get relevant, trustworthy answers in seconds.

Why This Matters Now

The Harness & InfoQ webinar, Architecting the AI-Powered IDP, explored how developer time is really spent today. Research from Gartner shows that only about 30-40% of engineering time goes toward writing code. The rest, nearly 70%, is spent testing, securing, deploying, and maintaining systems.

That imbalance has real costs. Manual approvals, disconnected tools, and outdated scripts make software delivery slower and more expensive. Developers end up managing complexity instead of building products.

AI coding assistants promise efficiency, but they’re also widening the gap. According to the 2025 DORA State of AI-Assisted Software Development Report, AI improves individual productivity by 19%. However, organizational throughput improves only 3%, and delivery stability actually drops 9%.

Basically, this shows that teams are producing more code than ever, but delivery systems haven’t evolved to handle the extra load. The result is a growing delivery bottleneck, with more pipelines to maintain, more configurations to fix, and more friction across the software lifecycle.

Platform Quality is the Multiplier

The same DORA report points to a clear solution. Teams with strong, automated platforms gain both speed and stability from AI. Teams with scattered tools and manual workflows see the opposite, and AI magnifies inefficiency instead of solving it.

Harness research confirms this trend. Companies with mature continuous delivery practices are twice as likely to report faster release cadences when using AI. The takeaway is simple: platform quality determines whether AI helps or hurts.

A strong platform, built on reusable pipelines, consistent templates, and solid governance, is what allows AI to deliver real business value. That’s exactly what the Harness IDP Knowledge Agent helps teams achieve.

What the Knowledge Agent Does

The Knowledge Agent acts as the voice of your platform. It understands your services, pipelines, environments, and scorecards, and it uses that context to help you make faster, better decisions.

Ask about a service, and it tells you what it does, who owns it, when it was last deployed, and how it’s performing. Ask how to raise your service score, and it recommends specific steps based on your team’s best practices.

Instead of switching between dashboards, tools, and tickets, developers can ask questions right in the portal and get instant, actionable answers. It’s like having an expert teammate available whenever you need one.

How it Works

The Knowledge Agent is powered by Harness AI and built on a software delivery knowledge graph, which is a connected data layer that links information from across your engineering ecosystem.

The graph brings together build data, deployment events, test results, security scans, cloud cost reports, and monitoring insights. This context allows the AI to understand relationships between systems and services rather than pulling information in isolation.

Ask, “Where are the docs for this service?” and it searches internal wikis, TechDocs, README files, and Harness IDP documentation to find the most relevant information.

Ask, “Create a new Kubernetes environment in AWS,” and it identifies the right workflow, fills in parameters, and executes it automatically.

This combination of intelligence and automation transforms the Harness IDP from a static catalog into a dynamic platform that understands your organization and helps it operate more efficiently.

What Makes it Different

Many AI tools can summarize text or surface data, but the Harness Knowledge Agent is built directly into the Harness platform. That means it operates on live, trusted data from your services, pipelines, and environments.

It also respects your security and access controls. Developers see only what they’re authorized to view, and every answer includes transparent reasoning so teams understand how conclusions were reached.

This balance of accuracy, visibility, and explainability builds confidence. It’s not just an assistant, it’s a reliable partner that helps teams move faster with control and consistency.

From Dashboards to Intelligent Platforms

Developer portals typically evolve through three stages:

1. Static portals help teams find information.
2. Automated portals enable workflows and self-service actions.
3. Intelligent portals understand intent and complete tasks automatically.

The Knowledge Agent powers that final stage. It connects a knowledge graph, AI-driven workflows, and autonomous delivery into one intelligent interface.

With it, developers can ask for a new service with a database, and the portal handles everything, like provisioning infrastructure, configuring pipelines, applying governance, and verifying deployment health.

This evolution transforms a portal from a reference point into an active engine that drives real work.

A Practical Example

Imagine you’re new to a team and need to learn about the “authentication-service.” You ask the Knowledge Agent to explain it. It shows what the service does, who owns it, when it was last deployed, and its performance score.

Next, you ask how to improve reliability. The Agent reviews your scorecards and suggests adding monitoring, improving test coverage, and adjusting deployment approvals.

Finally, you ask it to start the workflow for monitoring. It executes the right pipeline, provisions the necessary resources, and reports success.

A process that might have taken hours now happens in minutes, without switching tools or asking for help.

Why it Matters

AI has changed how teams build software, but it hasn’t changed how they deliver it. Code creation is faster, yet delivery processes remain slow and fragmented.

Intelligent IDPs close that gap by connecting systems, surfacing insights, and turning context into action. They bridge development and delivery, reducing manual steps and helping teams ship faster and with confidence.

The Harness Knowledge Agent is the next step in that journey. It combines the power of automation with the intelligence of AI to make developer platforms more proactive, transparent, and efficient.

For developers, that means fewer roadblocks and faster onboarding. For platform teams, it means better governance and a scalable model for supporting the organization. For engineering leaders, it means seeing real ROI from AI investments.

Getting Started

The Harness Knowledge Agent is available now in Harness IDP for anyone who wants to try it.

If you already use Harness IDP, contact your account team to enable it. If not, you can request a demo to see how it works in action.

And to learn more about how platform engineering and AI come together, read Platform Engineering: Beyond the Trough of Disillusionment and Industry Reports Agree: DevOps is the Key to Unlocking AI’s Potential.

Harness is helping teams build the next generation of intelligent developer portals that are powered by AI, designed for scale, and ready for real-world engineering challenges. The Knowledge Agent is a first step in that evolution.