Medicine

AutoSynthesis: An agentic system for automated meta-analysis

Moein Taherinezhad, Sebastian Maier, Gerardo Vitagliano, Francesco Pierri, Stefan Feuerriegel

Featured July 19, 2026

This analysis was generated by SciGrove. Upload your own PDFs or enter a DOI — and get the same AI breakdown on any paper.

Get started

AI-generated analysis — This is SciGrove's AI interpretation of the paper, not peer-reviewed content. Always refer to the original paper.

Simply

A new AI system called AUTOSYNTHESIS uses smart computer agents to automatically find, read, and combine scientific studies, quickly creating reliable summaries that used to take humans months.

In depth

The paper introduces AUTOSYNTHESIS, an end-to-end multi-agent system that automates the entire meta-analysis workflow. This system leverages large language models (LLMs) to perform tasks traditionally requiring extensive manual effort, such as literature retrieval, study screening, quantitative data extraction, and statistical synthesis, thereby making evidence synthesis more scalable and timely.

Key Takeaways

1
End-to-end automation: AUTOSYNTHESIS automates the full meta-analysis pipeline, from research question to final report, significantly reducing manual effort.
2
Multi-agent LLM framework: The system employs specialized LLM agents for reasoning-intensive tasks, combined with deterministic statistical modules for methodological rigor.
3
Scalable and reproducible evidence synthesis: It enables faster, more transparent, and auditable meta-analyses, supporting "living" reviews and evidence-based decision-making.

Conceptual Flow

HIGH LEVEL

Methodology (The "Logic")

The system uses many smart computer helpers, each doing a specific job like finding papers or checking numbers, to put together a big report.

Research Question

Many Smart Helpers Work Together

Full Report

Results (The "Impact")

The system's automatic summaries were very similar to those made by human experts, showing it can do a good job.

Human Expert Summary

AI System Summary

Compare Results

Very Similar Findings