Improving a Startup Discovery Workflow for Accuracy and Completeness
Hello Team, I'm reaching out about a workflow I’ve been building that is designed to generate a list of startups similar to a target company entered into a Notion database. Workflow Overview: The workflow is supposed to produce a structured list of 50-100 real, active, and accurate startups that match a target startup based on industry, customer segment, technology stack, business model, stage (e.g., Pre-Seed, Seed, Series A), and geographic proximity. It is important that:
All startups listed are legitimate and active.
The list reaches the full target count (100 startups) without stopping early.
Each entry includes fields like name, description, location, website, and a short similarity assessment with a similarity score.
Problems I'm Experiencing:
- 1.
Incomplete Lists:
- 2.
The workflow often stops well before reaching the full target of 100 startups, sometimes producing fewer than 50.
- 3.
Inaccurate / Non-Real Startups:
- 4.
Some startups being listed appear to be fictional, inactive, or otherwise not legitimate, which goes against the goal of compiling only real, verifiable companies.
Request: Could you help investigate why this is happening and suggest improvements or adjustments? I'd like to ensure that the workflow reliably produces full, high-quality lists without any fictional entries. Thanks so much for your help!
