Five hotel & travel distribution stories that caught our eye this week

21st of July 2026: Here are a few of the stories that caught our attention over the week with some of our takeaways.

Connectivity Stopped Being Plumbing: Inside the Data Expedia Left Out of Its Own Blog

Hospitality Net’s opinion piece “The connectivity gap: what sets high-performing hotels apart” makes the case that full connectivity has stopped being a back-office technicality and become the price of admission for competing in distribution. (Note: this is an older story)

The underlying research comes from Expedia Group, working with Censuswide, which surveyed 1,500 hotel decision-makers across six markets – the US, Canada, Mexico, France, Italy and South Korea. Fully connected properties are 31 points more likely to report improved occupancy, ADR or RevPAR than those running on basic connectivity, and 81% of fully connected properties say connectivity improved those metrics directly. Expedia is using this data to build the case for what it calls “autonomous distribution” – a three-phase software layer covering onboarding, management and optimisation that runs an increasingly large share of a hotel’s commercial operation through Expedia’s own systems.

The figure worth sitting with is a quieter one. Some 32% of unconnected hoteliers cite fear of losing control over pricing and inventory as their reason for holding back. That statistic appears in Expedia’s press release on the research but is notably absent from Expedia’s own blog post covering the same findings. A platform selling connectivity as pure upside has quietly dropped the resistance data from its own marketing – worth keeping in mind when the same pitch reaches your inbox.

The same week produced a second, related piece of positioning. RateHawk, the B2B platform belonging to Emerging Travel Group and marking its 10th anniversary, published a ten-trend report with Phocuswright on PhocusWire. One trend argues that hotel supply needs to be standardised into machine-readable formats before AI agents can book it at all – machine-readability becomes a precondition for being bookable, not an optional upgrade.

Read together, Expedia and RateHawk are making a similar underlying claim from different angles. Expedia is building the operating layer directly into a hotel’s commercial workflow; RateHawk is positioning its wholesale and API network as the plumbing agentic booking systems will need to function. Neither is disintermediation in the sense hoteliers have worried about for years. Both are re-intermediation – a new layer inserted one step removed from the guest relationship, owned by the same handful of large distribution players rather than by the property itself.

For commercial teams, the practical question isn’t whether to connect – that argument looks increasingly settled – but how much of the pricing and inventory decision-making sits inside a platform’s own software once you do. The connectivity data is real. So is the control question the research quietly sidesteps.

The Invisible Shelf: How Booking’s B2B Consolidation Could Undercut Your Parity

Hospitality Net’s opinion piece “Booking is scaling the machine that wholesales your rooms” lays out how Booking Holdings is bringing Booking.com, Agoda and Priceline together under a single B2B operation.

The unit is built to power travel for other companies – banks, airlines, loyalty portals – rather than to sell directly to travellers, and it’s led by Agoda chief executive Omri Morgenshtern, who built Rocket Travel by Agoda before taking on the wider role. The move is explicitly modelled on Expedia Partner Solutions, Expedia’s most durable asset and the division responsible for wholesaling inventory into other companies’ checkouts. Expedia’s B2B business has been growing at roughly twice the rate of its consumer-facing business, and that gap is what Booking is now moving to close.

A note on sourcing is worth flagging here. The confirmation that this is a formal legal entity was first reported by Skift on 10 July, citing a Booking Holdings spokesperson directly, but that specific confirmation sits behind Skift’s paywall. Treat the “confirmed” framing with appropriate caution if you can’t verify it yourself. The underlying announcement was made publicly at a Bangkok media roundtable back in March, so the direction of travel isn’t new – what’s new is the formal structure now being put around it.

The commercial mechanism already exists under the Rocket Travel by Agoda name – an accommodation-distribution API that bundles Booking-group and third-party wholesale supply into partner front ends, including Mastercard loyalty redemption, bank rewards portals and airline checkouts. The logic behind consolidating it is a hedge. If AI agents become the new front end for travel search, Booking would rather be the supply layer behind whichever interface wins than bet the company on owning that interface itself.

