Case study
Conversion Y – AI powered Surveys for B2B
Last updated:
Building a survey tool around the work of doing research well.
Conversion Y started with frustration at expensive, broad survey platforms that still left important research work to the user. My co-founder and I wanted an accessible alternative shaped by our background in CRO, user testing and statistics: a product that helps people create useful surveys and make sense of the answers.
Freedom in the editor, structure in the data
I work across product design, frontend and backend engineering. The editor allows highly customisable, free-form surveys, while the collection and evaluation backend is designed to keep responses valid, comparable and understandable. Balancing those goals is central: flexibility in presentation should still lead to data people can interpret.
Bringing AI into the whole workflow
The implemented capabilities cover survey creation, translation and evaluation. Surveys can also be embedded into websites, connecting creation, delivery and evaluation in one product. AI assistance supports repeated work such as drafting questions, preparing language versions and interpreting responses.
Automatically generated image description
This conceptual diagram titled 'Conversion Y' outlines a process for improving survey questions and evaluating responses with AI assistance. The layout is divided into two main sections: 'Before the Survey' and 'After Responses Arrive.' Each section contains three horizontally aligned boxes with arrows indicating flow. The first section shows a question draft, an improvement suggestion, and duration guidance in white, pale yellow, and dark green boxes respectively. The second section includes collected responses, automatic evaluation, and reviewable findings, all in white boxes. The background is light beige with minimalistic typography in black, gold, and orange accents.
Our research background also shapes the guidance and guardrails. A survey editor needs to help people understand the choices that affect their results. Simply offering every possible question type does not by itself help someone run a useful survey.
An MVP with a clear direction
Conversion Y is an actively developed MVP, with my contribution currently around one day a week. Our aim is to make a dependable research workflow accessible to smaller teams: flexible survey design, AI assistance and a collection and evaluation backend built around the quality of the resulting data.