Get cited by ChatGPT and Perplexity

Content/SEO lead / growth marketer / CMO / GEO specialist · Assemble in ~15 min; first run in <2 min once installed.

GTMEdit flowon X◐ readiness 47

step 1ready

Define prompt set & competitor frame

geo-platform-optimizer28 alts

step 2ready

Run AI-search citation scan for brand mentions

nexus-geo

step 3ready

Audit brand authority/citation surfaces feeding the models

geo-brand-mentions1 alts

step 4partial

Compute share-of-voice vs competitors

competitor-analysis12 alts

step 5partial

Correlate with real AI-referral traffic (GA4/GSC)

ai-traffic-tracking

step 6ready

Prioritize and apply content fixes to rise in citations

seo-geo28 alts

step 7connector

Store baselines and detect rising/falling trend across runs

⚡ bring your own tool

Did this recipe work for you?

What you get

▸AI-search visibility scorecard (where you're cited, vs whom)
▸Share-of-voice trend report
▸Prioritized content fixes for AI citation

tools: optional AI-rank tracker subscription

tokens: ~$1-3 per sweep

≈ $0-50/mo

What stays yours

· Publish the content changes
· Pick which competitors to track

The problem

When buyers increasingly research in ChatGPT/Perplexity/Gemini/Claude instead of Google, I want to continuously monitor whether my brand is cited, with what share-of-voice, against which competitors, and whether it's rising or falling, so I can manage a channel that today is a complete blind spot.

Readiness

The measurement core assembles well from the library: nexus-geo (333) does the live citation scan, geo-brand-mentions (277) explains authority, geo-platform-optimizer (282) frames per-engine, and seo-geo (133) closes the act-on-findings loop. But the three things that make this a TRACKING product rather than a one-off audit — share-of-voice math, real AI-referral traffic, and the rising/falling trend store — are all partial or BYO. You get a strong snapshot-and-fix engine out of the box; you must bring analytics access plus a persistence + scheduling layer to make it continuous.

Cybernetic layers

Memorybyo

Trend detection is the whole point of this job, so state MUST persist across runs, but no artifact carries a store. Bring an external time-series: a CSV/Sheet of per-run scores, a SQLite/Postgres table, or a GEO-tracking SaaS that holds baselines. Without it every scan is a one-off snapshot.

Background / schedulingbyo

Continuous monitoring implies unattended weekly/daily runs of the prompt-set scan. The scan artifacts run on demand but none self-schedule. Bring a cron/Claude Code scheduled routine to fire the scan + diff and surface deltas; the harness has scheduling primitives but they must be wired up.

Harnessready

Claude Code runs the loop: it can drive the GEO skills, call WebSearch to execute prompts against AI engines, read GA4/GSC, and write results to the store. Invoked as a scheduled agent that loads the geo skills and the persistence layer each run.

Click any step to see what fills it — and the ranked alternatives you can swap in.