For a hotel, the practical exposure sits in where net rates end up. Rates handed to Booking or Agoda can now reach partner shelves – a bank’s rewards tab, an airline’s hotel add-on – where a hotel has no visibility of, or control over, the final price shown against its own parity. That’s a channel a revenue manager can’t easily audit, because it doesn’t look like Booking.com or Agoda on the surface; it looks like whichever bank or airline happens to be doing the selling. Worth pairing with existing rate parity and wholesaler-content monitoring, since this exposure sits one layer further from view than the wholesale relationships hotels are already used to policing.

$40,000 a Year: The Arithmetic Behind Fora’s New Unicorn Status

Forbes’ piece “Can $1 Billion Fora Travel Challenge Expedia And Booking.com?” covers the New York host agency’s $60M Series D, announced 16–17 July.

Fora, founded in 2021, lets newcomers become travel advisors using its own client-communication and booking infrastructure. The round was led by Forerunner and Tactile Ventures, with continued participation from Thrive Capital, Insight Partners and Tribeca Venture Partners, taking total funding to date to $138.5M and pushing the valuation to $1B. Fora reports 15,000 active advisors across 180 countries who have collectively booked over $3B in travel since 2021, with roughly a third of that total booked in the last five months alone. The new capital is earmarked for Via, its AI assistant currently in beta that automates advisor admin work, plus geographic expansion and growth in cruise and air.

The detail that sharpens the story: 97% of Fora’s advisors joined as first-time travel agents. This is a recruitment and training platform monetised through membership fees and tech tools, not a consolidation of experienced advisors built on the traditional GDS or host-agency model. Forbes frames the raise as a distribution-war signal – human-plus-AI advisors positioning against Expedia and Booking by combining platform reach with the trust of a named advisor relationship.

The arithmetic is worth doing before accepting that framing at face value. $3B booked across 15,000 advisors over five years works out to roughly $40,000 per advisor per year – a trip or two annually rather than a full-time luxury desk – which sits in some tension with the unicorn coverage most outlets have run with.

For a hotel audience, the commercial exposure runs through Fora’s affiliations. The agency carries Virtuoso, Four Seasons Preferred, Rosewood Elite, Hyatt Privé and Rocco Forte Knights relationships across more than 4,500 preferred partnerships by its own count. Any of those 15,000 largely new advisors can generate a booking that carries a property’s preferred-partner amenity costs – upgrades, breakfast, late check-out, resort credit – regardless of how experienced that individual advisor happens to be. That’s a cost line worth watching separately from the funding headline, since amenity commitments scale with advisor volume rather than advisor tenure.

llms.txt Debunked: Google Says the File Hotels Are Paying For Doesn’t Work

Search Engine Journal’s report “Google’s Mueller Says llms.txt Can’t Help LLMs Differentiate Sites” covers comments made by Google’s John Mueller on the Search Off the Record podcast.

Mueller’s framing was direct: publishing an llms.txt file is essentially telling AI systems that your website is the best one out there – comparable, in his view, to the old keywords meta tag, in that it’s self-reported, unverifiable and not something any AI engine actually checks against the site itself. He allows one narrow legitimate use: helping an AI agent navigate a site cleanly once it’s already there, which is the job the standard was originally designed for by Jeremy Howard in September 2024, aimed at developer documentation. Stripe and Cloudflare use it exactly that way, for coding assistants reading their manuals – a different job entirely from being discovered or ranked by an AI engine.

Independent research backs up the scepticism. Search Engine Land tracked ten sites across finance, SaaS, ecommerce and other sectors for 90 days before and after llms.txt implementation and found no measurable shift in AI crawl frequency or referral traffic from ChatGPT, Claude, Perplexity or Gemini. Separately, Ahrefs’ server-log analysis of 137,000 domains found that 97% of published llms.txt files received zero requests.

