Airline & Distribution Leader Intelligence Brief — The Phocuswright Conference 2025 | ConferenceDigest
Airline & Distribution Leader Intelligence Brief
For: Airline executives, GDS leaders, NDC strategists, and distribution technology decision-makers
Airline & Distribution Intelligence Brief
The Phocuswright Conference 2025
Prepared for: Airline executives, GDS leaders, NDC strategists, and distribution technology decision-makers
Focus areas: Modernization, agentic AI distribution, offer/order transformation, competitive positioning
Conference: The Phocuswright Conference 2025, San Diego
Brief date: March 2026
Executive Summary
The Phocuswright Conference 2025 marked a decisive shift in how the travel industry frames the AI question — not "if" or "when" but "who controls the transaction." For airline and distribution executives, the conference delivered an unusually frank verdict: agentic AI has arrived as a live distribution channel, the IATA 2030 offer-and-order mandate will not be met on schedule, and NDC is simultaneously the prerequisite for airline modernization and the source of its most acute near-term operational pain.
The most consequential signal for airline distribution leaders came from Sabre CEO Kurt Ekert, who named agentic AI "a fundamentally new distribution channel, analogous to how OTAs emerged 30 years ago." This is not a product announcement — it is a structural diagnosis. Google confirmed it by announcing end-to-end agentic hotel booking partnerships mid-conference, with Booking.com and Expedia named as launch partners alongside IHG, Marriott, Wyndham, and Choice Hotels. Airlines were notably absent from that announcement. That absence is a strategic gap that distribution leaders should treat as urgent.
On the NDC and offer/order front, Ekert's acknowledgment that the IATA 2030 mandate is "unlikely to be met" should be read as a consensus view, not an outlier. Amadeus SVP Robert Buckman described offer and order transformation as "largely invisible" foundational work that determines whether personalization is possible at all — and confirmed most major carriers are still not on a committed order transition timeline. The industry is attempting open-heart surgery while simultaneously building the operating room. The airlines that move fastest on offer/order transformation will not merely comply with IATA standards; they will be the ones whose inventory is accessible to AI booking agents in 2026 and 2027, when the agentic channel transitions from pilot to mainstream.
The third dimension shaping this brief is the post-booking crisis that NDC has intensified. Acai Travel CEO Riccardo Vittoria quantified NDC's operational cost: call center handling times are up 20–30% specifically because of NDC complexity, while experienced human agents are retiring faster than they can be replaced. Wenrix/DeepFlow CEO Amir Balaish put a number on the efficiency gap — the industry has automated 40% of simple ticket servicing, but the remaining 60% of complex edge cases consumes 82% of agent time. These are costs that airlines bear directly and that their agency channel partners share. The near-term ROI case for post-booking AI automation is cleaner than any other AI investment category at the conference.
For GDS leaders specifically, the conference produced a composite picture of both threat and opportunity. Sabre, now 99% cloud-native, is moving to position itself as an intelligent caching layer for agentic AI — arguing that the look-to-book explosion driven by AI (which will generate logarithmically more search queries per transaction than NDC already has) creates a demand for petabyte-scale caching infrastructure that only a handful of players globally can provide. Amadeus is aligning with Anthropic on Model Context Protocol standards, describing it as a deliberate refusal to wall off its ecosystem. Neither GDS is treating agentic AI as someone else's problem.
Key Findings
Finding 1: Agentic AI Is Arriving as a Live Distribution Channel — Airlines Are Behind Hotels
Evidence: Google Travel Group Product Manager James Byers announced mid-conference that Google has enabled end-to-end hotel booking inside AI Mode, in partnership with Booking.com, Expedia, IHG, Marriott, Wyndham, and Choice Hotels. Airlines were not named. Separately, Booking.com CSO Rob Ransom disclosed a live ChatGPT app built via MCP-enabled queries allowing travelers to search inventory and complete transactions inside ChatGPT — again, accommodation-centric. ARC's Jennifer Watkins, speaking at the "AI Agents, Real Transactions" panel, noted that agentic commerce "was absent from airline distribution conference agendas in spring 2025 but dominated every panel by late 2025" — acknowledging the speed of the shift while also noting airlines' distribution channels are not yet ready to support agentic direct booking.
The gap: Watkins estimated that airlines are processing about 20% of transactions via NDC — a necessary precursor to the "orders" standard that would enable an AI agent to transact with an airline the way it can now transact with a hotel. She estimated orders could be supported by airlines in 2026. That "could" carries meaningful uncertainty.
Why it matters: Sabre's Ekert described the greatest distribution disruption as coming not from AI booking directly with suppliers, but from AI intermediating traffic that airlines currently count as "direct." If travelers shift their discovery to ChatGPT or Google AI Mode and those platforms route to hotel partners who have completed agentic integration, airlines risk ceding the booking relationship to a new generation of AI gatekeepers even for transactions that were previously direct. The window to establish airline presence on agentic platforms is compressing.
