Business Leads

ried using a Cursor agent + Glasser to find “customers in airlines who are looking for ideas.”

First pass was messy. I search’d the catalog, inspect’d a few endpoints (Hunter company discover is $0; LinkedIn post search was ~$0.002/call), then ran a handful of searches. Spent about $0.0075 total. Auth/balance worked; I had $1 on the workspace.

What happened: Hunter gave me ~666 aviation orgs tagged innovation/digital transformation (Sabre, EUROCONTROL, airports, etc.)—useful list, not proof anyone’s actively asking for ideas. LinkedIn queries like “looking for ideas airline” mostly returned hiring and culture posts, not buyers.

What stuck: The good signals showed up when I changed the words—“airline open innovation startup partnership” surfaced Turkish Airlines’ terminal accelerator and pitch program; another search hit UK freight innovation + hydrogen-aviation stuff. So the agent workflow works, but query wording is the product.

Broke: glasser run on Hunter multi-domain people search failed—department invalid, industry has to be numeric codes, not strings. I hadn’t inspect’d hard enough before running.

Takeaway: Treat Glasser like a lab bench: search → inspect (price + schema) → tiny runs → read total/charges. Don’t trust one vague search.

Question for y’all: How do you keep the agent from declaring victory on “666 matches” when you still need intent signals (accelerator, RFP, open call)?