# Elicit

> Elicit supports academic research with a Research Agent, research reports, and a structured workflow for screening papers and extracting evidence.

Website: https://elicit.com/

Elicit is an academic-research option to evaluate when your deliverable needs a selection record and a study evidence table. Begin with a question and eligibility criteria so you can judge whether the resulting papers answer the intended question.

## Current research workflows

Elicit's current help distinguishes Research Agent, Research Reports, and a paid Systematic Review workflow. These names describe vendor offerings, not demonstrated review completeness or accuracy. [Elicit workflow documentation](https://support.elicit.com/en/articles/14757543-getting-started-with-elicit-which-workflow-should-i-use)

Choose a workflow around the artifact you need. A research overview, a candidate bibliography, and a screened evidence table require different checks. Save what was actually searched and supplied, then keep proposed judgments separate from reviewed judgments. A polished report does not retrospectively create the missing search log or protocol.

## An illustrative research exercise

This is an **illustrative fictional scenario**, not an Elicit trial or a report of real findings. A team wants a bounded literature brief about measured onboarding task completion among newly hired staff. Prepare a criteria file and a bibliography, then inspect the candidate reports against the criteria.

Keep report IDs separate from study IDs. If a follow-up explicitly describes the same study, connect the reports before comparing outcomes. If an abstract omits the population or measure, record an unresolved eligibility question and seek the full source. Do not fill the gap with a likely-sounding method.

For selected sources, extract the method, population, outcome, unit, time point, and source location into a table. Open the relevant paper passages for the fields the conclusion relies on. If two reports conflict, retain both locators and the unresolved discrepancy. Do not average incompatible outcomes simply because they fit a common column.

The [AI literature-review workflow](/guides/workflow/ai-literature-review-workflow) supplies a protocol, logs, and checks for this process. [Evidence Synthesis](/glossary/evidence-synthesis) explains the difference between organizing several findings and summarizing one paper.

## Availability and limits

Elicit offers a free Basic plan alongside paid and organization options; this directory classifies it as freemium. Check the intended workflow's current entitlement. [Elicit pricing](https://elicit.com/pricing)

Do not treat plan classification or the name “Systematic Review” as proof that your work meets a formal standard. Your protocol, search coverage, eligibility decisions, source checks, and reporting remain part of the review. This profile does not establish exhaustive coverage or replace the expertise required for a formal review.

Record the limits of the actual source set. An abstract-only record should remain abstract-only; an unperformed errata or retraction check remains unverified. File structure checks can find missing cells and broken references, but they cannot prove that an extraction faithfully represents a study.

## Adjacent tools

[NotebookLM](/tools/notebooklm) is a possible substitute for studying a selected set of provided papers, with different search coverage and workflow scope. [Perplexity](/tools/perplexity) serves a broader web-discovery role. These are role comparisons, not accuracy rankings.

If you already have the selected sources, consider starting with a bounded document workflow and a reviewed evidence table. Evaluate Elicit when the additional research stages match your task, and keep the original sources available regardless of which interface helps organize them.

---

_Source: https://agentscamp.com/tools/elicit — Tool on AgentsCamp._