1 · Define objectives in the builder
Set survey-flow goals, tone, boundaries, and optional blueprint questions. Use the AI co-pilot to refine objectives, generate a draft blueprint, and attach context links or images the agent may show during chat.
Insight agent
One AI for interviews, fields, and notes
Replace separate “chat survey” and “feedback bot” stacks. The Insight agent conducts natural dialogue, fills blueprint slots for analytics, and writes structured notes when participants share real signal.
Build with AI —
Configure survey-flow objectives and optional feedback-session objectives in one builder. AI helps you tighten goals, tone, boundaries, and anchors before you publish.
AI at runtime —
At runtime the same model follows your guardrails, streams a modern chat UI, persists field captures, and updates analyst-facing notes — with embed-friendly controls when you need them.
- Adaptive follow-ups with blueprint capture for dashboards
- Structured notes for themes, sentiment, and follow-up actions
- Embeddable experiences with origin-aware policies when you deploy on the web
How it works
Configure, publish, then analyze
Insight agent studies move from brief to live participant chat to analysis-ready data without switching tools.
2 · Brand & preview the participant chat
Customize welcome title, subtitle, logo, themes, photo backgrounds, overlay, and text contrast. Run draft test preview from the builder before you publish a public link or website embed.
3 · Collect signal and review analysis
Participants chat naturally while the agent captures blueprint fields and writes structured notes. Open per-survey analysis for sentiment, themes, session transcripts, comparisons, and canonical topic clusters.
Dynamic builder
Everything you need before go-live
The Insight workspace is split into focused tabs — objectives, context, appearance, and blueprint — with AI assistance throughout.
Objectives & AI chat
Describe what you need to learn; refine primary goals, guardrails, and anchors with an in-builder AI assistant before participants see anything.
Blueprint & question depth
Approve a question blueprint for analytics slots, tune how deep the agent should probe, and version changes between draft and publish.
Context assets
Add links and images with titles and AI guidance — the agent can surface them mid-conversation when relevant (demo pages, screenshots, help articles).
Participant appearance
Control public title, subtitle, start label, logo, gradient themes, stock or upload photo backgrounds, overlay darkness, and light vs dark text.
Draft test preview
Load the real participant UI with ?test=1 from the live preview column — verify copy, tone, and objectives before sharing a public link.
Projects & company context
Tie studies to product projects and company profile so interviews stay grounded in your product facts and rollout context.
At runtime
A modern chat that still fills your schema
Participants see a polished conversational UI. Behind the scenes the same agent follows your objectives and persists analyst-grade structure.
Adaptive follow-ups
The agent probes naturally based on answers — not a fixed script — while respecting turn limits and wrap-up rules you configure.
Blueprint field capture
Structured slots fill as the conversation progresses so dashboards and exports stay comparable across sessions.
In-chat context cards
When appropriate, the agent shows linked resources or image previews inline instead of dumping raw URLs into the transcript.
Streaming structured notes
Analyst-facing notes update during the session with sentiment, insight types, themes, and participant quotes when signal appears.
CRM-aware sessions
Personalized invite links greet known recipients, support session resume, and feed group filters in analysis.
Voice & multilingual UI
Optional voice input where supported; participant-facing UI strings follow their locale while note language stays configurable per study.
After the conversation
Analysis built for qualitative scale
Insight studies do not end at the transcript. KogniFeed turns conversations into dashboards, filters, and clusters you can act on.
Per-survey KPIs: opens, completions, note volume, sentiment split
Session drill-down: full chat, field captures, and form-style aggregates
Compare date ranges or CRM groups side by side
Canonical topic clusters and notable quotes after post-interview embedding
Publish & integrate
Meet participants where they are
Share a link, embed on your site, or invite known participants from CRM — same agent, same analysis pipeline.
Public survey link
Branded slug URL with optional QR-friendly sharing. Works for ad-hoc research and tracked CRM cohorts alike.
Website embed
Launcher or inline embed with origin allowlist policies — ideal for in-product feedback without sending users elsewhere.
CRM invitations
Send tracked invites to segments, see who completed, and filter analysis by group or time window.
Start an Insight study
Create a survey, open the dynamic builder, and publish when the test preview looks right.