More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    GSC AI Performance: Search Console AI Report

    1

    Baseline and interpret Search Console's AI Overviews + AI Mode report honestly: what it measures, what no tool can tell you, probe reconciliation, and the opt-out call made on the record.

    $5

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    GSC AI Performance: Search Console AI Report

    GSC AI Performance: Search Console AI Report

    Example session with this skill installed

    Our SaaS docs site just got the AI Performance report. Leadership wants to know if AI answers are sending us traffic, and whether we should opt out while our compliance review is still running. First export attached — make sense of it.

    • Read your context and instructions
    • Compiled the gsc ai performance

    ONE-PAGE BRIEF — produced by applying the skill to the export (28 days, docs property).

    Baseline: 18,640 AI impressions. Top pages: the SSO setup guide (21%), the API rate-limits reference (14%), the billing FAQ (11%) — top 3 hold 46%, so visibility is concentrated in three pages. Country mix: US 48%, Germany 13%, India 11%. Device: 68% mobile.

    Traffic question, answered honestly: this report cannot say. No first-party tool publishes clicks or CTR for AI Overviews or AI Mode, and GA4 files any such visits under Organic Search with no split. Anyone who gives leadership a click number for AI answers has modeled it, not measured it. What exists is the impressions trend — next export in 28 days makes it a line instead of a point.

    Probe reconciliation: your 15-question probe set surfaced the brand in 4 answers, all on SSO topics. Verdict per the decision table: probe set too narrow — the report's second-biggest page (rate limits) is a topic the probes never asked about. Expand probes there.

    Opt-out decision: STAY IN, revisit at compliance review close. Scoring: docs are public (no paywall exposure), no reputation issue, and the compliance review concerns product claims, not the docs corpus. The one moderate row — corrections shipping on your own properties first — is handled by your docs CI, not by opting out and losing all AI visibility.

    Next actions: (1) extend probes to rate-limits and billing topics; (2) strengthen the SSO guide's citation-worthiness since it carries a fifth of your visibility; (3) re-export in 28 days and read the trend mix-first.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Google's AI Performance report is the first first-party window into how often your pages get used as grounding sources inside AI Overviews and AI Mode — and it is surrounded by people willing to invent the numbers it does not contain. This skill is the discipline for using it straight. It starts with the honest measurement map: impressions exist; queries, clicks, and CTR for AI answers do not, in any first-party tool, and GA4 cannot split AI-answer visits out of organic search. Every deliverable restates those limits instead of papering over them. The protocol then runs in order: confirm access (and record a missing report's documented reason as a finding, not a failure), export and archive a dated 28-day baseline, read the country and device mix before the totals, and reconcile the report against any synthetic AI-visibility probes with a decision table — report-over-probes gaps mean your probe set is too narrow; probes-without-report means the volume is not there. The opt-out decision gets its own scoring table (regulated content, paywalled material, active reputation management) and is never made without a named reason and a revisit date. From the second export onward, trend reading has guards: like-for-like ranges, mix-adjusted comparisons, concentration checks, and measurement breaks marked on the line instead of smoothed over. The deliverables are concrete: a one-page brief template, an archive naming pattern, a baseline checklist, and a brief trust rubric that blocks delivery if any figure cannot be traced to the export. If your team is reporting AI search visibility — or being asked to — this is how the numbers stay honest.

    • It covers Google's surfaces only (AI Overviews + AI Mode). ChatGPT, Claude, and Perplexity publish no first-party citation data; only labeled synthetic probing exists for them.
    • The report itself has no queries dimension and no click data — the skill will refuse to produce those numbers rather than estimate them.
    • A single export is a snapshot. Trend claims require the monthly series the protocol sets up; the first brief says so explicitly.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 24 days ago

    Permissions

    No special permissions declared or detected

    Listed24 days ago
    Updated21 days ago

    What's inside

    Frequently Asked Questions

    No — and neither can anything else. Google publishes impressions for AI answers but no clicks or CTR anywhere in Search Console, and GA4 cannot separate AI-answer visits from organic search. The skill's measurement map states this up front, and the brief is built from what does exist: impressions, page/country/device mix, and trend over monthly exports.

    Probably not. The documented reasons are too few AI-feature impressions for the property, or the site being excluded from Search's generative features. The protocol treats confirming which reason applies as step one and records it as a finding — a small or excluded property is a real answer, not a failed analysis.

    Only for a named reason: regulated or YMYL content whose E-E-A-T story is not audited yet, paywalled material the AI surface can partially expose, or active reputation management. Opting out costs you all AI visibility, so the skill scores your situation against a decision table, defaults to staying in, and attaches a revisit date to whichever way the call goes.

    New in Marketing & SEO