Wavefield Research Inc. is an Ontario, Canada company founded in July 2026. We build an AI-native survey platform for market research and political polling — surveys programmed by an AI agent, results analyzed to a standard you can defend.
Survey software has long forced a choice. Form builders make collecting answers easy, but they can't weight a sample, test significance, or manage quotas — the things that separate a defensible finding from a spreadsheet of percentages. Enterprise research platforms do all of that, behind annual contracts, per-seat licenses, and sales calls, priced for organizations that run research programs, not for teams that need one study done properly.
Wavefield covers the middle: one study at a time, at a flat per-project price. An AI agent programs the questionnaire from a plain-language brief — skip logic, quotas, rotations, translations — and every project includes the analysis layer that usually costs an analyst: weighting, crosstabs with significance testing, quality controls, and SPSS, Excel, and Word exports.
These aren't policies in a handbook — they're how the platform is built, applied identically to every project including our own:
Every report shows weighted percentages alongside unweighted base sizes. A weighted count doesn't correspond to actual people, so we never print one — and a report that hides its real bases deserves suspicion, including ours.
Crosstab significance letters and margins of error are computed on Kish effective bases — the sample size you actually have after weighting — not the raw count. Weighting always costs precision; pretending otherwise flatters the data.
Most online research uses quota samples, where the classic margin of error is a convention rather than a guarantee. We report it that way, computed on the effective base, instead of implying probability-sample precision that isn't there.
Speeders, straightliners, and failed attention checks are flagged for review — exclusion is always a deliberate, recorded decision by the researcher, and excluded rows stay in raw exports, labeled. A dataset you can't audit isn't a dataset.
Screen-outs and quota terminations never count toward a quota cell's fill, so your cells contain exactly the completed interviews they claim to.
Every invited respondent gets a personal, single-use link — no forwarded links, no duplicate completes, and screeners terminate without telegraphing the screen-out.
The reasoning behind these choices is written up in our guides: how to analyze survey results, quota sampling vs. random sampling and what survey scripting involves.
Wavefield is a young company, and we'd rather say so than imply otherwise. The platform is strictly self-serve with email support; there's no sales team, no onboarding calls, and no consulting tier. Our infrastructure providers hold SOC 2 certification; Wavefield itself is not yet independently certified, and we say exactly that on our security page along with what is and isn't protected today.
What we ask you to judge us on is the product itself: publish a survey, read the methodology in the reports, and check whether the numbers are presented the way this page promises.
Create an account, upload a brief, and see what the agent builds. You don't pay until you field.
Start a project