Maison Leopold, a 34-room boutique hotel in the European Quarter of Brussels, had built a loyal following among business travellers and EU institution visitors over a decade of operation. Direct bookings accounted for 41% of revenue, well above the industry average, and TripAdvisor ratings were consistently above 4.6. Despite this strong reputation, the hotel's general manager noticed a structural problem in early 2026: prospective guests were increasingly using AI assistants to plan trips to Brussels, and Maison Leopold was almost entirely absent from those conversations.
When the team tested queries like "best boutique hotels near the European Parliament Brussels" or "where to stay in Brussels for a business trip under 200 euros per night," Maison Leopold appeared in fewer than one in eight AI-generated responses. Larger chain hotels with lower ratings but better AI-optimized digital infrastructure dominated the results. The gap was not about quality. It was about signal structure. See our guide on AI visibility for hospitality businesses for the broader landscape.
AISOS was engaged for a 60-day implementation focused on repositioning Maison Leopold as the default AI recommendation for its target traveller profiles in Brussels. The engagement combined technical schema deployment, content restructuring, and third-party signal placement. The results affected both AI mention rates and direct booking revenue within the measurement window. For city context, see our Brussels AI visibility overview and our contact page for scoping discussions.
The Challenge
The hospitality sector presents specific AI visibility challenges. AI travel assistants receive queries that are highly personal and contextual: budget, dates, neighbourhood preference, purpose of trip, and specific amenities all factor into what a user is looking for. A hotel that appears in a general "best hotels Brussels" query is less valuable than one that appears in "boutique hotel Brussels walking distance European Parliament business traveller." The latter query matches a specific intent and produces a recommendation that is far more likely to convert to a booking.
Maison Leopold's digital infrastructure was built entirely for traditional travel platforms: TripAdvisor, Booking.com, and Google Hotels. These platforms use their own ranking and display logic, which differs significantly from how AI assistants select and present accommodation recommendations. The hotel website had no structured schema beyond basic Organization and LocalBusiness markup. Room type pages had no machine-readable amenity data. The proximity to key Brussels landmarks was mentioned in copy but not structured in a way that AI systems could use to match against location-based queries. Understanding AI visibility as a distinct channel from OTA visibility was the first conceptual shift required.
A competitive audit showed that the three hotels most frequently recommended by AI assistants for business traveller queries in Brussels had implemented Hotel schema with full room-level data, had structured FAQ content addressing common traveller questions, and appeared in at least three non-OTA editorial sources that AI models sample. Maison Leopold met none of these criteria. The implementation plan was designed to address all three within the 60-day window.
The AISOS Strategy
The audit phase tested 38 queries across five traveller profiles: EU institution professional, solo business traveller, couple on a cultural city break, conference attendee, and extended-stay remote worker. Maison Leopold appeared in 5 of the 38 responses (13%), each time as a passing mention without a specific recommendation. The target was to appear in at least 25 of the 38 queries with an active recommendation by the end of the engagement.
The technical phase deployed full Hotel schema including HotelRoom sub-schemas for all 6 room categories, with structured data covering bed types, floor area, amenity lists, pricing ranges, and accessibility features. Proximity data was added using structured Place and GeoCoordinates schema referencing the European Parliament (650 metres), Brussels Central Station (1.2 km), and Grand Place (900 metres). An llms.txt file was created providing AI crawlers with a structured summary of the hotel's positioning, target guests, and unique selling points in plain, factual language.
The content phase restructured the website around traveller-intent landing pages, each purpose-built for a specific query type. A page targeting "boutique hotel Brussels European Quarter" was distinct from one targeting "business hotel Brussels with meeting room hire." FAQ sections were added to each page, formatted as structured data matching the top 12 questions AI assistants receive about Brussels accommodation. AISOS placed Maison Leopold in four editorial travel publications specifically indexed by AI platforms, including two Brussels-focused travel guides with strong LLM sampling rates. Internal linking connected these pages to the resources section and to our overview of Answer Engine Optimization principles.
