See Trackingplan's AI Debugger in action, the AI agent that finds tracking warnings, explains errors, and improves data quality and tracking accuracy.
We’re super excited to show you Trackingplan’s new AI Debugger – the AI-powered tool that transforms raw debug warnings into clear, human-readable explanations.
No more spending hours digging through warnings and their possible root causes – now our AI agent helps you understand them quickly, see the cause, and take action fast.
Think of it as an AI-assisted debugging companion that tells you exactly what to fix, guiding you through the resolution process with clear, step-by-step instructions.Happy tracking!
Transcript
We're super excited to show you Trackingplan's new AI debugger, the AI powered tool that transforms raw debug warnings into clear, human readable explanations. No more spending hours digging through warnings and its possible root causes. Now you can understand them quickly, see the cause, and take action fast. To access Tracking Plans AI debugger, let's first open a warning from our warning list. here inside your warning management view. Now, you'll find a brand new tab called warning analysis. As you see, this functionality is only enabled if you or your clients explicitly choose to opt in. So, let's first activate it by going to our settings. In here, you can also configure instructions for the AI agent to adjust how AI generated explanations are written to better fit your team's context and way of working. We could add specific context about our plan, like for example that the data layer is not relevant for us. We could also define the language of the responses in case English isn't really your thing. Or we could also specify a particular tone or level of detail by telling our AI agent to be less technical and keep explanations simple and easy to follow. These instructions will be automatically included in the AI warning analysis prompts, allowing the AI agent to adapt its explanations to your needs. So, now that we've activated it and told our agent the way we want our responses, let's see it in action. Let's open the warning we want to debug again. And in just a few seconds, our AI agent will generate a human readable explanation of it based entirely on your debug warning data, but presented in a way that's easier to consume. Perfect for those who are short on time and need to act fast. By diving into the details, we see how the AI agent has given us a clear high-level overview of the issue, combining all the relevant debug data into a single easy to understand paragraph. It has also provided us with a clear interpretation of the most likely root cause written in plain language to make it easy to understand. It has estimated the potential business impact of the warning showing what this issue could mean for performance like its impact on conversions. campaign attribution or overall data reliability. This makes it much easier to decide what to fix first based on the real impact the issue could have if it's not resolved. And it's especially useful when you need to explain your work to others. For example, if you work in an agency or with stakeholders, this helps you clearly show what you've proactively fixed. And here goes the important part, the recommended actions. This section includes actionable suggestions on what to review or fix based on the detected warning patterns. Think of it as an AI assisted debugging companion that tells you exactly what to fix, guiding you through the resolution process with clear step-by-step instructions. Finally, it has also provided us with a concise summary of when and under which conditions the issue occurs, derived directly from tracking plans warning debug. With all of this, you already have a clear starting point that you can simply copy and paste into a readytouse ticket and share directly with the person responsible for fixing the issue without having to write anything yourself or translate technical details. But if you want to go one step further and validate these hypotheses using real user session tracks and correlation analysis, you can do that, too. But we'll leave that for another video. We'd love to hear what you think about this new feature. And if you have any questions, don't hesitate to reach out to our support team. Thanks for watching and happy tracking.


