Social intent monitoring skill

Social Intent Monitoring

Overview

Most outbound programs are volume-gated rather than timing-gated. Reps build lists and send sequences regardless of whether anything has actually happened. Social intent monitoring replaces that model: an AI agent watches public social conversations continuously, and an email or DM only fires when a qualifying event occurs in real time.

The mistake this skill prevents: treating social platforms as broadcast channels rather than signal feeds. Every competitor-mention comment, every "does anyone have a recommendation for X" post, every founder publicly describing a problem you solve — these are live intent signals that expire within 24-72 hours. Catching them and acting fast beats the best cold copy by a wide margin.

This skill builds the full monitoring architecture: what to watch, how to qualify signals, how to route them to enrichment and outreach tools, and how to measure signal quality over time.

When to Use

• "Set up social intent monitoring for our target accounts"
• "Monitor LinkedIn for people talking about problems we solve"
• "Build a Trigify workflow that routes competitor mentions to outreach"
• "Detect social buying signals before prospects fill out a form"
• "Create a signal infrastructure for our outbound motion"
• "How do I use social listening for lead generation?"
• "Set up keyword monitoring on LinkedIn and X for intent"
• "Automate outreach from social signals"

Authoritative Foundations

Max Mitchum — Signal, Context, Action Framework

Max Mitchum, Co-Founder & CEO of Trigify.io, published the Signal, Context, Action model as the operating framework for signal-led outbound. The three layers:

Signal — the public event that suggests an account is worth contacting now. Examples from Mitchum's 2026 playbook: competitor engagement on LinkedIn or X, hiring posts for GTM roles, founder or operator posts describing a problem you solve, funding reactions, pain-point mentions in comments, product launches in adjacent categories, public recommendation requests.

Context — the judgement layer. Who is this person? What company? Are they in ICP? Why does this specific event matter? Is the timing strong? A signal without context is noise; context without a signal is guessing.

Action — the downstream workflow. Enrich the person, find the email, research the company, draft the message anchored to that specific signal, push into the sequencer, alert the rep in Slack with the reasoning.

Source: https://maxmitcham.substack.com/p/48-meetings-from-120-leads-in-two

The core principle: the email is the commodity. AI can write decent emails. The moat is knowing when the email has a reason to exist.

Trigify Workflow Patterns

Trigify's public documentation describes four canonical agentic workflow patterns for sales and lead generation teams:

1. ICP-Filtered Post Trigger: New Post Trigger → Person Enrichment → ICP filter (country, headcount) → Email Enrichment → deliverability check → CRM or email campaign push.

2. Viral Engager Harvesting: Post with high engagement threshold detected → fetch all engagers → Person Enrichment → filter for VP/Director titles → CRM or agent memory.

3. Inbound Lead Routing by Seniority: Webhook trigger → Person Enrichment using LinkedIn URL → seniority check → high-value: Slack alert to sales team; others: nurture sequence.

4. Competitor Mention Monitoring: Track competitor mentions → enrich mentioner → ICP filter → route to warm outreach sequence.

Source: https://help.trigify.io/en/articles/8836198-draft-common-workflow-patterns

ColdIQ Signal Taxonomy

ColdIQ's published trigger-selling taxonomy classifies intent signals into four tiers: Hiring signals, Funding signals, Tech Stack changes, and Behavioral intent (website visits, content engagement, review site activity). Social listening adds a fifth tier — Public Conversation signals — not captured by traditional data providers. Each tier requires different outreach timing and angle.

Prerequisites

• Target account list (CRM, CSV, or HubSpot/Salesforce export)
• ICP criteria defined (firmographic + role criteria for signal filtering)
• Social listening platform configured (Trigify recommended; alternatives: Brandwatch, Mention, Keywordsio for simpler setups)
• Enrichment provider for email and firmographic lookup (LeadMagic, Apollo, Clay waterfall)
• Outreach tool connected (Instantly, Smartlead, Salesloft, or HubSpot sequences)
• Slack or CRM for internal signal routing and rep alerts

