Video summary

This AI Synthesizes 500 Papers So You Don't Have To (Elicit AI)

Main summary

Key takeaways

Product Review

Product Reviewed: Elicit AI (Academic Research Assistant)

Key Features Mentioned

Core tools available for free (some options greyed/locked)

  • Create research reports from a research question.
  • Find research papers by asking a question.
  • Chat with papers, including:
    • Uploading a paper
    • Asking questions about it

Research report workflow details (free tier demo)

  • Requires selecting a source type:
    • Research papers
    • Clinical trials
  • Provides question strength feedback and recommends refinement
    • More precise questions generally perform better
  • Lets you choose answer detail level:
    • Fast / Balanced / Comprehensive
  • Shows progress during execution:
    • Screening sources
    • Extracting data
    • Then producing the report
  • Report output includes:
    • Abstract, methods, results
    • Synthesized summaries and references
    • Table-style extracted information, including a “heat map”-like summary derived from extracted study characteristics
  • Runtime example cited: ~4 seconds for report generation (demo)
  • PDF export is implied for the demo, but later clarified as upgrade-gated (see Cons).

Paid tiers / advanced capabilities

  • Plus
    • Described as limited / not unlocking the “powerful stuff”
    • Includes additional downloads and some extra features
  • Pro ($49/month)
    • Unlocks the “powerful stuff,” including:
      • A full systematic review workflow (step-by-step: sources → screening → extraction → research report)
      • Workflow expansions, including items in beta

Systematic review workflow (Pro demo)

  • A guided, sequential process:
    • Define screening
    • Recommended steps
    • Abstract screening
    • Data extraction
    • Final research report
  • Supports sourcing from:
    • Elicit search
    • Your library
    • Connected Zotero
  • Also supports inputs like RIS / PDF / Bib

Data extraction limitation (in the paid workflow)

  • Extraction appears geared toward structuring findings into tables/columns
  • It does not directly produce outputs like “here’s the exact figure/table with numbers.”

Additional workflows mentioned (Pro / beta context)

  • Competitive landscape (beta)
  • Research landscape (requested to be highlighted)
  • Clinical trial access analysis
  • General research agent (ask a question, receive responses)

Alerts (paid tier; reviewer complaint)

  • Create instant email alerts for new sources on topics
  • Can filter using:
    • Journal quality (reviewer cares about Q1/Q2)
    • Keywords
    • Exclusions (e.g., require/deny terms such as “abstract does not contain …”)
  • Framed as saving time by reducing repetitive manual literature searching.

Pros (User Experience & Strengths)

  • Strong usability improvements over time (“OG” tool now “leaps and bounds” better).
  • Clean, modern interface with a larger input area than many chat bots.
  • Works well for free, including:
    • Generating structured research reports
    • Finding papers via questions
    • Chatting with uploaded papers
  • Built-in guidance:
    • Evaluates question strength
    • Helps refine prompts
    • Shows screening/extraction progress
  • Good report structure and readability:
    • Abstract/methods/results plus synthesized findings
    • Structured outputs like tables and heat-map-style summaries
  • Systematic review workflow described as “powerful” for guided, step-by-step use (especially for beginners or those without a process).
  • Alerts are positioned as a major productivity win (even though they’re gated).

Cons (Limitations & Frustrations)

  • Free tier is restricted:
    • Some options greyed out
    • Reviewer implies free access lacks key features like export/power tools
  • Question accuracy matters:
    • Demo showed at least one incorrect interpretation (example: “OPV” initially misunderstood, then corrected with another query)
  • Pricing perceived as expensive/stingy:
    • $49/month described as “a little bit cheeky” and “expensive”
    • Reviewer suggests PhD students may find it unsuitable unless funded
    • Plus tier provides little value
  • Alerts are gated (“a bit of a shame” there isn’t a free/cheaper option).
  • Extraction expectation mismatch:
    • User expected direct extraction of underlying numbers/figures/tables
    • Instead, Elicit emphasizes workflow-based structuring into columns.

Comparisons Made

  • Manual literature searching using tools like Google Scholar and repeated search cycles.
  • Mentions another tool reference (value comparison), described as:
    • “Elicit’s familiar workflow” / “size space as well”
    • Also states another tool is “probably a little bit better than this one” for value-for-money (no clear product name beyond that).

Ratings / Numerical Scores

  • Price mentioned: $49/month for the Pro tier
  • Runtime mentioned: report generation completed in ~4 seconds (demo)
  • Pro workflows usage: reviewer references usage limits like “12 workflows” (no additional quota details provided).

Unique Points Mentioned (Consolidated)

  1. Elicit described as an “OG academic AI tool” with major improvements.
  2. Free tier includes research reports, paper search, and chat-with-papers.
  3. Free UI shows greyed-out features indicating gating.
  4. Free research reports require selecting source type: papers vs clinical trials.
  5. Prompting guidance includes question strength evaluation and prompt refinement.
  6. Hovering shows whether the selected operation matches the question.
  7. Demo showed an initial wrong interpretation of “OPV.”
  8. Free reports offer answer modes: Fast / Balanced / Comprehensive.
  9. Live visibility into screening/extraction progress.
  10. Outputs include sections like abstract/methods/results plus synthesized references.
  11. Demo indicates reports can cover up to 50 papers (snapshot claim).
  12. Table-style extraction with heat-map-like summaries is shown.
  13. Paper search provides “too long, didn’t read” summaries.
  14. Export appears upgrade-gated (despite earlier PDF mention—no clear reconciliation).
  15. Paper chat is available after uploading a paper.
  16. Plus tier adds downloads/extra features but is “not much.”
  17. Pro tier unlocks advanced/beta capabilities at $49/month.
  18. Pro includes systematic review workflow (guided step-by-step process).
  19. Systematic review supports sources from search, library, Zotero, and RIS/PDF/Bib.
  20. Pro data extraction focuses on tables/columns, not direct figure/table number extraction.
  21. Other workflows: competitive landscape (beta), research landscape (requested), clinical trial access analysis, general research agent.
  22. Elicit’s library supports uploading and reusing papers across workflows.
  23. Alerts can email notify for new sources with filtering/exclusion rules.
  24. Alerts gating is described as unfortunate due to cost.
  25. Overall: Elicit is powerful but expensive; worth it mainly if you’re funded.

Speaker / Role Perspectives

  • Single main reviewer perspective
    • Highlights academic usefulness, UI improvements, and free-tier strengths
    • Critiques pricing/tier structure and gated productivity features
    • Notes expectations around extraction output format
    • Mentions budget considerations (e.g., PhD student affordability) and value comparison to another tool.

Concise Verdict / Recommendation

Elicit is a strong academic research assistant—particularly for structured research reports and guided systematic reviews—but it’s expensive. Key productivity features (notably alerts and full power) are gated behind higher tiers like $49/month Pro.

Recommended if you have funding (university/research group) or heavily rely on systematic review workflows; otherwise, consider alternatives or manual approaches if you’re budget-constrained and rely on free access.

Original video