Digital Fox’s reference guide, published 17 July and aimed squarely at a hospitality readership, supplies the sharpest hook for hoteliers specifically: vendors are charging hotels up to several thousand pounds a year for this file as an “AI visibility” fix. The honest position as of mid-2026 is that Anthropic, OpenAI, Google and Perplexity have all declined to confirm they use llms.txt for anything resembling ranking or citation.

The redirect for a hotel commercial team is straightforward. That budget belongs in Google Business Profile optimisation and review management instead. Independent research cited in the Digital Fox piece found that AI Mode still routes roughly 79% of hotel-link interactions through Google Business Profile rather than through OTA or brand-site citations. Before signing an invoice for an “AI visibility” file, it’s worth asking the vendor directly which AI engine has confirmed using it, and for what. On the evidence so far, none has.

Margin Squeeze, Not Disintermediation: What BTIG’s Research Means for OTAs

PhocusWire’s report “AI risk to OTAs is margin squeeze, not disintermediation” covers new research from investment bank BTIG, part of U.S. Bancorp.

BTIG surveyed 250 US leisure travellers on 26 June and separately audited the AI trip-planning offerings of OTAs – Booking.com and Trip.com – against general AI platforms including Google Gemini, ChatGPT and Anthropic’s Claude. Consumer appetite for AI in travel is real and growing: over 60% of travellers have used AI for travel planning, with 76% comfortable using it for inspiration, 80% for research and 65% for itinerary building. Some 38% would use AI to book a flight and 42% a hotel room.

BTIG’s core finding is that willingness has outrun functionality. Its verdict on the tools themselves is blunt: rudimentary, with little personalisation, only rough pricing indication and no actual booking capability – and the landscape has not moved meaningfully since BTIG ran the same exercise in 2023. The bank’s headline conclusion is that AI is more likely to become a customer-acquisition channel for OTAs than a disintermediation threat, because OTAs hold decades of behavioural data, millions of connected properties and the marketing budgets – over $20bn combined across Airbnb, Booking Holdings, Expedia and Trip.com in 2025, up from $17.8bn in 2024 – to remain the partner of choice. The more provocative reframe is that price-comparison sites and traditional search may face the bigger threat from AI platforms, not the OTAs themselves. Exposure today is small in absolute terms – AI platforms account for only about 0.4% of OTA traffic – though BTIG flags the open question of how much of OTAs’ roughly 40% traditional paid-search traffic that figure could eventually erode.

This lands on top of a debate PhocusWire has tracked all year. OpenAI quietly deprioritised its Instant Checkout feature in ChatGPT in March, moving purchasing into third-party ChatGPT apps built by Expedia, Booking.com, Skyscanner, Accor and Lighthouse instead. Phocuswright’s research manager Coney Dongre made almost exactly BTIG’s point four months earlier, noting that even shifting a narrow piece of the transaction flow is difficult in travel, since a booking sits on pricing volatility, fare rules, ancillaries and post-booking servicing rather than a simple product line. Mizuho’s Lloyd Walmsley called OpenAI’s retreat a “Waterloo moment” for disruption fears and turned bullish on OTA stocks as a result; Trivago’s chief executive Johannes Thomas drew the same distinction BTIG lands on now – AI is good at discovery, but the complexity begins once travellers move closer to an actual decision.

For balance, a BCG report covered earlier this year by PhocusWire took a harder line, arguing OTAs may need to fully reinvent their model in the face of AI-native search, and that even reinvention might not prevent disintermediation – so this remains genuinely unsettled rather than a settled consensus. Phocuswright’s own February survey supports the adoption trend BTIG describes: 56% of US travellers had used AI for planning, booking or in-destination help in the past year, up from 43% in the second half of 2025. This pairs directly with Story 1’s connectivity research – the same underlying question of whether AI pushes hotels toward or away from the major platforms, argued here from the investment-analyst side rather than the vendor side.