Consensus generates AI analysis in Pro messages and Deep reviews to help you understand the research behind your question faster.
Consensus AI analysis includes written summaries with in-line citations, so you can see which papers support specific parts of the answer. Depending on your question, the available papers, and the type of analysis requested, Consensus may also include figures, charts, tables, and custom visualizations.
These tools are designed to help you screen papers faster, understand the evidence more clearly, and verify the sources behind an answer.
Where AI analysis appears
AI-generated analysis appears in:
Pro messages: analyze up to 20 papers and provide a thoughtful summary. Pro is designed for faster, flexible research questions.
Deep reviews: analyze up to 100 papers and are designed for more in-depth, structured analysis across a larger set of research.
In-line citations
Consensus AI analysis includes in-line citations so you can see which papers support specific parts of the answer.
When you hover over an in-line citation, Consensus shows the exact quote used to support that part of the analysis. This helps you verify where the answer came from instead of taking the AI summary at face value.
Each citation connects the analysis back to the paper it came from, making it easier to review the source and decide whether it is relevant to your work.
When a citation includes a checkmark, this means Consensus used the full text of the paper in the analysis.
If there is no checkmark, Consensus relied on the paper’s abstract and available metadata.
Figures, charts, and tables
Consensus may include figures, charts, or tables from cited open access papers when they help support the analysis.
You can also ask Consensus to include figures by adding something like “show figures from the papers” to your question.
Figures help you review the original visual evidence from the paper, including charts, axes, sample sizes, and tables, without having to search through the full text manually.
Visualizations and structured outputs
Depending on your question and the type of analysis requested, Consensus may include visualizations or structured outputs to help you understand the evidence more quickly.
These outputs may appear automatically when they are useful for your search, or you can ask for them directly in your prompt.
Consensus Meter: The Consensus Meter is a visualization for yes-or-no research questions. When triggered, it shows where the research leans across the top papers. Papers and citations are color-coded by stance, helping you quickly see whether the evidence generally supports, does not support, or is mixed on a question.
Claims & Evidence Table: The Claims & Evidence Table breaks down key claims from the analysis and shows the strength of support, reasoning, and key supporting papers for each claim. This can help you compare different conclusions at a glance and understand which parts of the answer are strongly or weakly supported.
TL;DR Header: The TL;DR Header gives you a quick takeaway at the top of the page before you dive into the full analysis. This is designed to help you quickly understand the main answer, then review the supporting evidence in more detail.
Results Timeline: The Results Timeline shows when the top papers on a topic were published. This can help you see how research has evolved over time, identify influential studies, and spot trends in the literature.
Research Gaps Matrix: The Research Gaps Matrix helps map where research exists and where evidence may be missing. You can use it to compare dimensions like population, study method, outcome, or another custom category, then open individual cells to explore the papers and evidence behind each area.
Mermaid diagrams: Consensus may generate Mermaid diagrams when a visual flow, relationship, or process map would help explain the answer. These diagrams can help show how concepts connect, how a process works, or how different ideas relate to one another.
Key papers: Consensus may highlight key papers that are especially relevant, influential, or useful for understanding the topic. This can help you identify which papers may be worth reading more closely.
Top contributors: Consensus may also surface top contributors in a research area, helping you identify researchers or authors who frequently publish on the topic.
Research Gaps Matrix
The Research Gaps Matrix helps you identify where research exists and where evidence may be missing across a topic or Collection.
You can open a cell to view the papers behind it, review the supporting evidence, and see why those papers were placed in that section.
If a cell has little or no evidence, you can use it as a starting point for a more targeted follow-up search. This can help you investigate a specific gap, such as an underrepresented population, method, outcome, or study design.
You can use the Research Gaps Matrix across a general topic or within a saved Collection to understand what your current evidence base covers and what may be missing.
FAQs
Why didn’t a specific element or visualization appear in my analysis?
Not every AI analysis will include every feature.
The elements shown depend on factors like:
The question you ask
Whether you are using Pro or Deep Review
The number and type of papers retrieved
Whether full text is available
Whether the question is suited for a specific visualization
Whether figures or tables are available from open access papers
Can I ask Consensus to include a specific visual?
Yes. You can ask for specific outputs in your question, such as:
“Show figures from the papers.”
“Create a Research Gaps Matrix.”
“Show a timeline of the research.”
“Create a Claims & Evidence Table.”
“Use a Mermaid diagram to explain this process.”
What does the checkmark mean in the in-line citation?
In-line citation checkmark = full text was used for the analysis.
No checkmark = abstract and available metadata were used instead.