Skip to main content

PulseAI

PulseAI is a conversational analytics engine built into the Pulse dashboard. Instead of manually slicing data across dimensions and time ranges, you ask questions in natural language and PulseAI returns an actionable answer with supporting evidence.

How It Works​

PulseAI runs automated root-cause analysis across your sessions, cohorts, and experience metrics. When you ask a question, it:

  1. Identifies the relevant interactions, time ranges, and dimensions
  2. Analyzes the data for patterns, regressions, and correlations
  3. Returns a structured answer with evidence: charts, affected segments, and links to session replays

Example Queries​

QuestionWhat PulseAI Does
"Why did checkout conversion drop this week?"Identifies the interaction, finds the degraded metric, isolates the affected segment (e.g., Android 14 + specific carrier), and traces it to the likely root cause (e.g., a latency spike after v4.2.1 release)
"Which devices have the worst add-to-cart experience?"Ranks devices by Apdex score for the Add to Cart interaction, highlighting the worst performers
"What changed after the last release?"Compares interaction metrics before and after the release, surfacing any regressions
"How many users are affected by the payment flow error?"Counts distinct users who encountered errors in the payment interaction, segmented by device and region
"Show me sessions where login took more than 5 seconds"Filters sessions by the Login interaction with P99 > 5s and provides links to replays

Using PulseAI​

  1. Click the PulseAI icon in the left sidebar or the chat bubble in the bottom right
  2. Type your question in natural language
  3. PulseAI returns an answer with:
    • A summary of findings
    • Supporting data (charts, tables, metrics)
    • Affected segment breakdown
    • Links to relevant session replays
    • Suggested next steps

When to Use PulseAI​

  • Incident investigation — A metric dropped and you need to find the cause fast
  • Release validation — Check whether a new release introduced regressions
  • Executive reporting — Get a quick summary of experience health for a specific period
  • Exploratory analysis — Discover patterns you didn't know to look for

Tips​

  • Be specific about the interaction — "Why did checkout conversion drop?" is better than "Why did conversion drop?"
  • Include time context — "...this week" or "...after the v4.2.1 release" helps PulseAI narrow the analysis
  • Ask follow-up questions — PulseAI maintains context within a conversation, so you can drill deeper

Next Steps​