Skip to content
AskVault Docs
Try Free →

Buyer Radar

Buyer Radar was called Sales Pulse until October 2026. Only the name changed.

Buyer Radar turns your agent's conversations into a revenue workflow. Every night it reads the conversations your agent handled and answers two questions a support tool never asks: who was ready to buy, and what stopped the ones who didn't.

What it shows

  • Hot leads: conversations with high buying intent where a contact was captured (through lead capture or visitor identification). Each row carries a one-line reason: for example "Asked for bulk pricing on the Growth plan for 40 seats": so follow-up starts with context, not a transcript hunt. Click any row to open the full conversation.
  • High-intent conversations: everything the classifier scored medium or high, ranked by intent score. Anonymous high-intent visitors are listed too, with a nudge to turn on lead capture so the next one isn't anonymous.
  • Blocked sales moments: purchase-relevant questions your agent could not answer (it said it didn't know, or the answer wasn't grounded in your content), grouped by theme, pricing, shipping, availability, discounts, with real examples. The Fix gaps in AI Insights button takes you to the improvement loop where a drafted answer can be applied to your knowledge base.

What the classifier looks for

Each conversation is labeled with a small, fixed set of sales signals: pricing questions, explicit purchase intent, demo requests, availability and shipping questions, discount requests, competitor comparisons, bulk/enterprise interest, urgency, upgrade interest, and cancellation risk. Signals outside this list are never invented. Anything unrecognized is dropped.

Detection of "blocked" moments is grounded in measurement, not guesswork: answers your agent's own confidence checks marked unanswered or unsupported are flagged to the classifier, so a blocker means a real unanswered buying question.

How it runs

  • Analysis runs nightly on a fast, low-cost analysis model: it never uses your chat model and never draws down your message allowance.
  • Each night it reads up to 25 finished conversations per workspace on Free, 100 on Starter, 500 on Growth and 2,000 on Business and Enterprise, from the last 48 hours.
  • Conversation text is treated as untrusted input end to end: the classifier can only emit the fixed signal set above, and everything it writes is stored and rendered as plain text.

Privacy

Buyer Radar reads only your own workspace's conversations, and its results are visible only to your workspace's team. Contact details shown on lead rows come from your existing lead capture and visitor identification: Buyer Radar never collects new personal data on its own.

  • Leads, where captured contacts are stored.
  • AI Insights, for the unanswered questions behind each blocker.
  • Collect lead skill, which captures the contacts Buyer Radar ranks.
Was this page helpful?