Session Results

Documentation Index

Fetch the complete documentation index at: https://help.harmonica.chat/llms.txt

Use this file to discover all available pages before exploring further.

​

Overview

After participants contribute to your session, Harmonica automatically synthesizes their conversations into actionable results. This guide covers how to navigate, manage, and export your session results.

​

Result tabs

Your session results page has several tabs:

​

Summary

The Summary tab shows an AI-generated synthesis of all participant conversations. It highlights key themes, common viewpoints, and notable insights across all responses.

• Summaries are generated in real time — you don’t need to wait for the session to end
• A color indicator shows the summary status:
  • Green — up to date
  • Yellow — currently updating
  • Gray — needs update (new conversations have come in since the last summary)

​

Responses

The Responses tab lists all participant submissions. For each participant, you can see:

• Their pre-session form answers
• Completion status
• Timestamps

​

Chat

The Chat tab shows the full conversation logs from all participants. You can read through individual conversations to see the detailed exchanges between participants and the AI facilitator.

​

Custom Insights

The Custom Insights tab lets you add your own observations and analysis on top of the AI-generated summary. Use this to annotate results with your own interpretation or highlight specific findings.

​

Files

If participants or you have uploaded documents (PDFs up to 10MB), they appear in the Files tab with metadata and download options.

​

Reviewing facilitator behavior

After participants have finished their conversations, a Review tab appears on your session results page. It gives you an AI-generated audit of how the facilitator handled the session.

​

What it shows

Click Run review to analyze the session. The review reads across all participant conversations — looking at how the facilitator asked questions, whether it stayed on topic, whether it pushed too hard or too gently, and similar facilitation patterns. If it finds anything worth flagging, it surfaces a list of findings. Each finding includes:

• A short title describing the issue
• A quoted excerpt from a real conversation as evidence
• A suggested prompt rule that would address the issue in future sessions

​

Apply a finding to your session

For any finding you agree with, click Apply. This opens your session’s Session Design panel with the suggested rule pre-staged in the Facilitation Approach. Review the proposed change, edit it if you want to adjust the wording, and save it. The rule becomes part of your facilitation prompt going forward. You can re-run the review at any time if you’ve made changes and want a fresh assessment.

​

Quotes as evidence

The session summary can display direct quotes from participant conversations as evidence for each finding. Instead of describing what participants said, the summary shows what they actually wrote.

​

Where quotes appear

When grounding is enabled, each section of the rendered summary expands to show a Supporting voices block — a set of short verbatim excerpts, labeled with anonymized participant identifiers (Participant 1, Participant 2, and so on). The quotes are tied to the specific summary claim they support, not shown as a general appendix.

​

Enable grounding on a session

A Ground with quotes button is available on the session results page for hosts on Pro. Clicking it runs the grounding pass, which reads each participant’s conversation, extracts the relevant verbatim excerpts, and maps them to the appropriate summary sections. The result is saved and displayed immediately. You can run grounding on any session that already has a summary — including older sessions created before this feature launched.

​

Re-running the summary

Grounding is re-run automatically whenever the summary regenerates. The quotes stay in sync with the current summary.

​

Following up with another session

After a session ends and the summary has been generated, a Suggested next session card appears below the results tabs. It offers an AI-generated proposal for a follow-up session based on what emerged from this one.

​

What the card shows

Click Suggest a follow-up to generate a proposal. The card displays:

• A suggested Topic for the next session
• A Critical field — the key question to focus on
• A Context field — background drawn from this session’s outcomes

Click Try another to generate a different proposal.

​

One-click follow-up

Click Create follow-up to open the session creation flow pre-filled with the suggested values. The current session is automatically attached as a context source — so the new facilitator will have access to this session’s summary as background material. If the current session belongs to a project, the follow-up session is created in the same project automatically. You can edit any field in the creation flow before launching. The suggestion is a starting point, not a commitment.

​

Managing participants in results

You can control which participants are included in the summary:

• Include/Exclude toggle — Click on a participant to exclude them from the summary. This is useful if a response is a test entry or off-topic.
• When you exclude or include a participant, the summary will regenerate to reflect the change.

​

Editing the summary prompt

You can customize how the summary is generated:

1. Click Edit summary prompt from the results page
2. Modify the instructions that guide summary generation
3. Save to regenerate the summary with your new instructions

This is useful when you want the summary to focus on specific aspects of the conversations or use a particular format.

​

Exporting data

You can export your session data for further analysis:

• Export responses — Download participant answers
• Export chat logs — Download full conversation transcripts
• Export summary — Download the synthesized summary

Exports are available in machine-readable formats for use in spreadsheets or other tools.

​

Summary best practices

​

Related guides