We're introducing suggested queries in Elicit by merging the Brainstorm research questions task with the Literature review workflow! 

When you click on the query bar after running a search, you will find suggestions pop up below.

Literature review is the primary workflow for Elicit and the one you see prominently featured on the home page. But there are other tasks in Elicit, which you can access on the Tasks page. 

The most popular of these tasks is Brainstorm research questions. This task takes a research question and uses GPT-3 to generate related questions. Originally, it was designed to help people break down vague questions into more concrete subquestions. 


  1. The "non-literature review" tasks are significantly less developed and our team has not made updates to them recently.
  2. Queries are suggested independent of results. Right now, we are not suggesting queries that we think will get you better results. 

The vision
Our medium-term vision with Elicit is to surface the basic building blocks of reasoning to let you automate custom workflows, or spin up custom "research assistants."

Literature review is the first. Right now our core team is identifying and building out building blocks like search, classification, extraction, and summarization. We're stringing them together to produce literature review.

As we refine this infrastructure, the process will become smooth enough that you can mix and match them to automate your own reasoning workflows. 

Today you have to research one question at a time. Imagine an Elicit that enables you to launch parallel literature reviews for many questions simultaneously and synthesize the results across all of them! 

Integrating Brainstorm research questions into Literature review is the first exciting step towards this vision. 

Try the suggestions here and let me know what you think.

If you don't want to get these emails, you can update your preferences or unsubscribe from this list.