Show HN: Watchman – 4KB AI script to identify and qualify anonymous B2B visitors
We’re releasing Watchman – an AI agent that helps you capture hidden B2B buyers by automating the identification, research, and qualification of anonymous B2B visitors with just one line of code.
The average public SaaS company now waits nearly five years to earn back its customer-acquisition costs, and the same is true for most B2B companies. More than 70% of B2B buyer intelligence hides in the shadows – anonymous research behaviors and untracked engagements such as stealth web traffic. Watchman brings deep-science precision to capture dark-funnel buyers and provide predictive demand intelligence.
Several members of our founding team hold PhDs in fields related to statistical inference or machine learning. We came together because we believe modern GTM tools excel at collecting signals but fall short on the inference stack. Simply plugging an LLM into a dashboard won’t cut it — the real breakthrough requires a far sharper lens and much stronger predictive capabilities. That’s exactly what we provide.
We’ve improved and rethought every detail of legacy platforms – not only by embedding AI-native architecture and data science at the core, but also by refining execution down to the smallest elements. For example, our ultralight 4KB AI tracking script is more than 90% smaller than a standard Google Analytics tag, so pages load faster, bounce rates drop, and prospects stay focused on what matters. We’re committed to delivering what matters most: a smaller footprint, sharper intelligence, and faster action.
Watchman transforms anonymous B2B web traffic into pipelines of qualified accounts and person-level leads. Our agent identifies, researches, and qualifies hidden B2B buyers, syncs the enriched leads into tools such as your CRM, sends live notifications to Slack and email, and supports unlimited users so you can invite your entire team.
We’d love for you to try out Watchman and share your feedback! Check it out at https://joinwatchman.com