Meet Research Guide —
your AI research assistant.
A dashboard-first partner that listens to your research problem, thinks through it with you using your team's wiki and Qualz docs, drafts the study when you're ready, and helps you analyze results with cited quotes. Always grounded. Always a human on standby.
Research Guide
Attached: Onboarding v3 (survey · 214 responses)
Four capabilities, one chat surface.
Research Guide is the same conversation from problem framing through draft to analysis — not four separate tools stitched together.
Grounded Answers
Ask methodology or product questions and get replies cited back to the Qualz wiki and platform docs. No hallucinated feature claims.
Drafts On Demand
When the brief is rich enough, Research Guide proposes a survey or interview draft. Confirm, revise, or reject — nothing is written until you say so.
Chat Across Your Studies
Pin up to five surveys or interviews per session. Ask counts, tag-slices, quotes, and themes — or triangulate across studies, flag contradictions, and track questions longitudinally.
Compute In Chat
Mid-weight analytical questions (correlations, cross-tabs, themes across transcripts) run in a session-scoped sandbox and come back as quoted answers.
From fuzzy question to cited answer.
A single conversation carries you across the research workflow, with a brief accruing quietly in the background.
State The Problem
Click the purple Research Guide button from anywhere on your dashboard. Describe what you're trying to decide in your own words — no forms to fill.
Think It Through
The guide pulls from your Qualz research wiki and platform docs, cites sources, and proposes next steps — survey, interview, AI participants, or a mix.
Draft The Study
When the brief is specific enough, a draft preview appears. Read the survey or interview guide, revise, or reject with a reason.
Analyze With Citations
Attach a study once responses come in. Ask quotes, counts, and themes — every answer comes back with participant IDs and quoted evidence.
Everything wired into one agent.
Tool-calling loop, not a single-shot reply. The guide decides what to read, when to draft, and when to compute.
Knowledge-Grounded Chat
- Answers grounded in the Qualz research wiki
- Platform docs lookup for feature questions
- Every citation expands to the exact excerpt
- 'Talk to a researcher' escape hatch at any time
- Starter prompts tuned to real decisions
Drafting & Proposals
- Survey drafts with method-fit rationale
- Interview guide drafts with rubric-quality questions
- AI-participant plans flagged as simulation, not evidence
- Confirm / Revise / Reject on every proposal
- Guide chooses when to propose — no forced buttons
Multi-Study Attach & Q&A
- Attach up to 5 surveys or interviews per session
- Triangulate themes, flag contradictions, draft cross-study reports
- Longitudinal comparisons across studies run at different times
- Swap studies in-place without losing chat history
- Tag-aware refusal on slices the data doesn't support
In-Chat Analysis
- Quote specific open-ended survey answers
- Fuzzy + keyword search across interview transcripts
- Cross-transcript themes with quoted evidence
- Correlations, cross-tabs, charts in a session sandbox
- Primary vs. enhanced vs. translated transcript resolver
Fits the way you work, not the other way around.
Minimize to keep it out of the way. Drawer mode pairs it with a study detail page. Full-screen for deep analysis. State is remembered per user.
Tucked Away
A small button in the corner. Click to expand; the conversation survives navigation.
Side Panel
A panel on the right so you can browse the dashboard with the guide alongside.
Full Screen
Focus mode for long drafts or deep analysis. Your size preference is remembered.
Rigor over flash.
AI is only useful in research if you can trust the answer. Research Guide is designed around four commitments.
Everything Is Cited
Wiki references link to the source page. Data claims quote participant IDs and response rows. If it didn't come from your data or the docs, the guide says so.
Tag-Aware Rigor
Ask for a demographic slice on a tag the study didn't collect, and the guide refuses the answer instead of inventing it. Same discipline on survey and interview paths.
Transcript Discipline
When enhanced or translated-to-English transcripts exist, the guide reasons over those. Primary stays primary — the resolver is explicit about which variant won.
Humans On Standby
Whenever the topic drifts outside platform scope, or you sound stuck, the guide surfaces a Qualz researcher — service@qualz.ai or a Slack channel.
Pairs with the rest of Qualz.
Research Guide reads everything the platform produces — and writes only through the proposal/confirm flow.
Bring your next research question.
See Research Guide think through a real brief, propose a draft, and cite quotes from a live study. 20 minutes — no slides.
