Research pass included GraphRAG-style retrieval mismatch: correct knowledge exists, but retrieval returns wrong chunks or unverifiable summaries.
Tool profile · provisional profile
Microsoft GraphRAG
Useful corpus structuring for large bodies of text. Best viewed as a curation/indexing architecture, not an agent memory product by itself.
Provisional fit
74/100
Best for: Teams building corpus-level knowledge maps where relationship extraction and global summaries matter.
Avoid if: you need a fully governed, citation-complete knowledge architecture without adding policy, evidence capture, and review workflow around the tool.
Caution: Graph rebuild cost, update/delete lifecycle, source authority and retrieval mismatch remain open architecture questions.
Model signature: Curation primary · corpus scope · System
Layer coverage
Where this tool fits.
This is not a completed review. It is a provisional profile from public positioning plus known failure-mode mapping. Hands-on benchmarks, source snapshots, and citation-bound claims are still required before stronger conclusions.
Production
Curation
Storage
Context Assembly
Governance
Review rule: a tool does not get credit for a layer unless it exposes inspectable behavior, not just a marketing claim.
Evidence notes
What the provisional profile has applied so far.
Mapped to curation because it transforms corpus material into graph/community summaries rather than managing live agent memory.
Needs evaluation on incremental updates, provenance, stale summaries and operational cost.
Review packet
What a complete review must contain.
This page exposes the intended review structure. The current artifact is a profile, not a completed evidence-backed review.
Canonical source
Strengths
Teams building corpus-level knowledge maps where relationship extraction and global summaries matter.
Limitations
Graph rebuild cost, update/delete lifecycle, source authority and retrieval mismatch remain open architecture questions.
Dimension assessment
Scope, volatility, authority, lifecycle, resource economics, interoperability, and evidence quality must each get a rationale and citations before final scoring.
Open questions
- What can be verified from docs, code, issues, benchmarks, and changelogs?
- Where does the tool fail under stale, contradictory, private, or high-cost knowledge?
- Which claims are vendor claims versus independently observed behavior?
Benchmark critique
No benchmark number is accepted as architectural evidence unless it says which layer it tests and what it misses: lifecycle, scope boundaries, authority, context cost, and governance.
Related systems
Related tools should be connected by evidence-backed edges: competes with, integrates with, implements concept, evaluated by, or has governance gap.
Update history
Provisional profile created. Stale-review detection, source snapshots, and changelog watching are required before this becomes a durable review.