The problem: Companies spend millions on manual research teams to answer compliance / risk questions like, "does this business roll up to a state owned entity in China / Russia?". A bankruptcy processor we work with had 30 people just manually researching data points like this. An ex-Shell trader told us Shell has 800 people + $100M/yr spent with Deloitte to conduct manual research to ensure international trade compliance.
Why existing solutions are lacking: The dirty secret is that legacy providers like Dun & Bradstreet, Orbis, and S&P run on armies of manual workers and are built on decades old technology. At query time, you’re typically tapping into a quarterly generated, static, expensive database.
Our approach: We built this on top of our entity resolution engine - a deep research agent that anchors entities to our business graph, powered by direct integrations with government registrars and our web scraping infra. When you need hierarchy research, our AI agents spend on average 10-20 mins researching upward to identify the ultimate parent and build the complete ownership DAG with source citations. We auto-generate Mermaid diagrams so you can immediately render the results in your own app.
Technical backstory: I've spent 10 years building MDM, data enrichment, and entity resolution systems at B2B startups and enterprises, mostly in fintech. I actually built an in-house Clearbit replacement after they overcharged us and almost killed the startup I was at. The core challenge is always the same - taking messy business data and mapping it to web data and the actual legal entity. Most off the shelf solutions are mediocre because they skip the hard part: the gnarly 20-40% of records that force you into manual research.
Current state: Just three of us, bootstrap funded. We're getting requests for full subsidiary mapping too, but focused on upward hierarchy for now. We're still figuring out pricing (market rate seems to be a few dollars a lookup, but we aim to be a fraction of that). Deep research is expensive but we think we can make it work at scale.
Try it:
- Interactive demo: https://savvyiq.ai/products/entity-hierarchy (no signup)
- Free account: https://savvyiq.ai/signup
- API docs: https://savvyiq.ai/api-reference#tag/entity-hierarchy-beta
Would love feedback on the hierarchy API specifically, including the schema, gaps / edge cases, but also curious about other use cases where business identity resolution is a pain point for you.
Happy to run a few entity hierarchy lookups in the comments if anyone wants to see the API in action - I can share the results as Mermaid diagrams that you can paste directly into any Mermaid renderer.
