Activation: 01 August 2023
Scope: Organic Search, Local Search & AI Search (Helium, Zone & AI Source programmes)
Measurement Period: 1 August 2025 – 30 June 2026 (11 months)¹

Measurement note: the Helium programme (activated 1 August 2024) continued throughout the measurement period, and Zone (activated 1 August 2025) was suspended with April 2026 its final month. Year-on-year comparisons therefore measure the hotel’s second programme year (the year in which AI Source was added) against its first. AI-specific metrics (AI mentions, AI rankings, AI referral traffic) isolate AI Source directly.

The property is a centrally located economy hotel in Singapore, operating under the flag of a major international hotel group, in one of the densest and most competitive hotel markets in the world. OmniHyper began working with the hotel in August 2024, activating the Helium programme (organic search) immediately after the hotel’s new microsite went live at the start of the engagement. Prior to this, the hotel had no microsite of its own; its direct digital presence was effectively starting from zero.

The first programme year delivered strong results, and the hotel re-signed Helium at the twelve-month mark, adding Zone (local search) and AI Source (AI search / AEO / GEO) on 1 August 2025. Zone did not generate the measurable results we hold ourselves to and was suspended, with April 2026 its final month. From May 2026, the hotel has operated Helium and AI Source only. This case study measures the eleven months from AI Source activation: 1 August 2025 to 30 June 2026.

Singapore’s underlying accommodation demand remained strong and stable across the period. Google Trends indexed search interest for “hotel singapore” (Singapore) held between 73 and 100 throughout the measurement window, with no structural decline.² Demand was not the challenge.

The challenge was where that demand is resolved. In the quarter before AI Source commenced (April–June 2025), the hotel’s website generated 294,775 Google Search Console impressions at a 3.61% clickthrough rate. In the same quarter one year on (April–June 2026), impressions had grown 357% to 1,347,397, but click-through had fallen to 1.31%.³ Absolute clicks still grew 65.8%, but the widening efficiency gap is the interception problem in a single statistic: a hotel can appear in more searches than ever, yet if OTA comparison boxes, Google’s hotel booking module and AI-generated answers occupy the screen above the organic results, a shrinking share of that visibility converts to a direct website visit.

This is compounded by structural invisibility on Singapore’s highest-volume hotel queries. A precinct-level query covering the hotel’s own neighbourhood (27,100 estimated monthly searches), “hotel singapore” (12,100) and “affordable hotels singapore” (9,900) all return results dominated by Booking.com, Agoda and Expedia across paid, organic and AI-generated results.4 These are the broadest entry points to Singapore hotel discovery, and they sit beyond the reach of any individual hotel’s traditional SEO. Over the same period, aggregate estimated search volume across the hotel’s tracked keyword portfolio declined 17% even as Google Trends interest held stable: the signature of journeys being resolved without a click, inside AI answers and AI-generated search results.4

AI Source was deployed to create a parallel pathway: making the hotel the answer inside the AI-generated responses that increasingly intercept these journeys before an OTA listing, or even a traditional search result, is ever seen.

The hotel faced four core challenges:

• Singapore’s highest-volume discovery queries are structurally owned by OTAs, in paid, organic and, increasingly, AI-generated results.
• A young microsite (launched at activation) still building authority in one of the world’s most competitive hotel markets.
• Search itself restructuring: impressions growing sharply while click-through falls, as AI answers and booking modules resolve journeys before the click.
• Effectively no presence in AI-generated answers at activation, in a market where AI assistants increasingly resolve hotel discovery before an OTA listing or search result is ever seen.

The objective was clear: make the hotel the answer in AI-assisted discovery, and convert that visibility into measurable direct demand.

Rather than treating AI visibility as a bolt-on, AI Source layered a structured answer-engine programme over the Helium organic foundation:

This included:

• Expanding the microsite’s answerable content footprint across the queries travellers actually research: precinct guides for the surrounding neighbourhoods, major Singapore events and concerts, attractions and practical travel planning.
• Structuring content so AI systems can retrieve and cite it: direct answers, FAQs and entity-consistent property information.
• Building and monitoring the property’s AI brand performance: mentions, sentiment, share of voice and rankings across ChatGPT, Gemini, Claude, Perplexity and Copilot.
• Monthly reporting by Mae, OmniHyper’s AI digital marketing agent, to hotel and brand stakeholders, reconciled each month against the hotel’s own commercial reporting workbook.

