42 humans. 1,200 agent-hours a day. One shared inbox that no one reads unless the agent flags it.
Revenue per employee (2026 run rate)
€357k
Booking.com employs approximately 27,000 people. At comparable revenue, that's roughly €14k revenue per employee. Landr runs at 25× that ratio — not because we work harder, but because the agents absorb the work that used to require headcount.
Each agent owns one job. None are bolted-on features. They were in the product from day one — because the work never needed a human, only a judgment call on when to surface one.
Monitors flight status and pre-deploys eSIM, cab booking, and hotel instructions 90 minutes before landing. Watches for gate changes and delays in real time.
Captures every transaction in real time, converts currencies, and pushes a clean entry to Concur or the corporate expense tool before the traveller exits the terminal.
Detects cancellations and rebooking windows before the airline's own app updates. Acts without being asked — new boarding pass issued, hotel notified, cab window moved.
Builds a live preference profile across every trip — seat choices, hotel requirements, dietary needs — and applies it silently on every subsequent booking without prompting.
Reads the traveller's work calendar, detects an upcoming trip, and starts building the itinerary before they ask — often before they've thought to search themselves.
Surfaces curated restaurants and local activities near the hotel, matched to the traveller's taste profile — not ranked by commission, but by demonstrated preference.
€47 per completed trip. €0 on failure. No commission on lookups. No seats sold, no hotels pushed on margin. The incentive is to deliver the arrival — nothing else.
| Trigger event | Charge | On failure |
|---|---|---|
| Full trip delivered (eSIM + cab + expenses) | €47 | €0. Waived automatically. |
| Partial delivery (2 of 3 components) | €20 | Failed component refunded, incident logged. |
| Corporate account (annual) | €300k/yr avg | SLA credit applied. Account reviewed by Field Ops. |
| Inventory lookup / search only | €0 | — |
Revenue / trip
€47
avg blended, incl. corporate
Gross margin
74%
no inventory, no content ops
Trips closed without human touch
89%
target: maintain above 85%
CAC payback
4.2mo
individual travellers
89% of trips close on their own.
The 11% that require human involvement are the genuinely ambiguous cases — a traveller with a visa issue, a carrier that won't accept automated rebooking, a corporate policy edge case. Those go to a human within 4 minutes. Everything else, the agents handle end-to-end.
Every function below exists at our competitors. None of them exist at Landr — not because we're small, but because the work was never best done by a human in the first place.
What we skipped
Content operations team
Typically 80–200 people writing hotel descriptions, resizing photos, maintaining inventory accuracy.
What replaced it
Inventory is agent-structured from the moment it's ingested. The arrival agent reads machine-readable data directly — no human ever reformatted it for a listing page, because there is no listing page.
What we skipped
SEO & affiliate distribution team
Typically 60–150 people managing search rankings, affiliate partnerships, and click-based acquisition.
What replaced it
Distribution comes from the inventory being directly queryable by Claude, Gemini, and Copilot assistants. No SERP. No click. When a traveller asks their AI assistant to sort out a Hong Kong trip, Landr's structured inventory is what surfaces — not a search result.
What we skipped
Category management team
"Best hotels in Tokyo" curation. Editorial teams ranking, weighting, and updating destination content.
What replaced it
The Discovery agent rescores local inventory in real time based on each traveller's individual taste profile. There is no "Best hotels in Tokyo" — there's a list that changes every time based on who's asking. No human decides the ranking.
What we skipped
Corporate onboarding team
Permanent onboarding, support, and account management functions. Often 5–10 people per enterprise account in year one.
What replaced it
A 4-person Field Ops pair embeds with each new corporate client for 60 days. After that, the agents take the full handoff. No permanent onboarding team. The same 8 Field Ops people cycle across every new corporate account.
One real scenario, end to end. This is the full decision chain from signal to resolution — no human touched it until the morning review.
Flight delay signal detected
The Reroute agent picks up a 3-hour delay on CX234 Hong Kong → Amsterdam from the airline's API — 14 minutes before the airline's own app updates. It opens a rebooking window.
Reroute agent drafts 3 options
Reroute via Zurich (adds 2h), reroute via Frankfurt (adds 1.5h, business class upgrade available), accept the delay and keep original routing. The Pricing agent confirms availability and cost for each option.
Preference match: Frankfurt routing selected
The Memory agent confirms the traveller has a consistent preference for Frankfurt layovers and has accepted business upgrades within the corporate policy threshold of €180. Both conditions met. No human needed to approve.
Auto-execute: boarding pass issued
New boarding pass issued via airline API. Hotel departure time updated. Cab window shifted. Concur amended with the fare differential. Partner notified of adjusted arrival time. All automatic.
Traveller wakes up to a summary notification
"Your flight to AMS was rerouted via Frankfurt. New boarding pass attached. Cab adjusted. Nothing to action." The traveller never touched their phone during the disruption.