Booking Holdings AI Strategy: Five AI Bets and Results in the P&L

Booking Holdings has made an AI-powered traveler experience its long-term strategy, and it is placing several bets at once. Its brands run AI assistants, a new AI travel agent called Lola sells trips by chat, and AI agents handle cancellations on Booking.com. Unusually, it points to a result in its accounts: customer service cost per booking fell 10% in 2025. Traffic from AI chatbots is still under 1% of its business.

AI is written into the company strategy. Booking’s annual report states: “We are executing against our long-term strategy to create an ideal AI-powered traveler experience” (2025 10-K). CEO Glenn Fogel said in November 2026: “As AI transforms the ways to discover, plan, and book travel, we continue to make progress integrating AI across our brands, developing AI-native experiences” (Q3 2026 release). The goal behind it is the Connected Trip: flights, stays, cars and activities planned and booked in one place (2025 10-K).

The bet is that trust beats the chatbot. Asked about AI travel agents from outside the industry, Fogel said building one “wouldn’t be that hard if that’s all you had to do.” His argument is that travelers handing over card details and travel plans want a company they trust, with people behind it (Skift, September 2026).

The focus has four parts.

• Travelers: AI assistants in each brand, including Priceline’s Penny and Agoda’s assistant in 21 markets, and two new standalone assistants: Lola, launched in September 2026, and Vayo, in private testing since October 2026 (CX Dive, November 2026)
• Customer service: AI agents on Booking.com that handle cancellations and customer questions (November 2026)
• Partners: AI tools such as Smart Messenger and Auto-Reply that help hotels answer guests faster (2025 10-K)
• AI platforms: an advertising deal with OpenAI and a hotel booking integration in Google’s AI Mode (Skift)

The brands own AI, not a central team. Booking Holdings has no group-level AI chief. Fogel says the company is running at least five AI efforts at once, across Booking.com, Priceline, Agoda, Lola and an unnamed project on the US West Coast (Skift). Lola is built by KAYAK co-founders Steve Hafner and Paul English on a separate technology stack, and Hafner reports to Fogel (Skift). At Booking.com, the largest brand, the technology chief has led the rollout of AI tools to employees (Fortune, April 2025).

A cost program funds the AI work. In November 2024, Booking announced a Transformation Program to cut annual run-rate expenses by $400 million to $450 million, about one-third through workforce reductions (TipRanks). It raised the target to about $650 million in August 2026 (Q2 2026 remarks). The annual report says the program creates “capacity to invest in long-term growth” (2025 10-K).

Engineers were first in line for AI.

• Adoption: Booking.com said in 2025 that “almost 100% of our developers” had adopted generative AI, and daily users shipped 16% more code changes than non-users (CO/AI, June 2025)
• Output: about 9,000 technologists now get roughly 30% more code into production, counting only changes that pass testing, according to CFO Ewout Steenbergen (Fortune, October 2026)
• The new bottleneck: Booking.com’s interim technology chief Vipul Hingne said: “Now the bottleneck becomes humans reviewing the code” (Skift, June 2026)

Other staff get tools, not a standard. Booking.com has rolled out the AI search tool Glean to about 14,000 employees (Glean), alongside AI features from Zoom and Google’s Gemini (Fortune). Booking has published no AI training figures or skill standard. Its CFO leads by example: he uses an AI coach and talks about it inside the company, saying: “If I have to learn, and I have a coach, it’s very normal” (Fortune). Leaders using AI in the open is where real AI ROI starts.

The workforce has stopped growing. Headcount was about 24,300 at the end of both 2024 and 2025 (2025 10-K). In July 2025, Booking.com said it would cut close to 1,000 jobs because “there are too many layers in the organisation” and “people are managing the work of other managers” (DutchNews), the pattern we describe in how AI is breaking middle management. Booking did not link the cuts to AI.

Governance is not published. Booking’s annual report names the EU AI Act among the rules it must follow and warns that generative AI “may result in reputational harm or legal liability” (2025 10-K), but Booking has published no AI principles or governance body.

AI fluency, coded against the Lead with AI Fluency Matrix (September 2026 edition):

• Fluency floor: Not published. Almost all developers use AI and about 14,000 Booking.com employees have AI tools, but Booking has set no standard and published no training figures.
• Highest level evidenced: Level 3 (Integrating), company-claimed. AI agents handle cancellations on Booking.com, AI writes a growing share of code that reaches production, and customer service cost per booking fell 10%. Booking has not published what share of requests AI resolves on its own, so Level 4 is not evidenced.
• Portfolio design: Brand-led and experimental. Each brand builds its own assistants, a separate team builds Lola, and engineers carry the internal gains, with cost targets for each.

