Day 2: LeadRadar AI now qualifies leads automatically
Build Update — Day 2
Yesterday LeadRadar AI could collect Reddit posts.
Today it can decide which ones are actually worth paying attention to.
I added an AI scoring step that evaluates:
• purchase intent
• urgency
• budget signals
The goal isn't to replace human judgment.
It's to avoid spending time reading dozens of posts that have no commercial value.
Still validating the scoring prompt, but the workflow is starting to feel like an actual product instead of just an automation.
Still validating the scoring prompt is smart to flag early. The tricky part with purchase intent scoring is that signals look different depending on the industry and platform. A post asking about CRM migration has different urgency markers than someone comparing AI dev tools. Are you planning separate scoring logic per source or keeping it to one generalized prompt with adjustable thresholds?
That's exactly one of the questions I've been thinking about.
For the MVP, I'm keeping a single scoring prompt with a strict decision framework and configurable thresholds. I wanted to validate the overall pipeline before introducing more complexity.
Longer term, I don't think one prompt will fit every source. Reddit, LinkedIn, email inquiries, and job boards all express buying intent differently, and even different subreddits have different patterns.
My plan is to keep a shared scoring framework while allowing source-specific rules and examples where they actually improve precision, rather than maintaining completely separate systems.
I'm treating the current version as the baseline, then I'll specialize it once I have enough real data to justify the differences.