All ideas

In addition to the main behavioral question, I think allowing a user to add more context up front would help to generate the right survey questions. Perhaps even nudging them to add their own hypotheses.

For example, maybe I'm an architect who is designing a new floorplan. I'd like to run a DCE on new architecture floorplans to determine which features cause buyers to purchase my apartment unit. Another example, maybe I want to create programmatic...

Many test users would like their experiments to be private, others would like the ability to easily share their experiments. This could be either a live link to an R Shiny instance, or this could be an exportable PDF that we can use/share for expe...

Have different levels of visibility and ability to lock report

In order to build higher-quality population and market simulations. Integrate with third party data such as Canny, Clickup, Chorus.ai, G2, Github, Google Forms, HotJar, ProductBoard, Qualtrics, Reddit, Salesforce, Slack, Survey Monkey, Typeform. S...

People want to a/b test our capabilities against humans. There are some options including MTurk, and Cint.

One of the key issues is to convince an independent observer that a generated responder is producing results that are better than a random guess. Suppose that we have a study, and in the course of this study respondents complete a standard form. S...

People who run experiments have an idea of what they want to know and might not be thinking of attributes that should also be considered to test specific to their use case... providing additional information to users after the experiment has concl...

There is a huge demand for giving an LLM your own data for analysis. This is something my company is doing via RAG (i.e. customers can interact with their unstructured knowledgebase via a chat interface) but we do not link it to any other data. To...