Skip to main content
A briefing is the stimulus material you give personas to react to. It represents whatever you’re testing — a product brief, a campaign concept, a creative execution, a packaging redesign rationale, or a new feature announcement. When you attach a briefing to a simulation, Boses automatically extracts the text and injects it into every persona’s prompt alongside your question. Personas read the material in character and respond based on their profile, cultural context, and the content you’ve provided.

​
Supported file types

FormatExtensionNotes
PDF.pdfText is extracted page by page.
Word document.docxFull document text is extracted.
Plain text.txtRead directly as UTF-8.
Image.jpg, .pngAccepted; visual OCR extraction is in development.
PDF and .docx files produce the best results today. If you’re uploading an image of a creative execution, add a plain-text brief alongside it to ensure the content is fully extracted.

​
Uploading a briefing

Briefings are uploaded as multipart/form-data. Each briefing belongs to a project and can be reused across multiple simulations. 
curl -X POST https://api.temujintechnologies.com/api/v1/projects/<PROJECT_ID>/briefings \
  -H "Authorization: Bearer <access_token>" \
  -F "file=@./product-brief.pdf"
The response returns the briefing object with its id: 
{
  "id": "brf_01hw5m3zxe4gqv8bsc9f",
  "filename": "product-brief.pdf",
  "file_type": "pdf",
  "extracted_text": "We are launching a new line of oat-based energy bars...",
  "project_id": "proj_01hw3k9zxe7fmv2bqc4d",
  "created_at": "2026-04-07T09:14:22Z"
}
Save the id as BRIEFING_ID. Pass it when creating a simulation: 
curl -X POST https://api.temujintechnologies.com/api/v1/projects/<PROJECT_ID>/simulations \
  -H "Authorization: Bearer <access_token>" \
  -H "Content-Type: application/json" \
  -d '{
    "persona_group_id": "<GROUP_ID>",
    "briefing_ids": ["<BRIEFING_ID>"],
    "simulation_type": "concept_test",
    "prompt_question": "Based on the brief, how would you react to this product?"
  }'

​
How text extraction works

Boses extracts readable text from your file automatically when you upload it. That extracted text is then injected directly into the simulation prompt that each persona receives — no manual copying required.
  • PDF files are processed page by page to extract all embedded text.
  • .docx files have their full document text extracted, including paragraphs and tables.
  • Plain text files are read as-is.
  • Images are accepted; full visual text extraction is in development — add a plain-text brief alongside an image for best results today.
If your PDF contains only scanned images (no embedded text layer), extraction will return minimal content. Export your brief as a text-layer PDF or use .docx instead.

​
What to include in a briefing

The richer the content of your briefing, the richer and more specific the persona responses will be. A one-paragraph summary will produce generic reactions. A detailed brief produces nuanced, insight-rich responses. Strong briefings include:
  • Product or campaign description — what it is, what it does, what problem it solves
  • Price point — specific pricing anchors reactions around affordability and value
  • Target positioning — how you intend to position it relative to the market
  • Competitive context — which brands or alternatives it sits alongside
  • Visuals description — if you’re testing packaging or creative, describe what it looks like in text
  • Key messages or claims — the specific things you want personas to react to

​
Briefings are optional

For concept_test simulations, a briefing is optional. If you don’t attach one, the simulation relies entirely on your prompt_question to frame what personas respond to. This works well for early-stage idea testing where you don’t yet have a formal brief. For simulation types that involve scripts or structured questions (idi_ai, survey, conjoint), the briefing provides background context that shapes how personas interpret the questions.
When running multiple simulations to compare concepts, upload each concept as a separate briefing and run parallel concept tests against the same persona group. Then use the convergence endpoint to compare how consistently personas respond across runs.

​
Managing briefings

List all briefings for a project: 
curl https://api.temujintechnologies.com/api/v1/projects/<PROJECT_ID>/briefings \
  -H "Authorization: Bearer <access_token>"
Delete a briefing (also removes the stored file): 
curl -X DELETE https://api.temujintechnologies.com/api/v1/projects/<PROJECT_ID>/briefings/<BRIEFING_ID> \
  -H "Authorization: Bearer <access_token>"