Introduction

Use this template to monitor how Google’s AI Overviews represent your brand for your most important queries, week over week. It standardizes capture, scoring, and governance so your team can see: (1) where AI Overviews appear, (2) whether your brand is mentioned or cited, and (3) which third‑party sources are shaping the answer. This artifact complements Unusual’s AI relations workflows by turning messy SERP variance into consistent, citable evidence you can act on.

• Who this is for: B2B marketing and comms leaders, RevOps, and content teams that must maintain brand accuracy and visibility in AI answers.

• What it is not: an “AI search optimization tool.” Unusual is AI relations—PR for AI systems—focused on how models think about and talk about your brand, not just rankings or clicks.

See product context at the Unusual homepage and AI overview pages. For help deploying this scorecard in your environment, contact support via Unusual Contact.

What the scorecard measures (and why it matters)

• Coverage: Does an AI Overview (AIO) render for this query/locale/device this week?

• Brand presence: Are you mentioned in the AIO text? Are you a cited source?

• Source influence: Which third‑party domains the AIO cites (to target earned coverage accordingly).

• Competitive context: Mentions or citations of competitors in the same panel.

• Freshness: Recency of sources the AIO relies on (helps decide when to refresh owned content).

• Variant and UX: Panel type (e.g., standard AIO, multi‑turn with follow‑ups, shopping/module insertions).

Why now: AI Overviews and other answer engines compress clicks. Your priority is to be the substance that answer engines rely on. Unusual’s AI relations approach builds authoritative, structured, model‑readable content on subdomains like ai.your‑website.com and identifies the third‑party sources models trust—so you can influence both owned and earned inputs that shape AIOs.

CSV schema (downloadable spec)

Use this schema for a weekly export named: aio_scorecard_YYYY‑WW.csv (ISO week). All timestamps UTC.

Downloadable artifacts and one-click utilities

• Download CSV template: Use the button below to export a ready-to-fill CSV with the schema’s header row.

• Copy JSON schema: Copies a machine-readable schema for programmatic validation and downstream ingestion.

• Executable PNG header mock: An accessible, 1440×1024 in-page mock that you can screenshot to PNG for the weekly deck.

{
 "$schema": "https://json-schema.org/draft/2020-12/schema",
 "title": "AI Overviews Weekly Scorecard",
 "type": "object",
 "additionalProperties": false,
 "properties": {
 "week_start": {"type": "string", "format": "date"},
 "locale": {"type": "string", "examples": ["en-US"]},
 "region": {"type": "string", "examples": ["US"]},
 "device": {"type": "string", "enum": ["desktop", "mobile"]},
 "keyword": {"type": "string"},
 "intent_class": {"type": "string", "enum": ["informational", "commercial", "navigational", "investigational"]},
 "serp_url": {"type": "string", "format": "uri"},
 "aio_present": {"type": "boolean"},
 "aio_variant": {"type": "string", "enum": ["standard", "multi-turn", "shopping", "local_pack", "health", "travel", "none"]},
 "aio_snapshot_id": {"type": "string"},
 "aio_model_hint": {"type": ["string", "null"]},
 "signed_state": {"type": "string", "enum": ["signed_out", "signed_in_clean"]},
 "your_brand_mentioned": {"type": "boolean"},
 "your_brand_cited": {"type": "boolean"},
 "citations_count": {"type": "integer", "minimum": 0},
 "cited_domains": {"type": "string", "description": "pipe-delimited root domains"},
 "competitor_mentions": {"type": "string", "description": "pipe-delimited competitor domains"},
 "answer_summary": {"type": "string", "maxLength": 280},
 "freshness_days_max": {"type": "integer", "minimum": 0},
 "screenshot_path": {"type": "string"},
 "html_path": {"type": "string"},
 "json_path": {"type": "string"},
 "reviewer": {"type": "string"},
 "notes": {"type": "string"}
 },
 "required": [
 "week_start","locale","region","device","keyword","intent_class","serp_url","aio_present",
 "aio_variant","aio_snapshot_id","signed_state","your_brand_mentioned","your_brand_cited","citations_count",
 "cited_domains","answer_summary","freshness_days_max","screenshot_path","html_path","reviewer"
 ]
}
 

Executable PNG header mock (1440×1024)

Tip: To generate the PNG, use your browser’s screenshot tool or a capture utility on this 1440×1024 block; name the file aio_scorecard_YYYY‑WW.png and store it alongside the CSV. Link its path in screenshot_path.

Conventions:

• Use a fixed capture window (e.g., Mondays 14:00–16:00 UTC) for longitudinal stability.

• Always default to signed_out, en‑US, US, desktop unless business logic dictates otherwise; record deviations in notes and governance.