Booking reports AI results in its accounts. All figures below are company-claimed:

• Customer service: generative AI helped cut customer service cost per booking by 10% in 2025 while bookings grew 10%. Steenbergen said: “I’m very proud that we can point to a line item in the P&L where we actually do have meaningful results” (PYMNTS, February 2026). In August 2026, Fogel said the cost per booking was still falling “at a double-digit rate” (Q2 2026 remarks)
• Engineering: about 30% more code changes reach production (October 2026)
• Travelers: early users of AI tools book somewhat faster, convert at higher rates and cancel less (October 2026)
• AI channels: traffic from AI chatbots is “significantly below 1%” of room nights (September 2026)

The cost of AI is managed too. Booking sends simple tasks to cheaper models and measures engineering groups on total IT cost per code change (Fortune). Steenbergen cautions that even the companies building AI models “don’t really know what is going to be the ROI” (October 2026).

Next markers: whether Lola and Vayo win paying members, the share of bookings that comes through AI chatbots, the $650 million savings target, and whether Booking publishes what share of service requests AI resolves on its own.

Change log

• November 2024: Booking announces its Transformation Program.
• April 2025: Booking.com shifts its AI focus to employee tools as well as travelers.
• June 2025: Booking.com reports that almost all developers use generative AI.
• July 2025: Booking.com plans close to 1,000 job cuts to remove management layers.
• February 2026: Booking reports a 10% fall in customer service cost per booking for 2025.
• August 2026: The savings target rises to about $650 million.
• September 2026: Lola launches.
• October 2026: Vayo enters private testing.
• November 2026: Booking.com’s AI agents handle cancellations and customer questions.

• Disclosure: 3 of 4. Booking answers where it is going, where it will focus and how it will know, with results tied to a line in its accounts. What it will take is answered in part: it funds AI through a cost program, but names no group AI owner and no training standard.
• AI is the long-term strategy, written into the annual report as an AI-powered traveler experience.
• Booking places several bets at once, with assistants in each brand and two new standalone ones, instead of one central AI program.
• Its finance chief reports AI results in the P&L, which few companies in the index do.
• The say-do gap is people outside engineering. Booking measures what AI does for developers and customer service, but has published no standard or training figures for the rest of its 24,300 people.
• Watch the AI channel. The test is whether travelers who start in a chatbot still book through Booking’s brands.

What the numbers show: Revenue rose 58% from 2022 to 2025 and headcount 12%, so revenue per employee rose 40%, to about $1.1 million. Headcount was flat in 2025, the year Booking.com said it would cut close to 1,000 jobs to remove management layers, while revenue grew 13%. These figures do not show that AI caused any change.

Revenue, headcount and revenue per employee, FY2022 to FY2025

• FY2022: $17.09bn revenue, 21,600 employees, $791k per employee
• FY2023: $21.36bn revenue, 23,600 employees, $905k per employee
• FY2024: $23.74bn revenue, 24,300 employees, $977k per employee
• FY2025: $26.92bn revenue, 24,300 employees, $1.11M per employee

Change, FY2022 to FY2025: revenue up 58%, headcount up 12%, revenue per employee up 40%.

Notes: Revenue is total revenues, as reported. Headcount is approximate, at year-end; about 97% are full-time. Booking.com said in July 2025 it would cut close to 1,000 jobs to reduce management layers. The Transformation Program Booking Holdings announced in November 2024 includes a reduction in workforce.

These figures give context on scale and productivity. They do not show that AI caused any change. Sources: Booking Holdings 10-K filings.

What is Booking Holdings’ AI strategy?

Booking’s long-term strategy is an AI-powered traveler experience, the Connected Trip, in which one service plans and books a whole trip. Each brand builds AI assistants, Booking launched two new standalone assistants in 2026, and AI agents handle service requests on Booking.com.

‍

Who leads AI at Booking Holdings?

Booking Holdings has no group-level AI chief. AI work sits with the brands, including Booking.com, Priceline and Agoda, and with Lola, led by KAYAK co-founder Steve Hafner, who reports to CEO Glenn Fogel.

‍

Has AI reduced Booking’s costs?

Booking says generative AI helped cut customer service cost per booking by 10% in 2025, while bookings grew 10%, and that the cost kept falling at a double-digit rate in 2026.

‍

What are Lola and Vayo?

Lola is an AI travel assistant launched by Booking Holdings in September 2026 that plans and books trips through a chat, with a paid membership option. Vayo is a second assistant for leisure travel planning that entered private testing in October 2026.

‍

Is Booking cutting jobs because of AI?

Booking.com said in July 2025 it would cut close to 1,000 jobs to remove management layers, and its Transformation Program includes a reduction in workforce. Booking has not linked the cuts to AI.

‍

How is this profile put together?

Lead with AI profiles use public sources only, such as annual reports, earnings materials, filings and on-record interviews, dated at the time of writing. AI fluency is coded against the Lead with AI Fluency Matrix. Company figures are marked as company-claimed unless they have been independently verified.