What Is Maze?
Maze is an end-to-end user research platform whose defining architectural choice is a three-pillar design: Recruit, Research, and Analyze, connected inside one hub. The premise is that participant sourcing, study execution, and automated reporting should not live in three disconnected tools — Maze pulls them onto one surface so a design team can recruit, run, and report a study without leaving the platform. The company frames the value as research “at the pace of change,” turning insights into clarity a team can act on inside a sprint cycle.
The Research pillar runs a multi-method UX suite. Prototype testing is the legacy anchor — Maze grew up validating Figma prototypes, and a heavy Figma plugin still converts prototypes into testable studies in a few clicks. Around that anchor sit surveys, live website testing, mobile testing, card sorting, tree testing, and moderated interviews, plus a newer AI Moderator feature that runs moderated-style sessions. The Analyze pillar turns raw sessions into click and tap heatmaps, task-completion metrics, and AI-assisted synthesis. The Recruit pillar offers a 6M+ proprietary panel, bring-your-own-community connections, and AI-powered recruitment across B2B and B2C audiences.
The methodology is unmoderated by default. Maze’s instinct is to put a study in front of many participants without a researcher in the room, then read the behavioral residue — clicks, taps, completion times, abandonment points — back as signal. AI-powered automation threads through all three pillars: recruitment matching, the AI Moderator, and insight synthesis. The integrated recruit-research-analyze platform is the product in one phrase — a single hub for behavioral usability research, not a depth-interview instrument.
How Fast Does Maze Deliver Results?
A design team evaluating Maze usually feels the clock on prototype iteration: a Figma file is ready, engineering is waiting, and the question is whether the flow works before anyone commits sprint capacity. Maze is built for exactly that tempo — unmoderated testing collects responses in parallel, so a prototype study fields dozens of participants in days rather than the weeks a moderated calendar would take.
Speed depends heavily on whether the account is already set up. The two profiles diverge sharply:
- Established Maze account, Figma file ready, panel access configured. A prototype test launches in under an hour: the Figma plugin imports the prototype, the study builder adds task instructions, and the 6M+ panel or a connected community fills participants the same day. Results stream into heatmaps and completion metrics as participants finish. End-to-end on a 30-50 response prototype test: often 1-3 days.
- New buyer, no panel configured, AI Moderator not yet provisioned. The runway is longer. The free plan covers unmoderated testing immediately, but provisioning the AI Moderator requires a Business or Org plan and a sales conversation. Panel recruitment is a separate billing line that may need scoping. First study can slip a week or more behind a procurement step.
The honest read: Maze is fast for the unmoderated behavioral study it was built for, once the account exists. The bottleneck is rarely the test — it is panel provisioning and, for AI moderation, the plan upgrade. Behavioral throughput is quick; depth that needs the gated AI Moderator carries a procurement tail.
What Does Maze Cost?
Maze does not publish pricing. Per buyer-reported references, the structure has three layers. A free plan covers basic unmoderated usability testing — enough to validate a single prototype flow at low volume. Paid plans scale by seats and usage; the AI Moderator feature is gated to Business or Org tiers, estimated near $15,000 per year per buyer-reported references. Panel recruitment is billed on top of the subscription rather than bundled into it.
The practical consequence: a self-serve buyer can start free, but cannot forecast the annual cost of the capability they actually want — AI-moderated research — without a sales call. For the full cost-by-frequency breakdown at 1, 5, 10, 20, and 50 studies per year, see the Maze pricing reference guide.
Want to compare cost before a sales call? Run three free User Intuition interviews → — published pricing, no procurement step.
How Deep Does Maze Go in Each Interview?
The depth question is where a buyer learns whether Maze answers “does this design work” or “why do customers decide as they do.” Maze is built for the first question. Three sub-paragraphs frame what that means in practice.
Format — unmoderated by default. Maze’s native study is unmoderated: a participant works through a prototype, survey, card sort, or tree test alone, and the platform records the behavioral residue. There is no researcher probing a surprising answer in the moment. Depth, in the unmoderated model, comes from volume and instrumentation — many participants, precise click and tap heatmaps, task-completion rates, time-on-task — not from conversation. The data tells you where friction sits with statistical confidence; it does not tell you why the friction exists.
AI Moderator — the newer feature. Maze added an AI Moderator that runs moderated-style sessions and asks follow-up questions. It is a genuine capability, but it is an addition layered onto a behavioral platform, not the core. It is gated to Business or Org tiers, and buyer-reported references describe it as oriented toward open-ended Q&A after a task rather than systematic motivational laddering during one. It widens the question set; it does not convert Maze into a depth-interview instrument.
Synthesis behavior. The Analyze pillar synthesizes sessions into heatmaps, completion metrics, and AI-assisted theme summaries. The output is behavioral and task-anchored: which step lost users, which element drew attention, which variant completed faster. That is decision-useful for interface optimization. It is not the motivational architecture — the expectations, prior experiences, and identity-level drivers — that strategic product and positioning decisions need. Maze’s depth is task-based and behavioral by design; reaching motivational depth means a different methodology than the one the platform optimizes for.
