AI Insights

AI is no longer limited to code completion – it is steadily expanding into areas like logging, code review, and deployment pipelines. The next logical step is a context-aware browser. 

This is exactly what Comet from Perplexity AI offers. The assistant can see the content of your current tab without requiring any extra steps. There is no need to switch to a separate chat window, copy parts of documentation, or explain what you were looking at – Comet already understands the context.

The key difference compared to using Chrome with an external AI tool is not the model quality. It’s simpler than that: the “copy → paste → explain context” loop disappears entirely.

Let’s look at a few examples of how Comet can simplify a developer’s workflow.

Debugging in Azure Portal Without Context Switching

Working in the Azure Portal is rarely straightforward. Even seemingly simple tasks – like figuring out why a service has stopped receiving messages from Service Bus – can quickly become a series of navigation across multiple screens: topic metrics, subscription settings, dead-letter queues, and logs.

Each screen provides data, but the portal does not explain how these pieces connect. You see the numbers and configuration, and have to assemble the full picture yourself.

Comet changes this process. The assistant reads the current page, allowing you to ask questions directly within the context of what is displayed – without switching to another tool or re-explaining the situation from scratch.

Making Sense of Azure DevOps Pipeline Failures

A similar situation occurs in Azure DevOps, but here the challenge is often interpretation rather than configuration.

When a pipeline fails at a stage that previously worked, you are presented with a multi-level execution log containing nested jobs and non-obvious dependencies. Finding the single line that explains the failure among hundreds of log entries becomes a task in itself.

Comet allows you to ask questions like “Where is the actual cause of the error?” directly on the log page. Since the assistant reads the same output, it can help you identify the root cause much faster. This is especially useful for YAML pipelines, where parameter-related errors can surface in unexpected ways.

Limitations and Practical Constraints

Like any tool, Comet is only as useful as the problems it is designed to solve.

Its capabilities do not extend to the terminal, IDE, or local debugging, so it should not be expected to assist in those areas. The same applies to code generation. Perplexity is optimised for search and text understanding, not for writing code or generating Kubernetes manifests – in these cases, editor-integrated tools remain more effective.

There are also practical limitations. Linux is not supported, and in corporate environments using Azure AD, granting Comet broad access to a Google account may be problematic.

Where Comet Actually Adds Value

Comet does not change how you write code. However, it does change how you interact with information in your browser – and it does this well enough to be worth trying.

It is not a replacement for existing tools or a breakthrough in productivity. Instead, it is a focused improvement for specific contexts: Azure Portal, DevOps UI, technical documentation, and research workflows.

The more time you spend working in your browser, the more noticeable the benefits become.