Browse docs
Getting started
Setup
Workspaces
Prompts
Brands
Models & coverage
Results
Metrics
Chats
Sources
What's new

Understanding chats

How BrandAxis logs every AI conversation, the math behind your monthly chat volume, and how to read a logged chat in the dashboard.

Updated

A chat is one unique conversation between BrandAxis and an AI model — produced by one prompt running against one model on one tracking cycle. We log every chat in full, so you can audit exactly what each model said about your brand and where its answer came from.

Chats are the raw data behind every score in the dashboard. Visibility, Sentiment, Rank, and Sources are all derived by analysing what was actually said in your chats.

How chat volume is calculated

chats per month = prompts × models tracked × runs per month

Each prompt produces one chat per tracked model per run. So if you’re on the Starter plan (50 prompts × 3 models × 30 daily runs), you produce 4,500 chats per month.

Chat volume by plan

PlanPromptsModelsChats / 30 days
Free51 (ChatGPT)150
Lite2031,800
Starter5034,500
Pro150313,500
Advanced350331,500
EnterpriseUnlimited8+ (all available)Custom

Reading a chat

Open any prompt’s detail view to browse its logged chats. Every chat row carries a Mentioned tag if your brand was named in that conversation — so you can scan the list and instantly see which prompts the AI answered with you in the response and which it answered without you.

For each chat you’ll see:

  • A Mentioned indicator on the row (or no marker if your brand wasn’t named)
  • The exact prompt text sent to the model
  • The model’s full response, unedited
  • Which of your tracked brands were mentioned and where in the answer (Rank)
  • The Sentiment label assigned to each mention
  • Every source URL or domain the model referenced — see Mentions and sources for the matching indicator in the Sources panel and how to use the two together

Chats are your audit trail. If a Visibility score looks off, open the chats behind it to read exactly what the model said. Unmarked chats are the highest-priority reads — they show you what the AI was saying instead of you.

What chats power

Every aggregated number in BrandAxis traces back to chats:

  • Visibility — counted from chats where your brand was mentioned
  • Sentiment — classified per mention inside each chat
  • Rank — position of your brand inside the chat response
  • Sources — extracted from the URLs each chat references

If you need to investigate a metric change, the chats behind it are the source of truth.