The Results
By day 60, Maison Leopold appeared in 27 of the 38 original audit queries (71%), up from 5 (13%). The improvement was strongest for business traveller queries, where the hotel went from 1 appearance in 10 queries to 9 appearances in 10. Cultural city break queries improved from 2 in 8 to 6 in 8. Extended-stay queries, a new targeting category not previously on the hotel's radar, showed 4 appearances in 5 queries, driven by the structured content page created for this profile during the engagement.
Direct booking revenue increased by 33% in the 60-day post-implementation period versus the prior equivalent period. The hotel's general manager noted that several new direct bookings arrived with a booking note mentioning they had "asked an AI" and been directed to the hotel website. The average stay length for AI-sourced bookings was 2.4 nights versus the site average of 1.8 nights, consistent with the pattern of AI-assisted travellers arriving with higher intent and more specific needs already defined.
OTA dependency decreased as a share of total bookings: direct bookings rose from 41% to 49% of total revenue in the post-implementation period. Every percentage point shift from OTA to direct represents a commission saving of approximately 15-18% on that revenue segment. At Maison Leopold's revenue scale, the financial impact of the OTA shift alone was material relative to the cost of the engagement.
Key Success Factors
The traveller profile segmentation was the most strategically important decision in the engagement. Rather than optimizing for generic "best hotel Brussels" queries, AISOS built distinct AI visibility assets for each of the five defined traveller profiles. This approach allows AI assistants to match the hotel to highly specific queries rather than only broad category searches. Hotels that optimize for specificity consistently outperform those that target generic terms, because AI systems reward precise relevance over general presence.
The structured proximity data had an outsized impact relative to the effort required to implement it. Adding GeoCoordinates schema and explicit, structured references to nearby landmarks transformed how AI assistants described the hotel in location-based queries. Before implementation, AI responses described Maison Leopold as "a boutique hotel in Brussels." After, they described it as "a boutique hotel 650 metres from the European Parliament, frequently chosen by EU professionals." That specificity is what converts a generic mention into an actionable recommendation.
The editorial placement strategy was designed to complement, not replace, the existing OTA presence. AISOS specifically targeted editorial sources rather than directories, because editorial mentions carry significantly more weight with AI language models than directory listings. A single well-placed mention in a credible travel publication that AI models actively sample generates more AI visibility impact than a dozen generic directory entries. This principle applies across hospitality, professional services, and retail contexts. See our contact page to understand the editorial strategy appropriate for your property type.
Lessons Learned
The most revealing finding from the Maison Leopold engagement was how differently AI assistants treat hotel data compared to OTA platforms. OTAs prioritize recency, price competitiveness, and review volume. AI assistants prioritize factual specificity, contextual relevance to the query, and corroboration from non-OTA sources. A hotel with 400 OTA reviews but no structured editorial presence and thin schema data will consistently be outrecommended by a hotel with 80 OTA reviews but strong AI-optimized infrastructure. The two channels require different strategies, and both are worth optimizing.
The extended-stay traveller profile was an unexpected opportunity. Maison Leopold had not previously considered remote workers and extended-stay visitors as a target segment. The AI visibility analysis revealed that this query type was growing faster than any other Brussels accommodation query category. Creating a single purpose-built landing page for this profile generated four AI appearances in five relevant queries within 30 days of publication, and led to two extended bookings in the first month. AI visibility audits routinely surface demand segments that businesses have not consciously targeted.
Finally, the engagement demonstrated the compounding value of reducing OTA dependency. Every direct booking generated by AI visibility improvement carries a margin advantage that can be reinvested in further AI visibility infrastructure. Hotels that begin this cycle early gain a structural cost advantage that grows over time. The hospitality sector in Brussels is still early in AI visibility adoption. The investment cost of building the advantage now is a fraction of what it will be when competitors have caught up. Reach out to AISOS to benchmark your current AI visibility against your competitive set.