External intent signals

Your next customers announce themselves. Your system never hears it.

A Series B, three RevOps reqs, a new VP of Ops, a competitor's tag appearing in their stack — all public, all timestamped, all missed. That isn't a demand problem. It's a listening problem.

Map my signal sources Build a score

Diagram: external signals detected, matched to a CRM account, scored on weight and recency, then routed with an AI draft — How a public event becomes a matched account, a score, and a first touch inside 15 minutes.

01 · Detect

Funding news, job posts, exec changes, tech installs, social mentions and list appearances, pulled on a schedule from Clay and the sources under it.

02 · Match

Every event resolves to a company and a person already in your CRM — an account record, not an alert in someone’s inbox.

03 · Score

One weighted, time-decayed number per account: a Series B three weeks old outranks a blog mention from March.

04 · Play

Crossing a threshold assigns an owner, drafts outreach that cites the event, and starts the SLA clock.

01 · The foundation

The external signal object

Third-party intent dies in notification feeds: a funding alert in email, a job post someone happened to see on LinkedIn, a Clay export going stale in a spreadsheet. We model every external event as a first-class custom object in your CRM — the same way you model deals, contacts and companies.

Why it matters

Because the event is structured data on the account instead of an alert someone saw once, it can be scored, decayed, reported on, automated against and handed to AI with its evidence link attached.

Signal typeFunding round, exec hire, job change, promotion, open req, department hiring surge, tech install, social mention, industry-list appearance.
SourceClay waterfalls, job boards, news and filings, LinkedIn / X / Reddit / YouTube listening, install detection, curated industry pages.
Detected date & recencyWhen the event became public, so a 9-day-old raise outweighs a 6-month-old one and stale signals decay on their own.
Signal weightHow much this event predicts a deal — a Series B plus an ops req outweighs a founder’s viral post.
Match & evidenceResolved to the company and the buying-committee contact, with the source URL stored so a rep can read it before they call.

External Signal 1 record per event

Signal typeSeries B announced

SourceClay → news + filings

Detected2026-07-12

Signal weight+25

Matched companyNorthwind Systems

Matched contactDana Whitfield · VP RevOps

Recency19 days (×1.0)

EvidenceSource URL on record

Durable, queryable, matched to the account. Nothing sits in an inbox.

External signals detected on this account, last 8 weeks

8 wks agoThis week

The signal library

Signals we can wire, by category

Every row below is a signal we have pulled into a CRM signal object from Clay, HubSpot Breeze, or the provider underneath them. Filter it, then argue with the weights — that argument is the first workshop.

Filter by category — changes the list below

Signal	Where it comes from	Category	Starting weight

Starting weights come from prior builds, not from your data. Nobody wires 47 signals on day one — the first build usually takes six to nine, the ones that showed up before your last 20 closed-won deals.

02 · The prioritization layer

The account signal score

A feed of public events is still noise until someone answers one question: which accounts are worth a call this week? Toggle the signals below and watch the score, the band and the routing rule change.

Select detected signals on the account — changes the raw score

Detected when — changes the time-decay multiplier

The arithmetic

Bands: Cold 0–34 · Warm 35–69 · Hot 70+. Weights come from which events preceded your own closed-won deals — not a vendor’s default model.

0Cold

What the system does at this score

03570100

What the score changes

Without a signal score

Reps prospect alphabetically and by gut feel.
A funding round gets found six weeks late — by a competitor first.
Alerts pile up in inboxes and nobody owns the follow-up.
Clay exports go stale in a spreadsheet.

With the signal score

The list is ordered by events that preceded closed-won.
A raise or a new VP triggers outreach the same day.
One owner, one SLA, one number on the account.
Every signal keeps its evidence link for the rep.

03 · The conversion layer

AI outreach that cites the event

Knowing an account just raised is half the job; saying something they can’t dismiss is the other half. The AI reads the signal record — event, source, date — and writes the opener around it.

Detected signal — changes the drafted outreach

Draft · pulled live from the signal object Warm

Subject:

Reps approve and send instead of writing from scratch, so event-to-touch holds even at volume.

References the real event

"Saw the Series B and the two RevOps reqs" — not "hope you're well."

Matches the play to the signal

A funding round opens differently than a competitor’s tag appearing in their stack.

Carries the evidence

The source URL rides along, so the rep reads the announcement before they dial.

Fires the day it’s public

Drafted the moment the signal lands, queued to the owner with the record attached.

The queue

What a rep opens on Monday morning

One list, sorted by score. Nobody has to be watching LinkedIn for the system to work.

In-market accounts · this week

Accounts monitored 1,240 Hot crossings 24 Median event-to-touch 11 min

Account	Top external signal (last 30d)	Intent score	Next action

Illustrative data. In production this is a CRM view built on the signal object — no separate alert tool to log into.

Where you actually stand

Signal coverage, banded

Bands from JayOh engagements across B2B SaaS between $5M and $50M ARR. Find your row, then find your column.

Metric	Weak	Acceptable	Strong
Target accounts under monitoring	<20%	20–60%	>90%
Signals matched to a CRM record	<25%	25–70%	>80%
Event going public → first touch	>30 days	3–30 days	<48 hrs
Outbound citing a specific event	<5%	5–30%	>70%
Weights re-fit against closed-won	Never	Ad hoc	Every quarter

Two Weak rows is a listening and matching problem, not a demand problem. Three or more and the score would be untrustworthy even if you shipped it tomorrow.

The math

What the events you're missing are worth

Put in your own numbers. Every line of the arithmetic stays visible so you can check it against your CRM.

Accounts with a qualifying signal / month

Worked within 48 hours today

Average ACV

Close rate (%)

Conversion lift assumption — changes the opportunity math

Lift = the difference in opportunity-creation rate between an account contacted inside 48 hours of a public event and the same account contacted a month later. Pick the band you believe.

Arithmetic

Annual revenue left on the table

This is your money, not our fee. These events are already public — the only variable is whether anything catches them.

The result

One monitored list, not six notification feeds

Detect

Every public event caught on a schedule, not by luck.

Match

Resolved to the account and the person who matters.

Score

One number your team can work today.

Play

Outcomes feed back, so the weights keep sharpening.

See which accounts are signaling right now.

We'll inventory the external sources worth watching for your ICP, score them against your closed-won history, and show you the outreach that cites them. Nothing to install before the call.

Book the signal audit