Building software got easy. Finding the first customer didn't.

Description and use

Building software is now cheap and fast thanks to AI, but finding the first customers remains hard. The author built GTM Shortlist, a tool that generates a list of potential B2B SaaS customers by analyzing a company’s website and searching Google. After iterative improvements—filtering irrelevant pages and avoiding duplicate research—the tool became more reliable.

Problem

Building software is cheap and fast thanks to AI, but finding the first paying customers remains hard, often due to poor product‑market fit.

Outcome

After iterative fixes—searching for companies, filtering non‑company pages, handling invisible characters, ignoring menus, and preventing duplicate research—the tool produced a more reliable shortlist, reducing wasted credits and improving match accuracy.

Workflow

• Provide target company website to GTM Shortlist
• Tool reads website and writes a short description of the product and target buyer
• User confirms or corrects the description
• Tool searches Google for companies matching the description
• Tool visits each candidate company’s website and looks for proof sentences
• Tool presents a shortlist with proof sentences or "unknown" if none found
• User marks each company as yes/no, downloads spreadsheet, and starts outreach
• Tool never sends emails automatically

Evidence

GTM Shortlist, live at eygentic.tech.

Source

The app reads your website and writes a short description of what you sell and who might buy it.

The app never sends anything for you.

The app had a rule that every search had to work before it used any of the results, so it threw away all 8 and stopped with nothing.

I changed the rule: if one search fails, the app now keeps the companies the other searches already found and carries on.

I changed four things: Search for companies, not articles; Skip pages that aren't companies; Warn me when I ask for something a website won't say; Never research the same company twice for the same business.

Real company websites: there were 21. The old version let 16 of them through to research. The new one lets all 21 through.

The best page was ignored. Search often finds the exact page that proves the point, but research used to start from the homepage instead. Now it reads the page search found as well.