How Does Maze Scale to Your Research Volume?
Scaling Maze means asking three separate questions — how far the audience reaches, how the cost behaves as study frequency rises, and how access spreads across a team. Each has a different answer.
Audience scaling — 6M+ panel plus BYO. Maze’s Recruit pillar offers a proprietary panel of 6M+ participants across B2B and B2C audiences, AI-powered recruitment matching, and bring-your-own-community connections for teams that already have a user base to draw on. For a behavioral usability study, that is ample reach — most prototype tests need tens of participants, not thousands. The caveat: panel recruitment is billed separately from the plan, so audience scale carries an additional line item that grows with how much sourcing the team offloads to Maze.
Frequency scaling — tiered subscription. Maze is a subscription, not a per-study purchase. Once a team holds a Business or Org plan, running the second, tenth, or fortieth unmoderated study of the year adds no plan cost — the marginal study is effectively free against the seat. That makes Maze economical for teams running continuous, high-cadence usability testing. It makes Maze expensive for teams running a handful of studies a year, because the annual floor is paid whether the team runs four studies or forty.
Team scaling — seat-driven. Access spreads by seats. Adding researchers, designers, or PMs to the platform means adding paid seats, and pricing scales with seat count. For a centralized design-research team, that is predictable. For an organization that wants research democratized across many functions — marketing, support, product, each commissioning studies independently — the seat model becomes a budgeting constraint, because every additional commissioner is an additional seat.
How Useful Are Maze’s Insights — and Do They Compound?
Insight value splits into two questions: how good is the output of a single project, and does anything accumulate across projects into a durable asset.
Per-project insight quality. For its lane, Maze produces strong project output. A prototype test returns click and tap heatmaps, task-completion rates, time-on-task, misclick maps, and abandonment points — a precise, quantified picture of where an interface succeeds or fails. The AI-assisted synthesis in the Analyze pillar packages that into shareable reports a design team can act on inside a sprint. For the question “does this flow work,” the per-project insight is genuinely decision-grade. The limit is scope: the insight describes behavior on a tested artifact, not the motivation behind it.
Insight compounding. Maze’s outputs are organized per study. A prototype test from March and a card sort from September are separate reports; the platform does not, by buyer-reported references, maintain a queryable cross-study intelligence layer that lets a researcher ask a plain-language question across every study the team has ever run. Concretely: if a PM wants to know “across every onboarding study we have run in the last year, what consistently confuses new users,” Maze’s structure asks them to open and re-read individual reports. A platform with an ontology-indexed repository — User Intuition’s Customer Intelligence Hub is one — answers that question as a single query. Maze’s insight is fresh and useful per project; it does not compound into an institutional memory the way a cross-study intelligence layer does.
How Does User Intuition Approach the Same Dimensions?
User Intuition runs the same five dimensions through a different architecture. Where Maze is an integrated platform for unmoderated behavioral usability testing, User Intuition is a native-AI depth-interview platform: every study is an AI-moderated conversation using adaptive 5-7 level laddering, priced per interview, with a cross-study intelligence layer underneath. The contrast below is dimension by dimension.
Speed
Where Maze’s speed depends on a configured account plus, for AI moderation, a plan upgrade and a sales conversation, User Intuition starts the clock at signup. Three free interviews launch in five minutes with no card. A full study runs against the 4M+ vetted panel and returns themed results in 24-48 hours — depth research at the tempo Maze reserves for unmoderated behavioral tests.
Cost
Where Maze gates AI moderation to roughly $15,000-per-year tiers and bills panel recruitment on top, User Intuition publishes its pricing: $20 per audio interview, $40 video, $10 chat, with a 10-interview study at $200. There is no annual floor and no separate panel invoice — recruitment from the 4M+ panel is included. A buyer forecasts the full cost in a spreadsheet before any sales contact.
Depth
Where Maze’s depth is task-based and behavioral — clicks, completion, heatmaps — User Intuition’s depth is motivational. Every interview applies adaptive 5-7 level laddering, moving from a concrete behavior through functional consequences to emotional drivers and identity-level values. The deliverable explains why a user abandoned a flow, not only that they did.
Scale
Where Maze scales by seats and a separately billed panel, User Intuition scales by studies. Any team member can commission research with no seat fee; cost tracks actual interviews run. The 4M+ panel across 50+ languages reaches churned customers, prospects who chose a competitor, and category non-buyers — segments an unmoderated prototype test cannot recruit.
Insights
Where Maze delivers per-study reports, User Intuition feeds every interview into the Customer Intelligence Hub — an ontology-indexed repository answering plain-language questions across every past study. Insight compounds into institutional memory rather than resetting each project.
