/warm-lead-routing

name: "Warm Lead Routing"
description: >
  Lead-warmth scoring from intent and engagement signals — Warmly
  website-visitor identification, CRM form fills and email engagement, and
  Calendly booking activity — using an explainable Hot/Warm/Warm-Cool/Cool
  tiering, plus routing recommendations based on owner continuity, CRM
  routing rules, or rep capacity, and the degradation path to CRM-only
  signals when intent tools aren't connected.
when_to_use: >-
  When identifying which leads are showing real buying intent right now and
  who should follow up on them. Use when: warm leads, hot leads, who should
  follow up on this lead, lead routing, lead scoring, which leads are
  engaged, intent signals, website visitor identification.

Warm Lead Routing

Overview

Most inbound lead volume is noise — a form fill from a student researching a term paper looks identical in a CRM to a form fill from an IT director who just had an outage. Warmth is what separates the two, and it comes from signals CRM stage alone doesn't capture: is this account revisiting the pricing page, did the same person book a call, did an email actually get opened. This skill combines those signals into a practical warmth score and turns that into a routing recommendation — the right rep, contacted soon enough to matter, with the reasoning shown so a sales manager can override it.

This skill is explicitly tiered: the richer the connected tool set, the more precise the scoring, but a CRM alone (no Warmly, no Calendly) still produces a usable, if coarser, read. Never treat missing intent tools as a reason to skip lead scoring entirely — degrade the signal set, not the output.

Anti-triggers

• **Warmly's own visitor and account surface** — choosing between the

visitor and account lists, ICP filtering, and credit-burn checks belong to `warmly-visitor-intelligence`. This skill blends that signal with CRM activity into one warmth tier and a routing proposal.

• **Contact records and email engagement as data** — use `hubspot-contacts`

or `hubspot-activities`.

Signal sources and what they contribute

| Signal source | Signal | Weight (relative) | |---|---|---| | Warmly (if connected) | Anonymous or identified website-visitor activity — repeat visits, pricing/product-page views, session recency | High — this is the earliest, least-solicited intent signal available | | Calendly (if connected) | A booked call/meeting — the strongest possible intent signal, since the lead took a deliberate action | Highest — a booking should almost always push a lead into "warm" regardless of other signals | | HubSpot (or connected CRM) form fills | A submitted form — contact info plus whatever context the form captured (use case, company size, urgency language) | Medium — real intent, but lower-effort than a booking | | HubSpot (or connected CRM) email engagement | Opens, clicks, and replies on marketing or sales sequence emails | Medium-low individually, but a rising trend (multiple opens/clicks over a short window) is a meaningful signal | | HubSpot (or connected CRM) lifecycle/lead status | Existing lead score or lifecycle stage, if the CRM tracks one | Contextual — use as a prior, not a replacement for the signals above |

Discovering available tools first

Before scoring anything, call `conduit__search_tools` to determine which of Warmly, Calendly, and a CRM (typically HubSpot) are actually connected for this org. Do not assume all three — this is the skill in this pack most likely to run with partial coverage, since Warmly and Calendly are both explicitly optional enrichments per this pack's README. Scope scoring to whatever is actually live, and state which signal sources were used (and which weren't available) in every output.

Scoring approach

Keep this explainable rather than a black-box formula — a rep or sales manager needs to see why a lead is scored warm to trust routing it.

1. **Start from a base tier using the strongest available signal:**

• A Calendly booking in the last 7 days → start at **Hot**.
• Warmly-identified repeat visits (2+ sessions in 7 days) or a pricing/

product page view → start at **Warm**.

• A CRM form fill in the last 14 days with no other signal → start at

**Warm-Cool** (real intent, but lower urgency than a booking or repeat visit).

• Email engagement only (opens/clicks, no form fill, no booking) → start

at **Cool**.

• No recent signal of any kind → **Cold** — do not include in a warm-lead

routing output; this skill surfaces leads worth acting on now, not the full lead database.

2. **Adjust up or down based on secondary signals:**

• Multiple signal sources agreeing (e.g. a Warmly repeat visit *and* a

form fill within the same week) moves a lead up one tier — convergent signals are more reliable than any single one.

• A signal that's aging (e.g. a form fill 12 days old with nothing since)

should not be scored as fresh as one from yesterday — recency matters within a tier, not just which tier it lands in.

• Company-fit context, if available (existing client's employee vs. a

brand-new domain, or a company size wildly outside the org's typical client profile) can be noted as a qualifier, but should not override a strong behavioral signal — a bad-fit account that books a call is still worth a fast, brief qualifying response.

3. **Only "Hot" and "Warm" tiers get a routing recommendation** in normal output; "Warm-Cool" and "Cool" can be included in a fuller sweep if asked, but shouldn't clutter a default "who's warm right now" answer.

Routing recommendation

Once a lead clears the Warm threshold, propose an assignee:

1. If the lead is already associated with an owner in the CRM (e.g. an existing named account, or a rep already assigned to the deal/contact), route to that owner by default — continuity beats round-robin. 2. If unowned, and the CRM exposes territory, industry