We built a guest profile platform for a global resort group. It matches records from each property's PMS into one profile that front desks can pull up instantly. Re-bookings rose 28%.
Client: Zenith EnterprisesIndustry: HospitalityService: SaaS & Software
Overview
Client
Zenith Enterprises
Global resort group, 500+ properties in 40 countries
Engagement snapshot
The challenge
The client ran 500+ properties across 40 countries, each with its own local PMS. Because data stayed with each property, a VIP guest at a Paris resort was a stranger when they checked in at a Maldives location.
Preferences were trapped in local systems, so service varied from one resort to the next. Without live spending data, AI assistants could not suggest relevant upgrades, and sharing data between regions by hand broke GDPR and local privacy rules.
Preference data stuck in local systems meant guests had different experiences across the portfolio.
Without live data on guest spending, AI assistants could not suggest relevant room or spa upgrades.
Sharing data between regions by hand broke GDPR and local rules on personal information.
The solution
Each resort publishes stay events to a global backbone, while personal data stays under local control and in its home region.
A matching engine combines different records for the same guest into a single golden profile, using hashed identifiers instead of raw personal details.
A fast API gateway serves the full guest profile to front-desk applications in under 50ms.
Key features
AI merges each guest's records from separate systems into one profile with 99.2% accuracy.
Business impact
50ms p95 profile lookup
Models read feedback and spending to predict which amenities bring guests back to book direct.
Business impact
28% more direct re-bookings
Machine learning models study guest feedback and spending to predict which luxury amenities will bring guests back to book direct.
Business impact
28% more direct re-bookings
The personalisation engine works on anonymised tokens, so marketing can reach guests without exposing who they are.
Business impact
100% of personalisation on anonymised tokens
The system scales automatically for the busy check-in window between 12 PM and 4 PM, with no slowdown.
Business impact
Sub-50ms at p95
Before and after
Before
Local databases
Guest information stayed at the property where the stay happened.
After
Shared across the group
Staff get instant, secure access to stay history and preferences at all 500 sites.
Before
Asking every time
Staff relied on guests telling them their preferences at every visit.
After
Ready before arrival
AI assistants set up room settings and welcome amenities based on past stays.
Before
Batch syncing
Loyalty points and profiles took 24-48 hours to sync between properties.
After
Live events
Guest activity reaches every property in under a second, so data is always current.
Results
28% Re-booking Lift
Personal offers based on unified stay data brought more guests back to book direct, which cut OTA commission costs.
Before: OTA HeavyAfter: 28% Higher
Sub-50ms p95 Response Time
Front-desk applications now pull up the full guest profile in 50ms instead of 3-10 seconds, so check-in no longer waits on the system.
Before: 3s - 10sAfter: 50ms
15% Operating Cost Savings
Consolidating 500 separate regional data sources into one governed platform removed duplicate cloud storage and compute.
Before: HighAfter: 15% Lower
Testimonial
“Across 500+ properties, each guest used to exist as separate records in separate systems. Coretus built the matching that merges them into one profile at 99.2% accuracy. Direct re-bookings are up 28%, and data costs fell 15%.”
Elizabeth WatsonVP of Guest Experience, Zenith EnterprisesKey results
Trust and control
Stay data stays in regional nodes to meet local laws, and only tokens sync globally.
SecureCustom hashing keeps each golden profile unique without storing personal details in plain text.
Aligned with GDPRIf a property loses its network connection, the master guest record is unaffected.
Ready to growThe client owns 100% of the PMS connectors and the golden profile engine.
100% ownedFaster delivery
These modules saved about 9 weeks of setup, which went into the guest matching rules and regional data residency.
Provided the hashing and matching base for building one golden profile per guest.
Gave us tested adapters for pulling stay events out of each property's local PMS.
Watched the flow of events between regions, so sync problems were spotted before they reached front desks.
Tracked storage and compute as 500 regional data sources were consolidated, which is how the 15% saving was measured.
Each property's PMS publishes stay events to a shared global backbone. A matching engine then combines the different records for each guest into one golden profile, using hashed identifiers rather than raw personal details, with 99.2% matching accuracy. Front-desk applications retrieve that profile through a fast API gateway in under 50ms.
Direct re-bookings rose by 28%, which reduced OTA commission costs. Front desks now load a full guest profile in 50ms instead of 3-10 seconds. Consolidating 500 regional data sources into one platform also cut data operating costs by 15%, and profiles sync in under a second instead of 24-48 hours.
Stay data stays in regional nodes to meet local laws, and only anonymised tokens sync globally. Golden profiles are built with custom hashing, so no personal details are stored in plain text. The personalisation engine also works on anonymised tokens, replacing the manual data sharing between regions that had broken GDPR rules.
If guest preferences are stuck in each property's own system, we can help you build one profile that every front desk can see.