kai-brand-pulse

Multi-platform brand intelligence pulse - collect cited public reputation evidence across web, news, YouTube, X, LinkedIn, Reddit, and review sites, then turn it into objection mining, content angles,

SKILL.md

GitHub

kai-brand-pulse - Multi-Platform Brand Intelligence

Kai root note: knowledge/, harness/, and scripts/ paths in this skill live in the Kai install, not the user's project. Resolve them against the first ancestor directory of this SKILL.md that contains a knowledge/ folder (the Kai plugin root, ~/.claude/kai, or the kai-cmo-harness repo). MARKETING.md, memory/, and any output files live in the current project. If a referenced scripts/ command is not available in this install, say so, skip it, and continue with the file-based guidance — never fabricate its output.

Run a cited brand pulse across public reputation surfaces. This is the Kai-native version of the Brand Monitor Agent pattern: collect evidence first, analyze each platform separately, then synthesize into marketing actions.

When to Use

  • You need a current public read on a brand, product, founder, or client.
  • You want objection mining for copy, sales, ads, social, or lifecycle messaging.
  • You need competitor positioning evidence before /kai-brand, /kai-competitors, or /kai-surround-sound.
  • You want weekly delta monitoring for new mentions, complaints, comparisons, and content opportunities.

When Not to Use

  • You need a full marketing audit. Use /kai-audit.
  • You need technical SEO or agent-readiness checks only. Use /kai-seo-audit or /kai-surround-sound.
  • You do not have approval to collect or report on a sensitive individual.
  • You plan to publish claims without citations. Stop and collect source-backed evidence first.

Command Shape

/kai-brand-pulse <brand> [competitors]

Local runner:

python scripts/intel/brand_pulse.py "<brand>" \
  --domain "https://example.com" \
  --category "category buyers ask about" \
  --competitor "Competitor A" \
  --competitor "Competitor B" \
  --out "workspace/brand-pulse/<brand>-YYYY-MM-DD"

Optional:

python scripts/intel/brand_pulse.py "<brand>" --skip-fetch
python scripts/intel/brand_pulse.py "<brand>" --wiki-dir "<brain-wiki-folder>"
python scripts/intel/brand_pulse.py "<brand>" --json

Live search uses SERPAPI_API_KEY when present. Without it, the runner archives the query plan and writes data gaps instead of inventing findings.


Phase 0: Context and Provenance

  1. Read MARKETING.md if present. Pull brand name, domain, ICP, category, competitors, positioning, and voice constraints.
  2. For client-facing or quantitative recommendations, load harness/references/audit-data-provenance.md.
  3. If a domain is available, run the shared source collector before writing final claims:
python -m kai.source_data.collect \
  --url "https://example.com" \
  --firm-name "<brand>" \
  --workflow brand-pulse \
  --mode sales_external \
  --out "workspace/brand-pulse-data"

Use sales_external, onboarding_connected, or internal_demo. Cite collector sources for domain, schema, sitemap, or metric claims. Use _data-gaps.md for missing access.


Phase 1: Collect Evidence

Run the Brand Pulse runner. It creates a raw archive, platform packets, a synthesis shell, and a delta-tracking database.

Default surfaces:

Surface Collection Pattern Why It Matters
Web Brand, reviews, alternatives, pricing, own-domain entity queries General entity footprint and objections
News Brand and category news queries Authority, recency, PR angles
YouTube site:youtube.com search fallbacks Reviews, demos, creator narratives
X site:x.com and site:twitter.com search fallbacks Fast-moving complaints, praise, comparisons
LinkedIn site:linkedin.com/posts and company fallbacks B2B proof, founder/category narratives
Reddit site:reddit.com search fallbacks Raw objections and buying-language mining
Review Sites G2, Capterra, Trustpilot, Clutch, Yelp search fallbacks Social proof, complaints, competitor context

The runner writes:

workspace/brand-pulse/<run>/
├── brand-pulse-data.json
├── _brand-pulse.md
├── _content-angles.md
├── _objection-mining.md
├── _surround-sound-actions.md
├── _monitoring-plan.md
├── _data-gaps.md
├── raw/
│   └── query-plan.json
└── platforms/
    ├── web.md
    ├── news.md
    ├── youtube.md
    ├── x.md
    ├── linkedin.md
    ├── reddit.md
    └── reviews.md

Every evidence item has a citation id. Do not make a claim unless it points to a citation id or a collector source id.


Phase 2: Platform Analyzers

Analyze one platform at a time before synthesis. Use the packet in platforms/<platform>.md as the only source material for that platform.

For each platform, produce:

Analyzer Output Questions
Repeated claims What does the market keep saying about the brand?
Objections What pain, doubt, pricing, trust, or support language repeats?
Proof gaps What proof do people need that the brand does not visibly supply?
Competitor context Which competitors appear beside the brand and why?
Content angles What can Kai write, publish, pitch, or test next?
AEO actions Which citations, pages, or entity signals should feed /kai-surround-sound?

Keep platform conclusions separate until all packets have been reviewed. This prevents one loud platform from swallowing quieter but useful evidence.


Phase 3: Cross-Platform Synthesis

After platform analysis, synthesize:

  1. Narrative map - What the public web thinks the brand is, who it is for, and what doubts cluster around it.
  2. Objection bank - Exact objection themes with cited examples.
  3. Competitor positioning - Where competitors own attention, proof, or trust.
  4. Content angles - Blog, comparison, social, email, ad, and sales enablement ideas tied to evidence.
  5. Surround-sound actions - Third-party citation, directory, review, forum, and own-domain AEO moves.
  6. Monitoring deltas - What is newly observed since the previous run.

Do not blur "observed in search results" with "market share" or "sentiment share." Treat search output as a sampled evidence packet.


Phase 4: Recommended Actions

Turn findings into Kai work:

Finding Next Kai Move
Repeated pricing objection /kai-landing-page, /kai-write, or sales FAQ refresh
Repeated competitor comparison /kai-competitors plus comparison page brief
Reddit objections /kai-reddit-listen profile keywords and reply guardrails
Thin review-site footprint Review request system, directory cleanup, or /kai-surround-sound
Strong third-party praise Repurpose into proof assets, ads, case studies, and AEO citations
Missing own-domain entity clarity /kai-brand, /kai-seo-audit, then /kai-surround-sound

For phone-led businesses, apply the Phone Capture Fit Rule. Recommend a phone answering service or AI receptionist only when phone-capture evidence supports it, and compare alternatives.


Phase 5: Weekly Delta Monitoring

Run the same brand weekly. The local SQLite database at data/intel/brand_pulse.db tracks first-seen and last-seen mentions.

Example cron:

0 8 * * 1 cd /path/to/kai-cmo-harness && python scripts/intel/brand_pulse.py "<brand>" --domain "https://example.com" --category "<category>" --out "workspace/brand-pulse/<brand>-$(date +\%F)"

For Brain wiki ingestion, pass --wiki-dir to write a pointer page to the latest cited packet. Keep the full raw archive in the workspace.


Quality Rules

  • Cite every client-facing claim with a citation id or source id.
  • Store raw search responses and query plan before synthesis.
  • Use _data-gaps.md for missing APIs, private platform access, or unavailable exports.
  • Never report review counts, rankings, traffic, share of voice, sentiment share, or platform volume unless the source directly provides them.
  • Label sampled search evidence as sampled search evidence.
  • Gate any publishable copy generated from the pulse with /kai-gate.