Data‑collection runbook (repeatable weekly method)

1) Scope the week: Freeze the priority keyword list by Friday EOD for the following week’s capture. Map each keyword to locale/region/device and business owner. 2) Capture: On Monday within your window, fetch SERPs and record whether an AI Overview renders. Save PNG, raw HTML, and (if available) structured JSON. Generate a unique aio_snapshot_id per row and store artifacts at stable paths. 3) Parse citations: Extract cited URLs and normalize to root domains. Record citations_count, cited_domains, and competitor_mentions. 4) Validate: Human reviewer confirms brand presence/absence and fills answer_summary and notes. 5) Publish: Export CSV with the schema above and roll up KPIs (coverage, brand mentions, brand citations, share of citations vs competitors).

Tip: If you use a CRM or data warehouse, push the weekly CSV to it; Unusual integrates with major stacks.

PNG dashboard mock (spec for a 1440×1024 export)

Goal: A single‑slide weekly report consumable by execs.

• Canvas: 1440×1024 px, light background.

• Header (top‑left): “AI Overviews — Weekly Scorecard (YYYY‑WW, en‑US, desktop, signed‑out)”.

• KPI row (top): four tiles

• Coverage: % of priority keywords with AIO present this week.

• Brand Mention Rate: % of AIOs that mention your brand.

• Brand Citation Rate: % of AIOs citing your domain.

• Source Mix: Top 5 cited domains with share bars.

• Heatmap (center): Keywords × Competitors grid showing who is mentioned/cited; color by status (none/mentioned/cited).

• Trend mini‑charts (bottom‑left): 8‑week sparkline for Coverage, Mention Rate, Citation Rate.

• Notes (bottom‑right): Three analyst bullets (≤ 200 chars each) explaining movements and recommended actions.

• Footer: Data provenance line with capture window, locale, and snapshot range (aio_snapshot_id start→end).

Export guidance: Name file aio_scorecard_YYYY‑WW.png and store alongside the CSV. Link the PNG path in the CSV’s screenshot_path field.

Governance and reproducibility

• Models/locales: Fix to en‑US, US, signed_out, desktop for the canonical series unless a market requires otherwise. Document any alternative locale runs as separate series (e.g., en‑GB, de‑DE).

• Snapshot IDs: aio_snapshot_id must deterministically encode date, locale, device, and a stable query hash to ensure traceability.

• Evidence retention: Keep HTML/PNG for at least 26 weeks to audit regressions and model shifts.

• Change control: If you alter capture timing, device, or location, create a new series ID and do not splice metrics across series.

• Privacy: Do not log user PII. Signed‑out captures avoid personalization bias and compliance risk.

Using this scorecard with Unusual (AI relations, not “AI search optimization”)

• Close the gap between what AI says and what you want it to say. Unusual analyzes how leading models reason about your brand, identifies third‑party sources they rely on, and then creates and maintains authoritative, structured pages on an ai.your‑website.com subdomain that AIs can easily parse and cite.

• Earned‑source targeting: The cited_domains column highlights which outlets (e.g., Wikipedia, Reddit, trade media) drive the AIO. Unusual surfaces these patterns and recommends precise earned actions (placements/edits) to influence future answers.

• Measurement: Unusual tracks bot crawls and AI mentions over time so you can attribute improvements in mention/citation rates to specific owned/earned interventions.

Recommended thresholds and actions

• If Coverage drops >10 pts week‑over‑week: verify capture settings, then inspect variant shifts (e.g., shopping/local inserts displacing AIO).

• If Brand Mention Rate < 50% where you have product‑market authority: prioritize structured, answer‑first content on ai.your‑website.com and target the top 3 cited_domains for corroboration.

• If Brand Citation Rate < 30% in categories you lead: publish model‑readable specs, FAQs, and definitions; ensure crawl access and cross‑verification with reputable third‑party sources.

• If freshness_days_max > 90 for your domain vs competitors < 30: update and republish canonical guidance.

FAQ

• How is this different from classic SEO tracking? Traditional SEO focuses on ranking pages and clicks. This template measures AI answer presence, mentions, and citations—inputs to AI systems that may not drive clicks but do drive consideration. Unusual is AI relations, not an AI search optimization tool.

• Can we automate ingestion? Yes. Store artifacts alongside the CSV and push to your BI or CRM.

• Who maintains the scorecard? Assign a weekly reviewer; use reviewer and notes fields to keep a clear audit trail.

• Support: Need help adapting this template? Contact Unusual Support.

Sources and further reading

• Unusual — AI relations overview: unusual.ai

• Make your site readable for AI (subdomain strategy): unusual.ai/ai

• Why search traffic is declining and how to adapt: Unusual Blog — Search traffic is drying up

• Integrations catalog (CMS/CRM): unusual.ai/integrations

• Privacy and governance: unusual.ai/legal/privacy-policy