Step-by-Step Process

Phase 1: Define Your Signal Taxonomy

Decide which social signals are worth acting on for your business. Not all conversations are buying signals. The highest-value signals:

Signal Type	Description	Urgency	Outreach Angle
Competitor engagement	Prospect comments on, likes, or shares competitor content	High — 24-48hr window	Offer comparison, address common pain with that competitor
Problem statement post	Founder or operator publicly describes a problem you solve	High — act within 12 hrs	Reference their exact words; offer a specific solution
Recommendation request	"Does anyone have a recommendation for X?"	Very High — act within 2 hrs	Direct reply + personal DM
Hiring for GTM role	Post about hiring SDR, RevOps, VP Sales — roles your product supports	Medium — 3-5 day window	Tie outreach to the hiring intent; new hires evaluate tools
Funding reaction post	Company or founder posts about closing a round	Medium — 1 week window	Congratulate + connect to common post-funding challenges
Pain-point comment	Target prospect comments on someone else's post about a problem	High — act within 48 hrs	Reference the comment thread; show understanding
Content engagement pattern	Prospect consistently engages with content in your category	Low/Medium — warm long-play	Nurture track; add to account monitoring

Mitchum's rule: only act on signals where you can write a specific, non-generic opening that references exactly what happened. If the signal can't anchor the first line of the message, it's not a strong enough trigger.

Phase 2: Configure Social Listening

Using Trigify (full agentic workflow):

1. Create social listening searches for your target topics, competitors, and keywords across LinkedIn, X, Reddit, and other configured sources.
2. Configure Signal types in the Signals tab — buying intent, hiring signals, competitor mentions, product complaints, leadership changes.
3. Set score/severity thresholds to filter noise; only fire workflows above your minimum signal confidence bar.
4. Set date windows — only act on signals from the past 24-72 hours; stale signals lose their timing advantage.

Keyword search strategy:

• Competitor names + pain phrases ("Competitor X is too expensive", "left Competitor X")
• Problem categories your product solves ("struggling with [problem]", "anyone solved [problem]")
• Recommendation requests ("looking for a tool that does X", "anyone use Y")
• Adjacent category launches ("just launched our outbound program")

Avoid over-broad keywords that generate noise (company names alone, generic industry terms). Signal quality matters more than signal volume.

Phase 3: Build the Enrichment and Qualification Layer

When a signal fires, the workflow must answer three questions before any outreach is initiated:

1. Is this person in ICP? (job title, company size, industry, geography)
2. Is this company on our target account list? (or does it match ICP firmographic criteria if not pre-listed?)
3. Is there a reachable email? (run through email enrichment with deliverability verification before pushing to sequencer)

Typical workflow node sequence:

Signal Trigger → Person Enrichment (LinkedIn → firmographics + title)
  → ICP Filter (Boolean: title match AND company size match)
    → True: Email Enrichment → Deliverability Check
      → Verified: Push to Sequencer + Slack Alert with signal context
      → Unverified: Add to CRM without email trigger
    → False: Discard or add to low-priority nurture list

Use Trigify's built-in AI qualification agents to assess whether a post or mention is genuinely relevant before enrichment credits are consumed.

Phase 4: Craft Signal-Anchored Outreach

The outreach message must reference the specific signal. Generic openers destroy the value of the signal layer.

Fake personalisation (signal wasted):

"Hi Sarah, I noticed you're scaling your GTM team. Curious if you'd be open to a quick call about how we help companies like yours..."

Signal-anchored message (Mitchum framework applied):

"Hi Sarah, Saw your comment on Chris's post about the SDR ramp problem — specifically your point about the first 60 days. We helped three teams at similar stages cut ramp from 90 to 45 days. Worth a conversation?"

Rules for signal-anchored messages:

• The first line must be traceable to the exact signal event
• Do not ask for a meeting in the first message; offer context or a question
• Keep total message under 100 words
• No product feature lists; one specific proof point tied to their signal
• If the signal is a comment, quote or closely paraphrase what they said

Phase 5: Route and Alert

Qualified signals with verified emails enter the sequencer. Uncontacted high-value signals (CEO, VP-level with no email found) get a Slack alert to the rep with the signal context included — enough for a manual LinkedIn outreach.

