Felicis leads $7.5M bet on Unsiloed AI to make enterprise documents usable by AI agents
- ID
- 47778
- Status
- summarized
- Published
- 18 Nov 2026, 8:00 PM
- Fetched
- 18 Nov 2026, 8:41 PM
- Provider
- TechCrunch
- Category
- technology
- Original URL
- https://techcrunch.com/2026/11/18/felicis-leads-7-5m-bet-on-unsiloed-ai-to-make-enterprise-documents-usable-by-ai-agents/
- Source URL
- https://techcrunch.com/feed/
Summary
- Score
- 4.5
- Created
- 18 Nov 2026, 8:41 PM
- Tags
- Audience
- developersai_ml_learnersai_agent_userssaas_founders
What happened
Unsiloed AI raised a $7.5M seed round led by Felicis, with Y Combinator, Orange Collective, Uncorrelated Ventures, and angels Aarthi Ramamurthy and Beth Turner participating. The company extracts tables, charts, and complex layouts from enterprise documents into structured data for AI agents, and shows where extracted information came from in the source document. It says it has customers in finance, legal services, and healthcare, including a law firm running it on its own servers for contracts and old case files; the announcement gives no pricing, accuracy benchmarks, or latency numbers.
Why it matters
If you are building agents over contracts, bank statements, claims, or case files, the concrete bottleneck named here is table/chart extraction plus traceability back to the source page. Unsiloed's pitch is source-linked structured extraction and on-prem deployment for sensitive documents, but without accuracy, latency, or per-page pricing you should treat it as a vendor to benchmark on your own PDFs rather than a default choice. For SaaS founders, this is another funded document-ingestion layer, so decide whether to buy that layer or keep your own parsing pipeline as differentiation.
Discussion angle
What evidence would you require before trusting a document-extraction vendor for agent workflows: citation-level accuracy on your own PDFs, on-prem deployment, or price per page?