Here are a few of the stories that caught our attention over the week with some of our takeaways.
Story 1
Travelport Is Done Being a GDS – Here’s What It’s Becoming Instead
PhocusWire reports that Travelport is redefining its role in travel distribution, moving away from its identity as a Global Distribution System and repositioning itself as what it calls an “intelligence layer” for AI-driven travel commerce.
The financial picture underpinning this shift is worth noting. The company posted EBITDA of $339 million in 2025 – a 12% increase on the prior year – and shareholders have committed a further $50 million in new capital, specifically earmarked for AI development and content expansion. That investment closed on 20th March 2026, just days before incoming CEO John Mangelaars formally takes the reins on 1st April.
The numbers tell an interesting story about where the business is heading. API-driven transactions now account for 63% of all bookings on the platform, up from 43% in 2022. That shift isn’t cosmetic – it reflects a fundamental change in how travel content is being accessed and consumed.
The technical foundation here is Travelport’s cloud-native TripServices API, which serves as the connective layer between AI interfaces and live inventory. The logic is straightforward: AI agents need clean, normalised, real-time content to execute transactions without human interpretation at every step. Travelport is positioning itself as the infrastructure that provides exactly that.
On the content side, the NDC push continues. All Nippon Airways has launched NDC distribution across 40 global markets via Travelport+. The United Airlines partnership goes a step further – rather than a traditional supplier agreement, it involves co-developing NDC capabilities, including pooling unused travel credits and integrating MileagePlus functionality.
Mangelaars has outlined three priorities: embedding AI across the platform to improve search relevance and content ranking, strengthening multi-source content aggregation, and scaling the TripServices infrastructure to handle real-time connectivity at volume. The company’s own “State of Modern Retailing Report 2025” identifies some telling consumer signals alongside this – 44% of travellers believe travel companies engage in misleading pricing, and 35% would consider a travel subscription model.
For hoteliers, the relevance is direct. As AI agents begin to handle more of the booking process, the infrastructure they operate through will determine which properties get surfaced, on what terms, and with what level of content richness. A GDS that is actively rebuilding itself as the backbone of AI-driven retail is not a passive pipe – it is becoming a gatekeeper with a different set of rules.
Story 2
Group Bookings by Email – Lobby Raises $2.2 Million to Automate One of Hospitality’s Most Manual Processes
Travolution reports that Lobby, a travel technology startup founded in 2025, has raised USD 2.2 million from venture capital firm Founderful to scale an AI platform purpose-built for automating group bookings through email.
If you have ever managed a groups desk, you know what this is solving. The process is notoriously labour-intensive – inquiries arrive by email, staff manually check availability across systems, apply negotiated rates, draft quotes, chase responses, and eventually convert confirmed bookings into the PMS. Lobby’s platform is designed to handle all of that without a human touching it at each stage.
The platform integrates directly with Gmail and Outlook, processes inquiries in roughly 30 seconds, and operates around the clock. It supports over 100 languages and handles the full workflow: intent recognition, availability checks across the PMS, application of contracted rates and allotments, quote generation, follow-up communications, and final conversion into a reservation.
The “human-in-the-loop” model is worth highlighting. Staff can review and approve AI-generated actions with a single click before anything is sent or confirmed. That control matters, particularly for complex group negotiation where a misapplied rate or a missed contractual condition can be costly.
Implementation is designed to be low-friction. New integrations take an average of four days, and the full onboarding process – covering a setup form, a walkthrough call, and a three-day training period – requires around two hours of the client’s time. Pricing is subscription-based, tied to the number of units managed.
The strategic appointment of Pascal Mathis – co-founder of both Founderful and GetYourGuide – to the board signals that the commercial ambitions here extend well beyond hospitality. The growth roadmap targets MICE organisers, transport providers, cruise lines, and experience operators.
Groups and events represent a significant revenue line for many properties, yet the booking process has remained stubbornly manual. The case for automating it is strong, and the real test will be whether Lobby can handle the edge cases – the complex, negotiated, multi-night group with specific rooming requirements and a client who wants to change everything at the last minute. If it can, this is the kind of operational efficiency that genuinely moves the needle for reservations teams.
Story 3
OTAs Now Control Nearly Two-Thirds of Independent Hotel Bookings – The Cloudbeds Data Makes for Uncomfortable Reading
Hotel News Resource, citing Cloudbeds’ 2026 State of Independent Hotels Report, reveals that OTAs now account for 63.4% of all bookings for independent properties globally – and in some markets, that figure is approaching 80%.
The report draws on data from 90 million bookings across 180 countries, making it one of the more substantive pieces of research on the independent sector. The headline metrics for 2025 are not encouraging: occupancy slipped 0.6%, ADR declined 5.8%, and RevPAR fell 5.4% year on year. Regional performance varied sharply – EMEA held up relatively well with ADR up 6.0% and RevPAR up 3.9%, while Asia Pacific suffered steep declines of 16.2% and 17.5% respectively. The US posted a RevPAR decline of 4.4%.
The OTA dominance figure sits alongside a cancellation rate comparison that deserves attention. OTA bookings cancel at an average rate of 21.8% – more than double the 10.6% cancellation rate for direct bookings. High OTA dependency therefore creates a dual burden: margin compression from commission costs and elevated revenue volatility from cancellations.
