The 8-Step Workflow

The complete systematic review pipeline — from research question to published manuscript — in a single integrated platform.

The 8-Step Workflow

Other platforms give you pieces. mapped gives you the whole pipeline.

Systematic reviews follow a well-defined methodology: PRISMA, Cochrane, and GRADE lay out the standards. The problem has never been what to do — it's that doing it requires five disconnected tools, endless reformatting, and months of manual work. mapped puts all eight steps in one workspace, with AI acceleration and methodology enforcement built in.


Step 1: Research Criteria

Define what you're looking for — and whether it's worth looking.

Every systematic review begins with a clear research question. mapped's PICOS framework builder helps you define your Population, Intervention, Comparison, Outcome, and Study design with AI assistance. But mapped goes further: the Research Intelligence module scores your idea across 8 dimensions with 120 data points, cross-references PROSPERO for existing registrations, and gives you a mapped VALIDATION score before you invest a single month.

Most abandoned reviews fail because the question was flawed from the start. mapped catches that early.

Deep dive: Research Intelligence →


Search ten databases at once. Get back deduplicated, PRISMA-ready results.

Instead of running separate searches in PubMed, Cochrane, Scopus, Embase, and six more — then spending days merging and deduplicating — mapped searches all of them simultaneously from a single query. AI generates optimized search strings for each database's syntax, maps MeSH terms, expands synonyms, and automatically deduplicates results. Your search strategy is documented for PRISMA compliance without manual logging.

What used to take two weeks of librarian time now takes an afternoon.

Deep dive: Literature Search →


Step 3: Screening

Dual-reviewer screening with AI acceleration — rigor at speed.

Screening is where most reviews stall. Thousands of records, two reviewers, weeks of work. mapped's AI Classification Engine pre-processes your screening queue in batches, flagging likely inclusions and exclusions. Human reviewers then confirm or override. Dual-reviewer workflows with automatic conflict detection ensure PRISMA-compliant independence, and real-time progress tracking keeps the team on schedule.

The AI doesn't replace your judgment — it eliminates the obvious so you can focus on the borderline cases that actually matter.

Deep dive: Screening →


Step 4: Data Extraction

PDFs in. Structured spreadsheets out. No manual copying.

Extracting data from research papers is tedious, error-prone, and soul-crushing. mapped's Multimodal Extraction Engine reads PDFs the way a human would — seeing tables, figures, headers, and footnotes — and transforms them into structured, editable spreadsheets. Native Google Sheets integration means your team can collaboratively edit extracted data in real time. Every AI-extracted value goes through human validation before it enters your dataset.

No more squinting at PDFs with a spreadsheet open in a split screen.

Deep dive: Data Extraction →


Step 5: Risk of Bias

Four validated tools. One consistent, AI-assisted workflow.

Study quality assessment is one of the most subjective steps in a review — and one of the most scrutinized by peer reviewers. mapped supports all four major risk of bias scales: RoB 2.0 (randomized trials), ROBINS-I (non-randomized studies), Newcastle-Ottawa Scale (observational), and QUADAS-2 (diagnostic accuracy). The Advanced Reasoning Engine assists with consistent domain-level judgments, and you confirm every assessment before it's finalized.

Traffic light visualizations and summary tables are generated automatically for your manuscript.

Deep dive: Quality Assessment →


Step 6: Meta-Analysis

Publication-ready forest plots without writing a line of R code.

mapped's R-based statistical engine handles the analysis while you focus on interpretation. Run fixed-effects, random-effects, or mixed-effects models. Generate forest plots, funnel plots, and run publication bias tests (Egger's regression, Begg's rank correlation, trim-and-fill). Perform subgroup analysis, meta-regression, and leave-one-out influence analysis. All outputs are publication-ready and exportable to Word, PDF, or LaTeX.

The statistics are rigorous. The interface is not a command line.

Deep dive: Meta-Analysis →


Step 7: GRADE Assessment

Rate your evidence from Very Low to High — systematically.

The GRADE approach is the international standard for rating evidence quality, and it's notoriously manual. mapped walks you through each downgrade factor (risk of bias, inconsistency, indirectness, imprecision, publication bias) and upgrade factor (large effect, dose-response, residual confounding) for every outcome. The system generates Summary of Findings tables that go straight into your manuscript.

No more GRADE worksheets in separate files. It's integrated with your extracted data and risk of bias assessments.

Deep dive: Quality Assessment →


Step 8: Manuscript

From extracted data to publication-ready draft — with AI assistance.

The final step: turning your analysis into a manuscript. mapped generates structured drafts that follow PRISMA reporting guidelines, populates results sections from your actual data, and integrates figures, tables, and the PRISMA flow diagram. Collaborate on drafts in Google Docs, use pre-configured journal templates, and export to Word, PDF, or LaTeX.

You still write the discussion. The AI handles the sections where accuracy to data matters more than interpretation.


The Full Pipeline

StepWhat happensKey output
1. Research CriteriaPICOS definition, mapped VALIDATIONFeasibility report
2. Literature SearchMulti-database search, deduplicationPRISMA-compliant search results
3. ScreeningDual-reviewer with AI accelerationIncluded studies list
4. Data ExtractionPDF to structured dataExtraction spreadsheet
5. Risk of BiasRoB 2.0, ROBINS-I, NOS, QUADAS-2Bias assessment tables
6. Meta-AnalysisStatistical analysis with R engineForest plots, effect estimates
7. GRADE AssessmentEvidence quality ratingSummary of Findings table
8. ManuscriptAI-assisted writingPublication-ready draft

Every step feeds into the next. Data flows forward automatically — no exporting from one tool and importing into another.


Ready to start? Create your free account → or learn about the AI architecture that powers each step.