/digital-marketing-pro:geo-monitor

/digital-marketing-pro:geo-monitor

Script location. If your host does not set ${CLAUDE_PLUGIN_ROOT}, the scripts are in this plugin's scripts/ folder, next to skills/. Run them with Python 3.10 or newer: python, or python3 where python is not found (macOS, many Linux systems).

Purpose

Monitor and track brand visibility across generative AI engines. Systematically test how AI platforms respond to queries relevant to the brand, score visibility using a structured rubric, track changes over time, and identify opportunities to improve AI presence. This command provides a repeatable, quantitative framework for understanding where and how the brand appears (or fails to appear) in AI-generated responses — giving marketers the data they need to optimize for the emerging generative engine optimization (GEO) channel. Supports baselining, trend tracking, competitive benchmarking, and narrative alignment checks across all major AI platforms.

This skill is the RECURRING mode of the canonical AI-visibility scoring standard defined in /digital-marketing-pro:aeo-audit. It does not introduce a second scoring model: it applies the same per-platform 1-10 rubric + gates on a schedule and tracks it over time. The 0-100 GEO health score + A-F letter grade produced below is the trend view of that same data — a longitudinal roll-up for spotting momentum, not a competing scorecard. The 6 canonical surfaces (Google AI Mode, Google AI Overviews, ChatGPT, Perplexity, Gemini, Copilot) are defined once as the PLATFORMS constant in scripts/geo-tracker.py.

Probes vs first-party data (checked 2026-10-04)

The rubric scores below come from probes: queries we run and read. They show what an engine can say. They are not platform-reported numbers. Where a platform publishes its own visibility data, pull it into the same report as a separate, labeled column and never blend the two:

In the trend report, show the probe score and the first-party number in separate columns, each with its own source and date. If Bing Citation Share falls while probe scores rise, or the reverse, flag the divergence as a finding. Do not average it away.

Input Required

The user must provide (or will be prompted for):

• Target queries to test: Organized by intent type — brand queries ("What is [brand]?"), product queries ("[brand] [product] features"), comparison queries ("[brand] vs [competitor]"), and category queries ("best [category] tools"). Minimum 5 queries recommended for meaningful scoring. If not provided, the command will generate a default query portfolio based on the brand profile
• AI platforms to monitor: ChatGPT, Perplexity, Gemini, Google AI Mode, AI Overviews, and Copilot — default is all six (AI Mode added May 2026 — it is a distinct surface from AI Overviews and frequently selects different citations for the same query). The user can narrow to specific platforms if they only care about certain engines or have limited testing capacity
• Monitoring frequency: weekly or monthly — determines how often the brand should be re-tested and how trend data is bucketed. Weekly is recommended for active optimization campaigns, monthly for steady-state monitoring
• Competitor brands to benchmark against (optional): One or more competitor brand names to test with the same query portfolio — enables side-by-side visibility scoring to understand relative AI presence. If omitted, the report focuses solely on the user's brand without competitive context

Process

1. Load brand context: Read ~/.claude-marketing/brands/_active-brand.json for the active slug, then load ~/.claude-marketing/brands/{slug}/profile.json. Extract brand name, product names, category, key differentiators, and desired positioning to inform query portfolio and narrative alignment scoring. Also check for guidelines at ~/.claude-marketing/brands/{slug}/guidelines/_manifest.json — if present, load brand voice and messaging constraints. If no brand exists, ask: "Set up a brand first (/digital-marketing-pro:brand-setup)?" — or proceed with defaults.
2. Define query portfolio: Organize target queries by intent type — informational (what is, how does), navigational (brand-specific), transactional (buy, pricing, sign up), and comparison (vs, alternatives, best). If the user provided queries, classify them into these buckets. If not, generate a balanced portfolio of 10-20 queries from the brand profile covering all four intent types. Each query is tagged with its type for segmented scoring.
3. Test each query on each platform: For every query-platform combination, record the AI response and score brand visibility using the rubric — cited with link (10 points: brand is mentioned by name and a direct link to the brand's website or content is provided), mentioned by name (7 points: brand is explicitly named in the response but no link), concept referenced without attribution (3 points: brand's product, feature, or approach is described but the brand itself is not named), absent (0 points: brand does not appear in any form), misrepresented (-5 points: brand is mentioned but with incorrect, outdated, or damaging information). Record the full response text for narrative analysis.
4. Record results: Store each query-platform-result via geo-tracker (--result takes the rubric value: cited = cited with link, mentioned = named without link, concept-only = concept referenced without attribution, absent, misrepresented):

   python "${CLAUDE_PLUGIN_ROOT}/scripts/geo-tracker.py" \