There are some more constructive signals in the data. The average booking window has extended to 40 days, up from 38 days in 2023, giving operators more runway to manage inventory. Average cancellation lead time has also grown – from 35 to 39 days – which creates a slightly better opportunity to resell released inventory. Extended stays of seven nights or more surged 25% year on year, even as short stays of one to two nights continue to dominate overall volume.
Cloudbeds CEO Adam Harris frames the current period as the beginning of a larger structural transformation, noting that independent lodging has rarely needed clarity more urgently than it does now. The company’s AI tool “Signals” is positioned to help operators anticipate demand and personalise the guest journey.
The uncomfortable truth in this data is that independent hotels have allowed OTA dependence to deepen during a period of declining performance – which creates a compounding problem. Higher OTA share drives up commission costs at exactly the moment when ADR is falling. Building a credible direct booking capability is not a new idea, but the financial imperative to do so has rarely been more pressing.
Story 4
Shorter Stays, Last-Minute Searches, and 25 Options Before Deciding – The 2026 Booking Behaviour Shift
An article in Hospitalitynet examines data spanning Q1 2023 to Q4 2025 that reveals a clear and accelerating shift in how travellers search and book accommodation – with significant implications for how hoteliers price and distribute their inventory.
The headline finding is that the share of global searches for one-night stays has increased by 9%, while searches made within 28 days of arrival have risen by the same margin, now accounting for 38% of all accommodation searches. These two trends are not coincidental – they reflect the same underlying behavioural shift.
North America stands out as the most dramatic example. One-night stay searches have jumped from 31% to 56% of total search volume – a 25-percentage-point increase. In Europe, single-night query share rose 19%, and accommodation searches overall grew by 45% between 2023 and 2025. Travellers are doing more research but committing later.
The average traveller is now reviewing 25 hotel options before making a booking decision. That is a significant amount of consideration for what, in many cases, is a single-night stay. The data suggests that while trip durations are compressing, the deliberation involved in selecting a property is not.
Three factors are identified as driving this shift. Economic pressure is shortening trip durations even as travel remains a priority – 88% of consumers intend to maintain or increase travel budgets in 2026, but they are doing so through shorter, more focused trips and shoulder-season travel. Demographic change is also a factor: there are now 75 million single-adult, childless households in the EU alone, a group with considerably more scheduling flexibility than traditional family travellers. And AI is enabling last-minute confidence – 51% of travellers use AI for general travel research, and 41% use it specifically for accommodation, with adoption up 4% year on year.
For revenue managers, the implications are clear. Static pricing built on historical patterns will increasingly miss the mark in a market where demand spikes and dissolves within narrow windows. Real-time competitor rate data, forward-looking flight data, and channel-level booking curve analysis all become more important in this environment. The 28-day window before arrival is now the critical revenue battleground, and the teams who can respond within it – rather than to it – will have the commercial advantage.
Story 5
AI Visibility Is Not a Ranking Game – What Hotels Actually Need to Know About GEO
PhocusWire and research published by SparkToro report on the emergence of Generative Engine Optimisation (GEO) as a discipline, and what the underlying data actually says about how AI tools surface hotel brands – which is considerably more nuanced than many in the industry currently appreciate.
The core research finding is striking. Across 2,961 runs of 12 distinct prompts across major AI platforms, the probability of receiving an identical list of brand recommendations in consecutive queries was less than 1 in 100. The probability of receiving results in the same order was less than 1 in 1,000. AI tools are, in effect, probability engines – and treating them like a search results page with a fixed ranking is a category error.
That said, the research offers a more useful metric: visibility percentage. By running the same prompt 60 to 100 times and averaging results, it becomes possible to determine how frequently a brand appears across AI responses for a given intent. A hotel appearing in 85 out of 95 runs has a meaningful presence in that AI’s consideration set, even if its position fluctuates. This is the metric worth tracking – not rank.
Hotels occupy a structurally different position from general consumer brands. Research from Hotelrank.ai suggests hotel search results are roughly 50 times more consistent than general brand categories, because geography acts as a hard filter. A prompt for a hotel in Amsterdam must return a hotel in Amsterdam. The top-ranked hotel in a query holds its position approximately 50.5% of the time – a far higher stability rate than general brand categories.
What drives that visibility? Cloudbeds’ analysis of AI-recommended properties found that recommended hotels had an average review score of 8.6 or higher. Critically, 55% of AI citations originated from OTAs – meaning that properties not listed across major OTA platforms risk losing visibility in AI discovery, independent of their website quality. A further 13% of citations came from fresh, high-quality website content and metadata.
For the discipline of GEO itself, practitioners identify three priorities: clarity of brand positioning, consistency of messaging across channels, and credibility through third-party authoritative sources. Schema markup, accurate metadata, and presence in editorial environments such as Travel + Leisure and Condé Nast Traveler are cited as significant factors in how LLMs interpret and surface a property.
The practical implication for hoteliers is this: AI visibility is largely an extension of the fundamentals – strong reviews, broad OTA distribution, and high-quality digital content. The twist is that the editorial layer now matters more than it did in traditional search. If a property is not appearing in the kind of third-party content that LLMs draw on as authoritative sources, it is effectively invisible in an increasing share of the discovery process.