This created a measurable, auditable pathway from AI-era discovery to direct demand.

1 August 2025 to 30 June 2026, vs 1 August 2024 to 30 June 2025

This created a measurable, auditable pathway from AI-era discovery to direct demand.

Alongside visibility and engagement, the hotel’s own internal channel reporting, populated by the hotel team independently of OmniHyper and of Google Analytics, shows a clear commercial shift:

For completeness: indirect (OTA and third-party) revenue also grew strongly over the same period (+27.5%) in what was an exceptional demand year for Singapore, with major concerts and conferences driving citywide volume. Direct revenue growth is therefore reported here in growth terms rather than as share gain; lifting direct share remains the ongoing objective of the programme.10

A note on measurement: why hotel-reported figures are used

Google Analytics on the microsite is a last-click model that credits only bookings completed in the same tracked session. Over this period, that last-click view shows tracked purchases down 16.6% and tracked revenue down 8.2%, while on the very same property every upstream intent signal rose by 25–55%, and the hotel’s own systems recorded direct web revenue up 16.8%.6, 10

That divergence is not noise; it is the attribution problem this case study is designed to address. Guests increasingly discover the hotel through AI answers, search results and the microsite, then complete their booking on the brand’s central booking platform, by phone or email, or in a later direct session, none of which credit the microsite under a last-click model. This is precisely why commercial outcomes in this case study are drawn from the hotel’s own internal reporting, the system of record for what guests actually booked, with Google Analytics used for what it measures reliably: discovery, engagement and intent.

This case reflects a hotel that responded to a structural change in guest discovery ahead of its competitive set. Over eleven months, its search presence grew 4.3×, it moved from near-invisibility to 80 tracked mentions and top-3 rankings on 22 non-branded queries across AI platforms, and its own internal reporting recorded direct revenue up 16.4% year on year.

Attribution across channels remains inherently complex in a multi-touch, increasingly AI-mediated search environment. The measurement period also includes the ongoing Helium programme, Zone until April 2026, and marketing activity undertaken by the hotel and its parent brand, including strong social media growth over the same window. No single model can isolate one programme’s contribution, and this case study does not claim to. Instead, it triangulates five independently owned data sources, each accessible to the hotel group without any reliance on OmniHyper:

• Search presence: Google Search Console (Google’s own reporting).
• Local and map discovery: Google Business Profile (Google’s own reporting).
• Website engagement and intent: Google Analytics 4 (brand-accessible Analytics 360 property).
• AI visibility: AI platform results, verifiable across ChatGPT, Gemini, Claude, Perplexity and Copilot by the hotel group.
• Commercial outcomes: the hotel’s own internal channel reporting, populated by the hotel team.

TAll five move in the same direction over the same window, and the growth concentrates exactly where AI Source was engaged to act: non-brand discovery content, AI citations and answers, branded-search ownership, and direct channels.

The last-click paradox strengthens rather than weakens this case. If microsite last-click analytics were the only lens, this would look like a flat year. Booking-intent clicks up 25–55%, search impressions up 329.5%, AI mentions up 6.7× and direct revenue up 16.4% in the hotel’s own systems say otherwise. In an AI-mediated search market, direct performance must be measured across the systems that record what guests actually do, not just where the last click landed.

No single metric can prove attribution in modern search. But when Google’s search data, Google’s business profile data, the hotel’s analytics property and the hotel’s own commercial reporting all show sustained improvement over the same eleven months, concentrated in the areas AI Source was engaged to improve, they form a coherent body of evidence that is substantially stronger than any individual attribution model.

The methodology used in this case study was intentionally designed to minimise reliance on agency-reported metrics. Wherever possible, outcomes are drawn from systems owned by Google, the hotel group or the hotel itself, allowing every claim to be verified by the hotel group, and evidenced to prospective clients on request. All figures are verified against client-owned systems; underlying detail is available to prospective clients on request under NDA.