AI & Developer infrastructure
Multi-tenant hotel booking software
Key Achievements
Client
hivr.ai is an AI-powered sales platform built specifically for the Meetings, Incentives, Conferences & Events (MICE) segment of the hotel industry.
hivr.ai's AI agents automate the entire group-sales enquiry workflow for hotels — capturing, qualifying, and progressing enquiries so hotel sales teams can spend less time on admin and more time closing deals. The platform integrates directly with hotels' existing Property Management System (PMS) and CRM systems and is already used by 2,000+ hotels, including major hospitality groups such as Village Hotels and Radisson.
uxohtpbt8052ap4cuutv6j1rt3v7qv-small.webp)
6xdbc4z452vlmaaskki113q17kw7ho-small.webp)
Case study
The need
hivr.ai had already proven that AI agents could transform the backoffice side of hotel group sales. To close the loop end-to-end, hivr.ai needed a consumer-facing booking surface, a public search-and-book experience that event organisers and group travel buyers could use directly, with confirmed bookings flowing straight back into the hotel's sales pipeline as deals.
The missing piece was a white-label, per-client booking portal that could:
- Aggregate a single hotel group's own properties only (no shared marketplace exposure)
- Let end users search, configure, and price group bookings in real time
- Feed confirmed reservations back into the hotel backoffice automatically — no manual re-entry
Core issue
Delivering a production-ready booking platform for hospitality within a fixed 17-week window meant solving several compounding technical problems at once:
- Complex, multi-step booking flow — location/date/type search, space selection, per-session meeting room configuration (morning, afternoon, full day), room layout, food and beverage (F&B) packages, and real-time price/availability checks at checkout, all requiring tight coordination between frontend state and backend pricing logic.
- Multi-tenant, white-label architecture — each hotel group customer needed a fully isolated instance aggregating only their own properties, without duplicating the codebase per client.
- Real-time availability and pricing — prices and availability are date-sensitive and must be rechecked at checkout to prevent stale-data bookings, requiring careful optimistic-locking logic.
- An external payment dependency — the ADN payment gateway (hivr.ai's Stripe-equivalent) was being built in parallel by a separate team, creating a hard blocking dependency on checkout completion.
- A non-standard search layer — fast, filtered search across properties and spaces required integrating a custom Elasticsearch instance outside the standard .NET stack.
Our solution
Hotovo embedded a dedicated full-stack developer directly into hivr.ai's product team, taking full ownership of one of the platform's three core modules. Working within hivr.ai's chosen stack (.NET/Blazor), the developer contributed architectural input and integration design, not just implementation, in direct day-to-day collaboration with hivr.ai's internal team.
Rather than treating the booking UI as a thin layer over an API, the build centred on product-quality UX. This meant a guided multi-step flow with real-time feedback, date-aware pricing, cart-style space selection, and a checkout that gracefully handles race conditions on availability and price. The architecture was designed to be genuinely white-label from day one. Every new hotel group customer gets a fully isolated, configurable instance with no added engineering overhead.
What was delivered:
- Multi-step booking flow — full search → detail → cart → checkout → payment pipeline, filterable by location, date, space type (rooms/meeting/spaces/both), party size, and event type.
- Per-session meeting room configurator — room selection by time slot (morning/afternoon/evening/full day), including layout/shape and F&B package options.
- Real-time availability and price rechecks — checkout validates current price and availability against the backend before confirming, eliminating stale-data bookings.
- White-label, multi-tenant architecture — isolated, per-customer instances supporting hotel groups like Village Hotels and Radisson without duplicated engineering effort.
- Elasticsearch-powered search — sub-second, filtered property and space search across large portfolios.
- ADN payment gateway integration — completed transactions automatically surface as deals in the hotel backoffice.
- Super-admin / Adaptive Backoffice — internal tooling for tenant and property configuration, enabling fast onboarding of new hotel group customers.
- AI-assisted natural-language search (in progress) — foundations built for queries like "50 people for a celebration in London," deprioritised for initial release but ready to activate post-launch.
Key results
- On track for September 2026 release — the full public booking portal (search, space selection, meeting room configuration, checkout) delivered inside a 17-week timeline, with payment integration completing ahead of go-live.
- End-to-end booking journey delivered — all four stages (search → detail/cart → checkout → payment/deal creation) fully implemented, including real-time availability rechecks and automatic deal creation in the hotel backoffice.
- White-label architecture proven at enterprise scale — confirmed per-tenant isolation lets hotel groups including Village Hotels and Radisson run on separate branded instances with zero engineering duplication.
- AI search mode foundations in place — natural-language booking search pre-built and ready for post-launch activation, giving hivr.ai a clear AI differentiator on its roadmap.
Tech stack
Frontend
- Blazor and .NET
- Responsive web interfaces with Tailwind CSS
Backend
- ASP.NET Core and .NET
- REST APIs and modular backend services
- PostgreSQL databases
Search
- Elasticsearch
- Dedicated search service for properties, hotels, and meeting spaces
Payments
- Adyen payment gateway integration
Integration and platform services
- Keycloak authentication
- Azure Service Bus and event-driven integrations
- AWS integrations for selected supplier and messaging services
Team size
1 dedicated full-stack developer, embedded in hivr.ai's product team for the full 17-week engagement — specialising in Blazor/WASM, multi-step booking UX, Elasticsearch frontend integration, and white-label architecture.
AI-native team: All developers work AI-natively, using agentic coding tools (Roo Code, Cline) as standard practice across every Hotovo engagement – enabling faster delivery without sacrificing quality.
1y8bw9zwfkekqjq4zuqkimrxonh0ju-small.webp&w=3840&q=75)
bpbmnhkyjx1n9mc6e5yeyhtcluq9u0-small.webp&w=3840&q=75)
5bu3ff6jrd3rkfexehtdzzir8p1hs7-small.webp&w=3840&q=75)
ffejbupeoflmiy7b1rg817ztopi7hl-small.webp&w=3840&q=75)
9u80rif7ydl8z0gol4yocpxdjxa79j-small.webp&w=3840&q=75)
