What Maze does
Maze is a product-research platform for unmoderated prototype and live-site testing, surveys, card sorting, recruitment, reporting, and AI-assisted study creation and synthesis.
Maze shortens the loop between a design question and observable participant behavior. Teams can test prototypes and websites, run surveys, card sorts, and tree tests, recruit participants, and inspect paths, misclicks, completion, timing, open text, and reports. Templates and AI assistance help draft studies and summarize responses, while integrations keep research close to design workflows.
The free tier supports evaluation, while paid plans vary by studies, responses, collaborators, recruitment, recordings, integrations, and enterprise governance. Participant-panel incentives or recruitment can be separate from platform access. Teams should confirm current response definitions, overages, prototype limits, data retention, SSO, and export before estimating a continuous-research program.
Unmoderated metrics do not explain motivation by themselves. Task wording can lead participants, prototype fidelity can create false failure, and completion rates can hide confusion. AI summaries may erase minority behavior or turn a small convenience sample into certainty. Use neutral tasks, pilot the study, obtain consent for capture, minimize PII, segment intentionally, inspect sessions and raw responses, and triangulate with moderated or production evidence.
How Maze works
Researchers connect a prototype or live experience, define tasks and questions, recruit or invite participants, and collect paths, clicks, completion, time, responses, and recordings where enabled. Maze aggregates metrics and AI can assist study design or summarize qualitative responses for human interpretation.
Turn a decision into neutral tasks
Define audience, hypothesis, prototype fidelity, success evidence, and sample. Draft tasks that do not reveal the expected path and disclose recording, AI processing, PII, incentives, and withdrawal.
Collect behavior and responses
Participants navigate a prototype or live experience and provide clicks, paths, completion, time, text, and eligible recordings. Technical issues and task interpretation can look like usability problems.
Inspect metrics and raw sessions
Maze aggregates results and AI can summarize qualitative input. Researchers inspect failures, paths, clips, segments, contradictions, outliers, and sample quality rather than treating completion as explanation.
Report evidence and uncertainty
Findings state method, sample, fidelity, exclusions, and alternative explanations, link to source evidence, and define follow-up research or production telemetry. Sensitive participant data remains restricted.
How to set up Maze
Write the decision question
Define the product decision, target participants, evidence needed, risks, sample limits, and what result would change the team's direction.
Prepare consent and stimuli
Use an authorized prototype or site, disclose recording and AI processing, avoid real credentials, and remove unnecessary personal data.
Build and pilot neutral tasks
Avoid revealing the expected path, test the study internally, verify branching and success criteria, and check mobile and assistive experiences.
Review raw behavior
Inspect paths, clips, failures, open text, segment differences, and technical issues before accepting aggregated metrics or AI themes.
Report qualified findings
State method, sample, uncertainty, exclusions, and contradictions; link claims to evidence and define follow-up research or telemetry.
Maze FAQs
Is Maze free?
Maze offers free evaluation access with limits. Paid plans expand studies, responses, collaboration, recruitment, integrations, and governance.
Can Maze recruit participants?
Maze offers participant-recruitment options, with availability, audience criteria, incentive, and pricing separate or plan-dependent.
Does task completion prove usability?
No. Prototype fidelity, wording, guessing, technical issues, and sample bias affect completion; inspect paths and qualitative evidence.
Can AI write the whole study?
It can assist drafts and summaries, but researchers must verify neutrality, validity, consent, accessibility, sample fit, and evidence.
Listing reviewed 2026-07-15. Product details and pricing can change; verify important terms on the provider's website.
Related Research AI tools
Related AI guides
Reviews
Tell the community what you made, what worked, and what you wish you knew before starting.