Inquiry intelligence + CRM

Inquiries pass a zero-cost heuristic pipeline first: corporate-email detection, country inference, product-keyword hits, then A/B/C grading — obvious spam short-circuits without spending a cent of LLM. Spam also clusters on three axes (same IP, same email variants, same message template) for one-click batch cleanup.

The board carries a full activity timeline — notes, calls, mail in and out, tasks, stage changes, eight event types. Assignment rules match country, product and source to the right rep with round-robin pools; the leaderboard computes reply rate, win rate and average first-response hours; the boss overview compresses week, month and year into one screen.

The question every sales manager asks — did the rep actually follow up? — gets a quantified answer on the team board. Follow-up rate measures real contact coverage across open leads. Hard-touch rate counts only objective evidence — emails the system actually sent, or a customer reply — so self-reported notes and calls can't inflate it. And any open lead untouched for over seven days with no pending task surfaces in a no-next-step alert, sorted by score, one click from the lead itself.

The rep's day runs on this system too: install it as a phone app and the customers due for follow-up today, or inquiries unanswered past 24 hours, push straight to the lock screen. Reps self-serve AI-drafted replies within a daily quota, and saved replies swap in per-customer variables. Unassigned public-pool leads move to a claim-and-approve model: reps request, ops decides, and the system assigns atomically so no two reps collide. Replies to cold outreach flow into the inquiry pool automatically — a real human reply lands on the timeline with attribution and a rep alert, while auto-replies and out-of-office only pause the sequence, and mail carrying quote or sample buying signals is never mistaken for an unsubscribe.