Visibility Wizard: 4 Setup Steps, 100 Tracked Queries, Zero Guesswork
The setup wizard walks every new domain through brand identity, target query selection, competitor confirmation, and engine activation - so what gets tracked is what the customer actually wants tracked, not a generic template.
One of 48 criteria in AEO Rank, the citation-readiness score we run against every site we audit.
By Alex Shortov
Quick Answer
The Visibility Wizard is the onboarding flow that turns a raw domain into a fully configured tracker. Step 1 confirms brand identity (services, audience, niche, CTA URL). Step 2 reviews the auto-generated target query set, pulled from Google Search Console plus niche enrichment, and lets the operator promote/demote/add queries up to a 100-query cap. Step 3 confirms competitors so brand-vs-competitor citation comparisons are accurate. Step 4 activates tracking across ChatGPT, Perplexity, Google AI Overview, and (where applicable) Reddit, then schedules the first recurring run.
Audit Note
In our audits, we've measured Visibility Wizard: 4 Setup Steps, 100 Tracked Queries, Zero Guesswork on live sites, we've compared implementations, and we've audited the gaps that keep scores low.
What does the wizard ask me to confirm before tracking starts?
Step 1 confirms brand identity including services, target audience, niche, geographic focus, and the CTA URL that future article CTAs will link back to.
How are the initial target queries chosen?
Step 2 pulls candidate queries from Google Search Console when connected, runs a commercial-intent classifier, and lets you promote, demote, or add queries up to 100.
Can I add or remove queries during setup?
Yes, you can add, promote, demote, or remove queries at any step and re-open the wizard later without losing the historical visibility data already collected.
What if I do not have Google Search Console connected?
Without GSC connected, the wizard falls back to a brand-profile-driven query generator that seeds the list from your services, niche, and competitor patterns.
How long does the whole setup take?
Full wizard setup takes 5 to 15 minutes of operator time across the four steps, with system processing typically under a minute at each stage.
Summarize This Article With AI
Open this article in your preferred AI engine for an instant summary and analysis.
What this article answers
- What does the wizard ask me to confirm before tracking starts?
- How are the initial target queries chosen?
- Can I add or remove queries during setup?
- What if I do not have Google Search Console connected?
- How long does the whole setup take?
Key takeaways
- The wizard is the single setup flow that captures everything the trackers need - brand identity, query set, competitor list, engine choices - in 5 to 15 minutes of operator time.
- Step 1 captures brand identity: services list, target audience, niche, geographic focus, and the CTA URL that future article CTAs will link to. The same data feeds the article writer’s brand profile so wizard answers shape every future article.
- Step 2 generates a target query list from Google Search Console (if connected) or a brand-profile-driven fallback (if not). Operator promotes/demotes/adds queries up to the 100-query cap.
- Step 3 confirms competitors so brand-variation matching in the trackers can distinguish “you got cited” from “a competitor got cited” on every query result.
- Step 4 activates tracking on every engine the customer’s plan supports - ChatGPT + Perplexity + Google AI Overview by default, with Reddit + Mentions for higher plan tiers - and schedules the first weekly run.
- Wizard answers are never one-shot. The customer can re-open the wizard at any time to refine queries, swap competitors, or pause an engine - all without losing historical visibility data.
Why a Wizard Instead of Sensible Defaults?
The temptation when building a visibility tracker is to skip onboarding entirely - take the domain, auto-generate a query list, start tracking. Faster activation, lower friction, prettier marketing copy.
We tried that. It failed for a specific reason: the auto-generated query list is wrong about 30% of the time. Even with good GSC data + niche enrichment, the top-100 target queries often include queries the customer doesn’t actually want to track (“brand history” on a query about Wikipedia mentions, generic head terms the customer has explicitly chosen not to compete on, queries that the customer’s own SEO team has flagged as off-brand). Auto-tracking ships visibility reports the customer either ignores or distrusts.
The wizard exists to fix that error rate at setup time instead of after the first weekly report. Five minutes of operator review saves an hour of “why are we tracking this query?” support conversation later, plus prevents wasted API spend on queries that won’t drive content decisions.
The wizard also captures brand identity that the trackers cannot infer reliably: which subsidiaries count as the same brand, which legal/trademark variations are valid, which competitors are actually competitors vs which are just similar-domain noise. Customer knows these. The classifier doesn’t.
Step 1 - Brand Identity
The wizard surfaces inferred brand name, niche, services, audience, and geography from a quick site crawl, then asks the operator to confirm or correct each field.
