Blog
Using Storytell to Build Storytell
blog category

How the Storytell Crew Uses Storytell to Enrich User Profiles

By
Louise Maniego
June 27, 2025
How the Storytell Crew Uses Storytell to Enrich User Profiles

As part of our How We Use Storytell to Build Storytell series, we're sharing how our team enriches user profiles by analyzing what people actually ask Stoytell and how you can apply the same approach to understand your own users better.

Understanding users isn't just about job titles or demographics. The real signal comes from what people do with your product: what they ask, how they interact, and what problems they’re trying to solve. At Storytell, we use our own platform to turn user prompts into structured, actionable profiles that help us prioritize features, shape onboarding, and improve outreach.

Try it yourself

  1. Export user prompts
  2. Create a classification rubric
  3. Upload both files to Storytell and run a structured analysis prompt
  4. Apply the output in your internal tools
  5. Repeat regularly

Why we enrich user profiles from prompts

The prompts users send give us direct insight into what they’re trying to do. Instead of relying on static labels, we classify users based on how they actually engage with Storytell. This helps us make more informed decisions about what to build, who to build it for, and how to communicate along the way.

Step 1: Export user prompts and organize data

We start by exporting a CSV from PostHog. It includes:

  • Prompt text from users
  • Timestamps
  • Identifiers like email or user ID

Step 2: Create a rubric CSV for classification

Next, we prepare a single rubric file that combines persona and seniority classifications. In this PersonaID and SeniorityID file we created for our process, the structure is hierarchical:

  • W =  any work-related category
  • W-1.0 = Sales
  • W-1.1 = Account Executive
  • W-1.1.1 = Enterprise Account Executive
  • W-2.0 = Marketing
  • and so on

Step 3: Upload both files and run the analysis prompt

In a new Storytell Collection, we upload:

  • The user data CSV
  • The rubric CSV

These files are then referenced together in a single structured prompt.

With the files uploaded, we write a prompt like this:

Prompt
Analyze all prompts provided by the user in @Collection and categorize this user according to the rubric in @PersonaIDs + SeniorityIDs - PersonaID.csv. 

For each prompt, identify relevant codes and their corresponding categories and subcategories from the rubric. 

Then determine the user's primary and secondary category based on the frequency and thematic relevance of prompt content. 

Present the output first in bullet point form, and then below that as a table with these columns: 

(1) Primary Category - include the code and full hierarchical category/subcategory path (e.g., "W-2.1.3: Marketing: Brand Marketing: Brand Development"), 

(2) Secondary Category - similarly detailed, 

(3) Category Reasoning - explain why these categories best represent the user, citing the nature and content of the prompts, 

(4) Implied Seniority - assign the most fitting seniority band (e.g., "M3: Manager") based on prompt complexity and focus, and 

(5) Seniority Reasoning - justify the seniority assignment with reference to prompt types, responsibilities implied, and task complexity.
  

Storytell runs the analysis row-by-row, referencing the rubric and returning structured outputs we can use immediately.

Step 4: Enrich user profiles across systems

We feed these outputs into our internal tools like our CRM, user tracking docs, or GTM workflows. This helps us:

  • Segment users based on behavior
  • Prioritize feature requests from specific groups
  • Tailor messaging based on what different personas care about

For example, when we tag someone as W-8.1.1 – Product Manager, we know they’re likely focused on user research, cross-functional alignment, and roadmap clarity so we surface features and language that match.

Step 5: Track evolution and validate insights

We re-run this process over time to see how user behavior changes. This helps us:

  • Validate that our classifications still match usage
  • Understand how personas evolve as the product grows
  • Adjust onboarding, outreach, or product direction accordingly

What this enables

This workflow gives us a consistent, low-effort way to turn raw user input into structured insights. It makes our product, outreach, and planning better aligned with how people actually use Storytell—while using Storytell to do it.

Gallery

Changelogs

Here's what we rolled out this week
No items found.