Side-by-side at a glance
| Dimension | Maze | User Intuition |
|---|---|---|
| Core methodology | Unmoderated usability testing | AI-moderated depth interviews |
| Defining architecture | Three-pillar recruit-research-analyze hub | Native-AI laddering + intelligence hub |
| Depth mechanism | Click/tap heatmaps, task metrics | Adaptive 5-7 level laddering |
| Pricing model | Not published; tiered subscription | Published: $20/audio interview, $200/study |
| AI moderation access | Gated to Business/Org (~$15K+/yr) | Included; 3 free interviews on signup |
| Panel | 6M+ panel, billed separately | 4M+ vetted panel, included |
| Languages | Not disclosed on site | 50+ languages |
| Speed to results | 1-3 days unmoderated, once configured | 24-48 hours end-to-end from signup |
| Cross-study layer | Per-study reports | Customer Intelligence Hub (queryable) |
| Best-fit research object | Prototype / interface validation | Customer motivation and decision drivers |
How Do Maze and User Intuition Compare on Security and Compliance Posture?
Security posture is where procurement teams gate a purchase, and both platforms should be verified directly during scoping rather than taken from a marketing page.
| Posture dimension | Maze | User Intuition |
|---|---|---|
| Documented today | SSL encryption, GDPR, SSO, AWS infrastructure, RBAC, private workspaces | GDPR-compliant, ISO 27001-aligned, SOC 2 audit in progress |
| Certification to verify | SOC 2 / ISO / HIPAA not explicitly on homepage — confirm in scoping | Confirm current SOC 2 attestation status in scoping |
| Sub-processor disclosure | Request from vendor | See the User Intuition security overview |
| Participant PII surface | Panel + BYO community data; AWS-hosted | 4M+ vetted panel; fraud-prevention layer on participant data |
| Customer data export | Confirm export formats and ownership terms in scoping | Study data and transcripts exportable |
| AI training on your data | Confirm in scoping | Confirm in scoping; not used to train external models |
Maze presents an enterprise-grade governance posture — SSL, GDPR, SSO, role-based access control, AWS infrastructure, and private workspaces are all documented. What is not explicitly visible on the homepage is formal SOC 2, ISO 27001, or HIPAA attestation. That absence is not evidence Maze lacks them — many platforms hold certifications they surface only in a security packet — but it does mean a regulated-industry buyer must request the current attestation list directly before contracting. User Intuition is GDPR-compliant and ISO 27001-aligned with a SOC 2 audit in progress; its sub-processor list and data-handling terms are documented in the security overview. For both vendors, the correct move is the same: ask for the current certification list in writing during scoping rather than inferring it from a homepage.
How to Choose Between Maze and User Intuition
The choice resolves cleanly once three questions are answered: what is the research object, how often does research run, and what operating model does the team want.
By research question:
| Your question is about… | Better fit |
|---|---|
| Whether a prototype or interface flow works | Maze |
| Why customers decide, churn, or choose a competitor | User Intuition |
| Behavioral metrics across design variants | Maze |
| Motivational drivers behind a behavior pattern | User Intuition |
By research frequency:
| Cadence | Better fit |
|---|---|
| Continuous, high-volume usability testing | Maze (subscription amortizes) |
| Occasional or project-based research | User Intuition (per-study, no floor) |
By operating model:
| You want… | Better fit |
|---|---|
| One hub for recruit-research-analyze, seat-based | Maze |
| Published pricing, no annual contract, democratized access | User Intuition |
Two-platform answer. For many design-led organizations the honest recommendation is not either-or. Maze earns its seat for continuous prototype and interface validation inside the design sprint — that is its lane and it runs it well. User Intuition earns its place for the strategic, motivational research that decides what to build next: why users churn, why positioning fails, what would bring a lost customer back. The two answer different questions, and a research practice that runs both gets behavioral precision on the interface and motivational depth on the decision. The cost structures make this practical — User Intuition’s per-study pricing means adding it alongside Maze is a variable line item, not a second annual floor.
Evaluation Questions for Your Maze Demo
Bring these to a Maze demo, organized by the dimension each one probes.
On the three-pillar architecture:
- How tightly do Recruit, Research, and Analyze actually connect — can a study move from recruitment to reporting without manual handoffs?
- Which capabilities live in the free plan, and which require a Business or Org upgrade?
On the AI Moderator: 3. Does the AI Moderator run systematic follow-up laddering during a session, or open-ended Q&A after a task? 4. What plan tier is required for AI Moderator access, and what is the annual cost? 5. Can the AI Moderator probe motivation, or is it scoped to usability follow-ups?
On pricing: 6. What is the total annual cost for our seat count plus the AI Moderator plus panel recruitment? 7. How is panel recruitment billed — per participant, per study, or as a pool? 8. Is there a per-study or per-interview cost we can forecast, or only an annual contract?
On panel and reach: 9. Can the 6M+ panel reach churned customers and category non-buyers, or only general usability participants? 10. How many languages does the panel support for non-English research?
On security: 11. Can you provide your current SOC 2, ISO 27001, and HIPAA attestation status in writing? 12. Where is participant PII stored, and is our study data used to train any AI model? 13. What data-export formats and ownership terms apply if we leave the platform?
Three free interviews. No card. 5 minutes to launch. 5/5 on G2 and Capterra. Try User Intuition → · Compare Maze vs User Intuition → · Maze pricing reference → · 7 Maze alternatives compared → · Migration guide →