The wizard opens by surfacing what the system already inferred from a quick site crawl: brand name, niche, services list, target audience, geographic focus. The operator confirms or edits each field.
| Field | What it captures | Used by |
|---|---|---|
brand_name | The canonical name customers know | Brand-variation matcher (HIT detection in tracker responses) |
niche | Category descriptor (“live chat software,” “boutique insurance brokerage”) | Topic relevance filter in GSC query classifier |
services | Specific services or product lines offered | Topic-coherence enforcement in editorial calendar; article writer’s brand profile |
target_audience | Who the customer is selling to | Tone selection in article writer; FAQ question selection |
geographic_focus | Country/region/local | Geo-targeting in AIO tracker; SERP localization |
cta_url | Where article CTAs should link by default | Article writer’s CTA block; B-23 / B-29 link targets |
Brand variations and abbreviations get captured here too - “HelpSquad” and “Help Squad” and “HSQ” all map to the same brand for matching purposes. Aliases prevent the tracker from reporting MISS on a query that actually cited the brand via a common abbreviation.
A correctly-filled Step 1 takes about 2-3 minutes. Wrong answers here propagate everywhere downstream, so the wizard surfaces a confirmation summary at the end of the step before moving on.
Step 2 - Target Query Selection
This is the most operator-time-intensive step and the most valuable. The system generates a candidate query list and the operator reviews it.
If Google Search Console is connected: the wizard pulls the most recent GSC snapshot (up to 1,000+ queries with impressions, clicks, position metadata) and runs the keyword classifier described in how-we-track-chatgpt-visibility. The output is the top-ranked target queries from real search demand.
If GSC is not connected: the wizard runs a brand-profile-driven fallback. Codex CLI generates topic-phrase candidates from a site crawl + niche descriptor, scoring each by commercial relevance. Result is a list of plausible queries based on the customer’s content, not search data. Lower quality than the GSC path but workable.
Either way, the operator sees three buckets in the wizard UI:
- Target queries - what will be tracked (default top 100, capped at 100)
- Brand queries - excluded by default but visible (the customer probably wins their own brand searches)
- Noise queries - filtered out (adult content, foreign-language fragments, internal-search artifacts)
The operator can:
- Promote a query from brand/noise to target (manual override)
- Demote a target query to noise (we don’t care about this one)
- Add custom queries (typed in, optionally seeded by paste of a query list)
- Reorder to set tracking priority (top of list = most important)
The cap is 100 queries because that’s roughly the limit where tracker API spend stays under $5/domain/month at weekly cadence with three engines. Operators on higher plan tiers can negotiate cap increases per domain.
Time investment: typically 5-10 minutes for an operator who knows their business. A site with very specific commercial intent (one product, narrow audience) might finish in 2 minutes; a multi-line business with overlapping services might need 15.
Step 3 - Competitor Confirmation
The classifier infers competitors from crawl, niche, and GSC overlap, then asks operators to confirm, flag partners, add brand variations, and include any missed names.
The classifier infers competitors from the site crawl (linked external brands), the niche (common players in that vertical), and GSC data (sites also ranking for the same target queries). The wizard presents this candidate competitor list and asks the operator to confirm.
For each candidate competitor:
- Operator marks competitor / not a competitor / partner (partners get tracked separately for relationship insights, not as competitive threats)
- Operator can add brand variations for each competitor (so “LiveChat” and “LiveChat Inc.” and “livechat.com” all match the same entity)
- Operator can add competitors the classifier missed (typed in by domain)
Why this matters: the trackers use the competitor list to populate “your competitor was cited here” data in every visibility report. A wrong competitor list produces noisy reports - “your competitor X got cited 14 times this week” is useless if X is actually a partner or a non-competitor.
Time investment: typically 2-5 minutes. Customers with established competitive intelligence finish fast; customers in fragmented markets need longer to think through who actually counts.
Step 4 - Engine Activation
The final step picks which AI engines to track and starts the first run. By default, every plan tier gets at least:
- ChatGPT (via OpenAI API or ChatGPT Pro Codex CLI - cost-controlled depending on plan)
- Perplexity (via Sonar API - cheap enough to always include)
- Google AI Overview (via DataForSEO
google/ai_overviewendpoint)
Higher plan tiers add:
- Reddit monitoring - tracks where the customer’s domain or competitors get mentioned in Reddit threads, captures top-rising questions in the customer’s niche
- Mentions - off-page brand presence tracking across listicles, directories, comparisons (separate engine, fed back into the article pipeline via the same evidence-library mechanism described in original-data-pipeline)
- Bing/Copilot - tracked via the same provider that powers our ChatGPT tracking
The operator can disable specific engines if they don’t matter (e.g., “we don’t sell in the US, skip Google AI Overview which is US-en-en-US dominated”). The choice is per-domain, not global.
