WJX has been building forms for two decades. The library is large, the editions are clear, and exams and assessments come with it, so a conventional survey lands safely there. WeJot does something else: the agent writes question types and logic as code, and once the survey closes you can interview the same respondents.
Both platforms collect responses. The difference lies in instrument freedom and how far a study can go. WJX draws question types from a preset library and stays on the survey and exam side; WeJot generates types and logic, and continues into AI interviews with the same respondents once fieldwork closes. On sample, the same 19-question study with women 50+ and 100 completes came to RMB 2,289 against RMB 260.
WJX is the first survey tool many teams ever used. Its edge is not one feature but two decades of maturity: plenty of templates, careful quota management, clear enterprise tiers, and adjacent modules for exams and 360 reviews.
The product assumes a questionnaire is assembled in a UI: question types from a preset library, logic from a settings panel, capacity and collaboration from a higher edition. That works well for conventional surveys. The cost shows up when a study needs an interaction the library lacks. You fall back to the nearest type and absorb the measurement error yourself.
WeJot replaces that assumption. Question types, logic, routing and presentation are written at generation time, so the questionnaire matches the research goal. The same agent stays on, and once fieldwork closes it can interview the same sample to answer what a summary table cannot.
WeJot generates the survey page itself, so a question type is implemented rather than selected. Conjoint, Kano, MaxDiff, card sorting, quadrant drag-and-drop, image hotspots and conditional visual questions can all be produced, with their data structures and analysis contract written at the same time.
Source：Question-type count from the WJX home page and topic pages
WeJot takes a sentence, such as "if they chose A and rated question three below 4, jump to the follow-up page and shuffle the options", and executes it as code. Complex rules stop being limited by how many controls a panel offers. The trade-off is real: logic lives in code, and it should be walked through with test responses before launch.
Source：WJX logic capabilities from its own site and hands-on use
WeJot also prices by length and label, starting at RMB 0.8 per base demographic label with scarcer labels stepping up, and with no membership gate or small-volume surcharge; the quote is visible before ordering.
Source：Live quote from the WJX sample request page on 2026-09-15
WeJot treats self-serve fieldwork as one set: audience labels, crossed and independent quota, start and end times with daily windows, pause and resume, full-quota alerts, identity and liveness checks, longitudinal waves with automatic gap filling, and connected overseas panels. There is one screen to configure it on.
Source：Our capabilities from the panel capability list and the ordering screen
WeJot keeps both ends in one project. The survey establishes what is happening; respondents from the same pool who opt in then run an AI voice or video interview to explain why. Both kinds of evidence sit in one report, and each conclusion links back to its source.
WeJot works on both layers. On the panel side: identity and liveness checks, security verification, attention checks, location limits, in-session supervision, logic consistency, clustering limits and history-based exclusion of repeat respondents. On the survey side: crossed and independent quota, with progress, timing and daily windows visible as they run.
Competitor capabilities come from their own sites and help centers. Sample prices come from quoting the same questionnaire on the same day.
The questions that arise most often during an evaluation, answered directly.
Ninety-plus question types covers conventional surveys. When a study needs an interaction the library lacks, the nearest type becomes the compromise, and the researcher absorbs the error. WeJot generates the type and its execution logic on the spot, so conjoint, Kano, MaxDiff, card sorting and image hotspots are available with their data structures defined up front.
WJX has both an AI interview mode and human depth interviews inside its sample service, so both routes exist. WeJot differs in form and integration: voice, video and text, resumable sessions, voice selection, on-camera interaction and emotion analysis, with anyone who opted in during the survey invited directly and interview findings sitting beside the survey data in one report. After fieldwork, regression, reliability and validity analysis, cross-tabs and clustering run in the same conversation, and deliverables come out as Word, Excel, PPT, PDF or HTML on request.
A WJX survey caps at 100,000 completes (30,000 for polls), quota rules stop applying beyond 50,000 completes and a survey is limited to 50 active rules, and Word/PDF response export is batch-limited to 2,000 on the free edition. WeJot sets no response ceiling, no quota expiry threshold, and generates exports and deliverables on demand, with a data export for further modelling in SPSS or similar.
Once responses are in, regression, reliability and validity analysis, cross-tabs and clustering run in the same conversation, with conclusions, charts and deliverables produced together: Word, Excel, PPT, PDF or HTML on request, or a data export for further modelling in SPSS or similar. Survey, interviews and data stay in one project, and every step is traceable.
Tencent Survey ships with a million-strong respondent group. WeJot differs in two places: question types and logic are generated on demand rather than limited by the available settings, and audiences beyond your own channels can be ordered with targeting, quota and quality controls attached. On the same 19-question study with women 50+ and 100 completes, fieldwork came to RMB 1,700 against RMB 260.
Read the comparison Panel and survey platform · ChinaCredamo’s strength is granularity: every quality control switches on individually and is priced separately, which suits the tightly separated conditions common in academic work. WeJot makes the opposite trade: quality control runs as one default set, pricing is linear and itemised, and every targeting label is available on a free account. On the same 19-question study with women 50+ and 100 completes, the real cost was RMB 563 against RMB 260.
Read the comparison AI interview platform · ChinaBoth products run AI depth interviews, with different levels of price transparency: Mizzen publishes no price list and gates trials behind review, while WeJot prices per minute and per item, visible before ordering. The larger difference is scope: WeJot keeps the survey and the interview in one project, so the distribution comes first, the reasons follow, and one traceable report closes the study.
Read the comparison
Create a project and keep the questionnaire, fieldwork, interviews and report inside it, with costs visible before you order.
