Run a transcript to see how a design review really worked. This page shows how to use the SIGNAL system on a real conversation. Take a design review call, run it through a set process, and see how decisions formed.
Instead of relying on memory or gut feel after the call, you can see:
-
Where the conversation was strong
-
Where it lost clarity
-
Where decisions started to form
-
What needs to improve next time
The goal is to make the conversation easier to learn from.
Why this matters
Most teams run design reviews, but few have a steady way to learn from them. After a call, people often leave with different versions of what happened:
-
What problem was actually discussed
-
What success was supposed to mean
-
What evidence shaped the direction
-
Why a decision felt weak
-
Who owned the next step
-
Whether anything was truly locked
Running a transcript makes that visible. It turns a conversation into something the team can review, score, and improve. You are looking at how the conversation worked, where it held together, and where it lost strength.
This helps improve the next review, not just grade the last one.
What you need before you start
Start with one design review transcript. It works best when it includes:
-
Speaker names
-
The full discussion
-
The design problem being reviewed
-
Any signals, data, or feedback mentioned
-
The closing conversation about ownership and next steps
You do not need a perfect transcript. You need enough of the conversation to see how the review moved from problem to action.
How the transcript is reviewed
The transcript is reviewed through the SIGNAL flow:
Surface → Identify → Ground → Navigate → Align → Lock
Each step looks for a different part of the conversation.
-
Surface: Was the real problem made clear?
-
Identify: Did the group define what success should look like?
-
Ground: Were signals used to shape the discussion?
-
Navigate: Did a clear direction start to form?
-
Align: Was ownership made clear?
-
Lock: Were next steps confirmed?
The review shows where the conversation created clarity and where it broke down.
3 ways to run a transcript
There are three ways to use the system, depending on how you want to work.
1. Use the skill
Best for fast, repeatable analysis in your AI tool of choice. The Design Review Skill runs the SIGNAL framework on a transcript with AI. It applies the rubric for you and gives structured feedback across all six steps.
Use this when you want:
-
Quick analysis
-
The same kind of output every time
-
Results you can compare across calls
-
A faster way to review design conversations
How to use it:
-
Copy your transcript.
-
Run the Design Review Skill.
-
Review the output from Surface to Lock.
-
Choose one step to improve in the next review.
👉 Get the skill: Design Review Skill
2. Use an AI prompt
Best for learning how the rubric works. A prompt lets you run the rubric yourself. You walk through each step and judge the conversation one step at a time.
Use this when you want:
-
To understand the SIGNAL system more deeply
-
To train your own judgment
-
To review an important call by hand
-
To see how the rubric works before using it at scale
How to use it:
-
Paste your transcript into your preferred AI tool.
-
Use one of the example prompts on the Design Review Skill page, or ask it to score the call with the SIGNAL Call Rubric and the Scoring Model.
-
Work through each SIGNAL step in order.
-
Review where the conversation gained or lost strength.
👉 Example prompts: Design Review Skill
3. Use Ray
Best for team workflows and ongoing analysis. Ray is the agent in the Glare app. You can upload transcripts, run the SIGNAL analysis, and track patterns across calls. Use Ray when you want to review more than one conversation and see where the team keeps getting stuck.
Use this when you want:
-
Team-level insight
-
Ongoing review of calls
-
The same evaluation across projects
-
Patterns across many design reviews
How to use it:
-
Upload your transcript to Ray in the Glare app.
-
Run the SIGNAL analysis.
-
Review patterns and results across calls.
-
Use the results to improve future reviews.
What a run gives you
After running a transcript, you should have a clearer view of how the design review went. The output should show:
-
How each SIGNAL step showed up
-
Where the conversation created clarity
-
Where the review lost strength
-
How decisions formed
-
Which weak point needs attention next
-
What to improve in the next review
Use the output to find where clarity first dropped. That is usually the best place to improve before the next call. The Scoring Model lists what to do when each step is weak, and the techniques give you the moves to use.
You do not need every call to be perfect. You need a clear read on where the conversation lost strength and what to improve next.
Running a transcript brings the Design Review system together.
-
The rubric shows what to look for.
-
The scoring model shows where the conversation gained or lost strength.
-
The techniques show what to improve next time.
Take one real conversation and run it through the system. Find where the flow broke down. Use that step to improve the next review. Over time, this creates clearer conversations, stronger decisions, and design reviews that keep moving forward.
