> ## Documentation Index > Fetch the complete documentation index at: https://docs-evisyn.grabmywork.com/llms.txt > Use this file to discover all available pages before exploring further. # Quality Assessment > AI-assisted RoB2 and GRADE assessments with evidence citations Quality assessment evaluates how trustworthy each included study is. EviSyn supports the two most widely used assessment tools — **RoB2** (Cochrane Risk of Bias) and **GRADE** — with AI-generated assessments backed by quoted evidence. ## Setting Up Assessment Navigate to **Quality Assessment** from your project dashboard. Quality assessment setup ### Stats Bar | Metric | Meaning | | -------------------- | --------------------------------------- | | **Finalized Papers** | Total papers eligible for assessment | | **Assessed** | Papers with completed assessments | | **Pending** | Papers awaiting assessment | | **Confirmed** | Assessments verified by human reviewers | ### Choose Your Assessment Tool Select from the dropdown: | Tool | Best for | Domains | | ------------------------ | ----------------------------- | ------------------------------------------------------------------------- | | **RoB 2 (Risk of Bias)** | Randomized controlled trials | Randomization, deviations, missing data, measurement, selective reporting | | **GRADE** | Overall certainty of evidence | Risk of bias, inconsistency, indirectness, imprecision, publication bias | ### Managing Questions Click **Manage Questions** to view and customize the assessment domains and questions. Each domain contains specific questions. For example, **RoB2** includes: **Bias arising from the randomization process:** * 1.1 Was the allocation sequence random? * 1.2 Was the allocation sequence concealed? * 1.3 Did baseline differences suggest a problem? **Bias due to deviations from intended interventions:** * 2.1 Were participants aware of their assignment? * 2.2 Were carers aware of participants' assignment? * 2.3 Were there deviations from intended intervention? * 2.4 Did deviations likely affect the outcome? You can: * **Add questions** to any domain * **Remove** irrelevant questions * **Add domains** for custom assessment criteria * Choose rating scales (Yes/No, Low/Some concerns/High) Click **Save Questions** to save your configuration. ## Running AI Assessment Click **Run AI Assessment** to start. For each paper, the AI: 1. Reads the full text 2. Evaluates each question in each domain 3. Provides a **rating** (Low risk / Some concerns / High risk) 4. **Quotes the specific text** from the paper that supports the rating 5. Includes **page and section references** 6. Explains its **reasoning** in plain language ### Human Confirmation Every AI assessment starts as a draft. Reviewers must: * Review each domain rating * Check the quoted evidence against the source * Confirm or override the AI's judgment ## What's Next? After quality assessment, proceed to [Meta-Analysis](/user-guide/meta-analysis) for quantitative synthesis, or skip to [Manuscript Writer](/user-guide/manuscript-writer).