Signal routing tiers:

Account Tier	Signal Strength	Action
Named target account	Any qualifying signal	Immediate Slack alert + email enroll
ICP-match not on named list	High signal (recommendation request, problem post)	Email enroll + add to CRM
ICP-match not on named list	Medium signal (competitor engagement, hiring post)	CRM add + nurture track
Non-ICP	Any signal	Discard

Phase 6: Measure Signal Quality

Track signal performance separately from overall outreach metrics. The signal layer needs its own feedback loop:

• Reply rate by signal type — which signal types generate the highest reply rates? Reallocate monitoring budget to the highest-performers.
• Signal-to-meeting rate — of all signals that triggered outreach, what percentage booked a meeting?
• Signal freshness distribution — what percentage of signals fired within 24 hours vs. 24-72 hours vs. older? Older signals should be deprioritized.
• False positive rate — signals that triggered enrichment and outreach but were not actually relevant. Review and tighten keyword search or ICP filter criteria.

Review signal performance weekly during ramp; monthly once stable.

Output Format

Social intent monitoring system documentation:

1. Signal taxonomy — table of signal types with urgency, outreach angle, and platform sources configured for each
2. Listening configuration — keyword list, competitor list, topic categories, and score thresholds per platform
3. Qualification workflow — node-by-node flow diagram or table (trigger → enrichment → ICP filter → email verification → action routing)
4. Message templates — 2-3 signal-anchored message variants per top signal type, each under 100 words with first-line tied to the signal
5. Routing rules — account-tier-to-action mapping
6. Measurement dashboard spec — KPIs, signal-type breakdown, and weekly review cadence

Quality Check

• Signal types defined with urgency windows (not just categories)
• ICP filter criteria specified before enrichment step (prevent credit waste)
• Email verification in workflow before sequencer push
• All message templates reference signal in first line — no generic openers
• Slack or CRM alert configured for high-value uncontacted signals
• Keyword searches reviewed for noise — signal quality over volume
• Signal freshness thresholds set (24-72 hr maximum for high-urgency signals)
• Measurement plan covers reply rate by signal type, not just overall rate

Common Pitfalls

1. Acting on stale signals. A competitor mention from last week is not the same as one from yesterday. Set hard freshness windows: recommendation requests expire in 2 hours, most other signals in 24-72 hours. Signals older than 72 hours should go to a low-priority nurture track, not hot outreach.

2. Generic outreach from signal data. Running a signal-triggered sequence with the same generic opener you'd send to a cold list defeats the entire point. The signal must appear in the first sentence. If you can't reference it specifically, don't send the email.

3. Enriching before filtering. Running every signal through enrichment before checking ICP fit burns credits and slows workflows. Filter by job title and company size using free platform data first; only enrich confirmed ICP-matches.

4. Keyword search too broad. Monitoring your own company name or a one-word competitor name without context generates hundreds of irrelevant signals. Use phrase-level queries and negative keyword filters.

5. No feedback loop on signal quality. Running signal-led outbound without tracking reply rate by signal type means you never know which signal types are actually worth acting on. Some categories that seem strong (funding signals) often underperform industry-specific problem posts. Measure weekly.

6. Confusing social listening with social selling. Social listening is about monitoring what others say publicly (inbound signal collection). Social selling (the social-selling skill) is about building your own presence and engaging with your network. They work together but require separate workflows and metrics.

Execution Artifacts

• references/framework-notes.md — Named frameworks and reference tables
• templates/output-template.md — Deliverable shell for agent output
• scripts/check-output.py — Lightweight deliverable validator

Related Skills

• signal-scoring: Score and tier accounts across multiple signal sources; complements social monitoring with hiring, funding, and tech stack signals
• lead-enrichment: Enrich the people detected by social monitoring
• hiring-signal-play: Specific play for hiring-intent signals from job boards
• multi-channel-outreach: Coordinate email + LinkedIn + call after signal fires
• social-selling: Build your own LinkedIn presence that generates inbound signals
• cold-email-strategy: Write the signal-anchored emails that social monitoring triggers