Once Step 4 is confirmed, the wizard:
- Persists every captured setting to
aeo_brand_profiles+ per-engine config tables - Enqueues the first visibility run (typically completes in 5-15 minutes per engine)
- Schedules the recurring run on the customer’s plan-tier cadence (typically weekly)
- Surfaces a “First report will land in your email within 24 hours” confirmation
What Happens to Wizard Answers Long-Term
Wizard answers are not frozen. The wizard re-opens any time from Studio’s Settings > Brand Profile and Visibility pages. Operators can:
- Add new target queries quarterly as the business expands into new products
- Drop queries that stopped being commercially relevant
- Add new competitors as the market shifts
- Pause an engine for a billing cycle without losing historical data
- Refresh GSC connection and re-classify queries (especially useful after a major site redesign that changed GSC query distribution)
History is preserved across wizard re-runs. If you tracked query X for 8 weeks then demoted it from the target list, the 8 weeks of data stays in the report - you just stop adding new data points to it.
How Long Does the Whole Setup Take?
Total operator time runs 10-20 minutes across all four steps, with the first usable visibility report landing 30-45 minutes after wizard completion.
| Step | Operator time | System time |
|---|---|---|
| Step 1 - Brand identity | 2-3 minutes | <5 seconds |
| Step 2 - Target queries | 5-10 minutes | <30 seconds (GSC import + classification) |
| Step 3 - Competitors | 2-5 minutes | <5 seconds |
| Step 4 - Activate engines | <1 minute | 5-15 minutes (first visibility run runs in background) |
Total operator time: 10-20 minutes. Total wall-clock to first usable report: 30-45 minutes, because the first visibility run kicks off at the end of Step 4 and takes 15-20 minutes to complete across all 3-4 engines. The first weekly digest email lands within 24 hours of wizard completion.
The visibility wizard runs four steps end-to-end, each with a different time budget and reviewer requirement.
| Step | Typical Time | Reviewer Required |
|---|---|---|
| Brand Identity | 3-5 minutes | No |
| Target Query Selection | 5-10 minutes | Yes, owner |
| Competitor Confirmation | 3-5 minutes | Yes, owner |
| Engine Activation | 2 minutes | No |
What Goes Wrong If You Skip the Wizard
We do not allow tracking to start without the wizard completing for a specific reason: tracking without wizard answers produces low-quality data that customers ignore. Specifically:
- Without brand identity capture, the matcher can’t recognize brand mentions in tracker prose, so HIT counts under-report by 20-40%.
- Without target query review, 20-30% of tracked queries are off-brand or noise, so the visibility report’s average HIT% is meaningless.
- Without competitor confirmation, “competitor mentioned” counts are random - sometimes including partners, sometimes excluding actual competitors.
- Without engine activation step, customers don’t know what they’re paying to track, which produces support tickets after the first invoice.
The wizard is the minimum viable spec to make the tracker output worth reading. We treat it as required.
External Resources
- AEO Studio Visibility - https://studio.aeocontent.ai/visibility
- Google Search Console (for GSC integration) - https://search.google.com/search-console
- AEO Brand Profile documentation - https://www.aeocontent.ai/knowledge/aeo-rank-methodology
Related topics
Key takeaways
- The wizard is the single setup flow that captures everything the trackers need - brand identity, query set, competitor list, engine choices - in 5 to 15 minutes of operator time.
- Step 1 captures brand identity: services list, target audience, niche, geographic focus, and the CTA URL that future article CTAs will link to. The same data feeds the article writer's brand profile so wizard answers shape every future article.
- Step 2 generates a target query list from Google Search Console (if connected) or a brand-profile-driven fallback (if not). Operator promotes/demotes/adds queries up to the 100-query cap.
- Step 3 confirms competitors so brand-variation matching in the trackers can distinguish 'you got cited' from 'a competitor got cited' on every query result.
- Step 4 activates tracking on every engine the customer's plan supports - ChatGPT + Perplexity + Google AI Overview by default, with Reddit + Mentions for higher plan tiers - and schedules the first weekly run.
- Wizard answers are never one-shot. The customer can re-open the wizard at any time to refine queries, swap competitors, or pause an engine - all without losing historical visibility data.