Lead scoring for a small business on WhatsApp: what hot, warm and cold actually mean, the signals an AI agent reads from a conversation, how to use the score without overthinking it, and why a three-level score beats a hundred-point one
Lead scoring sounds like something for enterprise sales teams with dashboards. For a small business it is simpler and more useful: of the thirty people who messaged this week, which five should I call today? A three-level score answers that, and an AI agent can set it from the conversation.
Quick answer
Lead scoring is labelling each enquiry by how likely and how soon it is to buy, so you spend your time on the right ones. Three levels are enough: hot (clear intent, soon, budget fits, specific), warm (interested but later, unsure, or missing a detail), cold (browsing, out of scope, or no reply). An AI agent reads those signals from the WhatsApp conversation (what they asked, when they need it, how they responded to price, how specific they are) and records the level on the lead. Act on hot today, warm within the week, cold on a monthly review. ReplyKit tags leads this way automatically.
The three levels
| Level | Signals in the conversation | Action |
|---|---|---|
| Hot | Asks about a specific service or product, gives a date or "this week", does not flinch at the price, asks how to book or pay | A person follows up today; book or quote |
| Warm | Interested but "next month" or "just comparing", asks many questions, hesitates at price, needs a partner's say-so | Follow-up within the week with a reason; nurture |
| Cold | Vague interest, out of service area or scope, no reply after the first answer, price far outside range | Monthly review; no active chasing |
The questions that surface these signals are in lead qualification questions.
How the agent reads it
As the conversation runs, the agent notes intent (what exactly they want), timing (when), fit (service area, budget, scope), engagement (reply speed and depth) and specificity (a named service beats "info please"). It records the level on the lead and updates it as the chat develops; a cold lead who comes back with a date becomes hot. Bookings move leads automatically. The lead record also carries the answers, so the person following up starts informed; the CRM side is in what a small business needs from a CRM.
Using it without overthinking
- Each morning: filter hot. Call or message every one. That is the whole sales routine for most small businesses.
- Mid-week: filter warm with a follow-up date due. One nudge each with a reason; the sequence is in following up silent leads.
- Monthly: skim cold. Close the dead ones with a reason; a few will have warmed.
- Never let the score replace reading the chat. It is a sort order, not a verdict.
Why point systems fail small teams
- Nobody maintains the weights. A 100-point score built in a spreadsheet is abandoned by month two.
- False precision. A 63 and a 67 are the same lead; three levels are honest about that.
- Gaming. Staff learn which fields raise the score.
- What works: three levels, set from conversation signals, reviewed by a human who reads the hot ones. The measurement that does matter is in measuring lead response time: a hot lead answered in a minute converts; the same lead answered tomorrow is warm at best.
Frequently asked questions
What is lead scoring?
Labelling each enquiry by how likely and how soon it is to buy, so you follow up the right ones first.
What makes a lead hot?
Specific intent, a near-term date, budget fit, and asking how to book or pay.
Can an AI agent score leads?
Yes. It reads intent, timing, fit, engagement and specificity from the conversation and records a level on the lead.
How many levels do I need?
Three. Point systems are rarely maintained by small teams and imply precision that does not exist.
What do I do with cold leads?
A monthly skim. Close the dead ones; a few will have warmed.
Hot leads first, every morning
The agent reads the signals and tags the lead; you call the hot ones. 7-day free trial, no card.
Sources and further reading
- Harvard Business Review, The Short Life of Online Sales Leads (Oldroyd, McElheran, Elkington, 2011) · firms that responded within an hour were about seven times more likely to qualify the lead
- Lead Response Management Study (2007) · the five-minute window for contacting a new lead
ReplyKit is an independent product and is not affiliated with or endorsed by Meta or WhatsApp. This article is general guidance for running a small business, not legal, financial or regulatory advice.