From 12% to 67% AI Share of Voice in 90 Days

In March 2026, Stayoa's head of marketing noticed something troubling. When she searched for "boutique coastal hotels in southern Europe" on ChatGPT — the kind of query their best customers use — Stayoa didn't appear. Not even close to the top. The results were dominated by Marriott's Edition brand and two aggregator sites.

Their traditional SEO metrics were healthy: page one Google rankings for core terms, a 34 DA, solid review scores. But in AI search, they were invisible.

By June 10, 2026 — 90 days later — Stayoa had a 67% citation rate on their target query set. This is the playbook.

Starting Position

Before starting, Stayoa ran a baseline audit using Genlytic across their 60 highest-priority query-engine combinations. The numbers were worse than expected:

• Blended AI SoV across 4 engines: 12.3%
• GPT-4o SoV: 8.1%
• Perplexity SoV: 4.4%
• Claude SoV: 17.6% (their strongest engine, attributed to a long-form destination guide published in 2024)
• Gemini SoV: 19.1%
• Branded query citation rate: 61% (they were being displaced even when users asked for them by name)

Their competitors in the same query sets averaged 34% SoV. The gap was 22 percentage points — meaningful but closeable.

Step 1: Prompt Tracking Setup (Days 1-7)

The first priority was visibility into what was actually happening — not just the final SoV number, but which queries, which engines, and which competitors were displacing them.

Stayoa set up Genlytic tracking across:

• 60 target queries spanning their three market segments (romantic getaways, family coastal trips, remote work travel)
• 4 engines: GPT-4o, Perplexity, Claude 3.7, Gemini 1.5
• 8 competitor brands most frequently appearing in their target query slots

Within 72 hours, they had a query-level breakdown showing exactly where they were being displaced and by whom. The data revealed something actionable immediately: on Perplexity, two competitors consistently outranked them despite lower traditional SEO authority. The differentiator was FAQ schema — both had it, Stayoa didn't.

Key action: Map your query-engine matrix before doing anything else. Knowing which engine has the biggest gap tells you where to invest first.

Step 2: Competitor Audit (Days 7-14)

With the query tracking live, Stayoa conducted a structured competitor audit — not of their competitors' websites, but of their competitors' AI footprint.

For each of the 8 tracked competitors, they documented:

• Which queries each competitor appeared in and at what position
• What content was being cited (specific pages, not just domains)
• What structural signals those pages had (schema, content length, update recency, external citations)

The audit took about 12 hours across the two-week window, mostly running Genlytic's competitor snapshot reports at different points in the day to catch variation.

Three findings emerged:

1. Schema gap confirmed: 6 of 8 competitors had FAQPage schema on their property pages. Stayoa had none.
2. Content depth gap: The top-performing competitor had 14 indexed pages per property (detailed amenity pages, nearby-activity pages, booking FAQ pages). Stayoa averaged 4 pages per property.
3. Recency signal gap: The two competitors with the highest Perplexity SoV had both updated their main property pages within the past 45 days. Stayoa's most important pages hadn't been touched in 8 months.

Step 3: FAQ Schema Implementation (Days 14-28)

With the competitor audit complete, Stayoa's developer spent one sprint on FAQ schema implementation. The scope was deliberately narrow: their seven property pages and their central booking FAQ page. Eight pages total.

For each property page, they wrote 6-8 Q&A pairs covering:

• What makes this property different from alternatives in the area
• Booking logistics (cancellation policy, check-in, early arrival)
• Room-specific questions (views, accessibility, family setup)
• Local area questions (transfers, nearby beaches, dining)
• Seasonal questions specific to their markets

The schema was implemented via JSON-LD in the page <head>, validated through Google's Rich Results Test before deployment, and pushed live on Day 22.

Step 4: Content Brief Execution (Days 28-70)

The schema work closed the technical gap. The next phase was closing the content depth gap.

Stayoa's marketing lead used Genlytic's query data to identify 12 content gaps — query types where they were consistently absent and competitors were consistently present. These became content briefs:

Destination guides (6 pieces): Long-form guides (1,200-1,800 words) covering their six target coastal regions. Not generic travel content — specific guides written from a boutique property perspective, with embedded FAQ sections and explicit coverage of the queries Genlytic identified as high-traffic in their segment.

Comparison content (4 pieces): "Boutique vs. chain hotels in [destination]" style content that directly addressed the query formats where aggregators and chain brands were most often displacing them. Counterintuitively, writing content that compares you favorably to chains is more effective than writing content that ignores them.

Seasonal content (2 pieces): Two pieces targeting high-intent summer travel queries with explicit date freshness signals built in.

All 12 pieces were published with Article schema including datePublished and dateModified markup, and structured with H2/H3 hierarchies matching the query patterns from Genlytic's tracking data.

Step 5: Monitoring Cadence (Days 70-90)

The final phase was establishing a sustainable monitoring rhythm. GEO is not a one-time project — AI model behavior shifts, competitors react, and query patterns drift seasonally.

Stayoa settled on a weekly review cadence with monthly deep audits:

Weekly (30 minutes):

• Review SoV trends across the 60 tracked queries
• Flag any query where SoV dropped more than 5 percentage points week-over-week
• Note any new competitor appearances in target query slots

Monthly (2-3 hours):

• Full competitor audit refresh
• Content update pass on the highest-impact pages (freshness signal maintenance)
• Query expansion — add new query variants based on emerging patterns in the tracking data

By Day 90, this monitoring caught one material shift: a competitor had deployed FAQ schema on pages that had previously lacked it, and was gaining on Perplexity. Stayoa responded by expanding their own FAQ coverage from 6-8 Q&As per property to 10-12, covering a wider range of query types.

Final Results — Day 90

Engine breakdown at Day 90:

• GPT-4o: 8.1% → 58.4% (+50.3pp)
• Perplexity: 4.4% → 71.2% (+66.8pp) — largest gain, fastest mover
• Claude: 17.6% → 68.9% (+51.3pp)
• Gemini: 19.1% → 69.7% (+50.6pp)
• Branded query citation rate: 61% → 97%

The 97% branded citation rate is the metric Stayoa's team is most proud of. When someone asks for them by name, they appear — reliably, consistently, in the primary answer slot.

The Five Levers, Ranked by ROI

Based on Stayoa's results and our analysis of attribution:

1. FAQ schema — fastest payback, lowest effort, highest Perplexity impact. Do this first.
2. Content depth on property/product pages — the most durable SoV driver, medium effort.
3. Destination/category guides — high GPT-4o impact, medium effort, 4-6 week delay to citation.
4. Freshness signals — low effort, prevents decay, important for Perplexity maintenance.
5. External citation seeding — slowest ROI but compounds; don't skip it, just start it last.

Stayoa is an anonymized case study. All performance data is from Genlytic's tracking platform. Results are specific to Stayoa's competitive context and query set — individual results will vary based on industry, competition level, and execution quality.