Finding 2: The IATA 2030 Offer/Order Mandate Is Unlikely to Be Met — And That Has Real Consequences
Evidence: Sabre CEO Kurt Ekert stated directly: "Offer-and-order delivery will take far longer than anticipated. The IATA 2030 order mandate is unlikely to be met on schedule." He described it as "open heart surgery for airlines," noting that most top 20–30 carriers have not committed to a formal order transition timeline. Amadeus's Robert Buckman described offer and order transformation as the "foundational building block for true personalization" — work that is "largely invisible at conferences dominated by distribution and travel-seller audiences" but which determines whether personalization is achievable at all.
The secondary cost: Ekert specifically flagged NDC's look-to-book problem as a distribution infrastructure crisis that agentic AI will dramatically worsen. AI agents will generate logarithmically more search queries per transaction than human users — preliminary NDC deployments already show significantly degraded look-to-book ratios compared to EDIFACT — and without intelligent caching at petabyte scale, the cost of serving those queries could render certain distribution architectures economically unviable. Ekert positioned Sabre's caching infrastructure as one of only a few globally capable of absorbing this load cost-effectively.
NDC agency adoption: Outside OTAs, NDC adoption among brick-and-mortar agencies and TMCs remains slow. The attendant workflow changes, back-office integration requirements, and agency retraining costs have proven to be organizational barriers as much as technical ones. Caroline Strachan (Festive Road) at the Business Travel panel noted mid-market TMCs have "cause for concern" specifically because these integration costs fall most heavily on smaller distributors with less technical capacity.
Finding 3: NDC Is Directly Increasing Post-Booking Costs — The Operational Crisis Is Quantifiable
Evidence: Acai Travel CEO Riccardo Vittoria, whose company won the Phocuswright People's Choice Award, opened by establishing that NDC complexity has pushed call center handling times up 20–30%. He quantified the total post-booking cost burden at $200 billion per year across travel companies. The workforce handling this complexity is simultaneously contracting: experienced agents are retiring faster than new ones are being trained, and offshore cost arbitrage is decreasing as a solution.
Wenrix/DeepFlow CEO Amir Balaish provided complementary data from the inside: the industry has automated roughly 40% of simple in-flight ticket servicing tasks, but the remaining 60% — the complex edge cases involving airline-specific policies, fragmented systems, and unstructured data — consumes 82% of agent time. He cited that DeepFlow achieved 93% automation with one major travel company — "more than double the stated industry average" — trained on over 50 billion real-world servicing data points across GDS, NDC, and LCC channels.
Amadeus's signal: Robert Buckman specifically named travel disruption management as "the next major frontier likely to be meaningfully solved — potentially within a few years," citing Acai Travel — a minority Amadeus investment — as the exemplar. That Amadeus is backing the startup solving NDC's operational mess is a pointed signal about where the industry's sharpest minds see the near-term ROI.
The airline cost implication: Airlines bear this cost in two directions: directly through their own contact centers and disruption management operations, and indirectly through agency and TMC channel partners whose servicing costs are passed through in distribution negotiations. Airlines that deploy AI post-booking automation reduce their direct costs and strengthen their agency relationships simultaneously.
Finding 4: GDS Modernization Is Genuine but the Narrative Is Contested
Evidence: Sabre CEO Kurt Ekert pushed back forcefully against calls to "blow up" the GDS model, arguing Sabre has already transformed from a mainframe-based system to a cloud-native, modular, AI-infused platform: "99% of Sabre's compute is now in the cloud." Sabre has achieved 15% developer productivity improvement using Google Gemini and Vertex AI. The company now carries more NDC connectivity than any competitor, alongside low-cost carrier content and AI-powered algorithmic shopping.
Ekert introduced Concierge IQ — an AI conversational commerce layer built on Sabre's AI IQ platform — now in full production with Virgin Australia, enabling research, search, booking, and post-trip support in a single conversational experience across WhatsApp, airline.com, or agency channels.
Amadeus's Buckman described the company's alignment with Anthropic on Model Context Protocol standards as "a deliberate signal that Amadeus is not walling off its ecosystem." He demonstrated a live internal MCP-based voice agent that changed a travel reservation in natural language while switching between English and French mid-conversation.
The contested claim: The Street Talk financial analysts were skeptical of the GDS modernization narrative's commercial implications. Lloyd Walmsley (Mizuho) noted that even with cloud-native infrastructure, if AI drives traffic back through Google — where OTAs pay roughly 80 cents per dollar of acquired traffic — distribution economics deteriorate independent of GDS platform quality. Emma Taylor (Barclays) focused on back-end AI ROI as the most measurable near-term value: call center automation, software development velocity, and advertising optimization, rather than front-of-funnel booking changes.
Finding 5: Agentic Payments Infrastructure Is Being Built Now — Airlines Must Engage
Evidence: The "AI Agents, Real Transactions" panel (Visa, Stripe, ARC) was the most technically detailed session on agentic commerce infrastructure. Clara Liang of Stripe described the Agent Commerce Protocol (ACP), co-developed with OpenAI, which defines how buyers, sellers, and AI platforms communicate programmatically at checkout. Key features: shared payment tokens that allow credentials to be used securely without exposing card data, spend limits, time limits, and revocation capabilities. The design explicitly preserves the original seller as merchant of record — maintaining pricing control, inventory management, fulfillment responsibility, and post-checkout upsell relationships.
Gloria Colgan of Visa described the Trusted Agent Protocol (TAP), announced as part of Visa Intelligent Commerce. TAP addresses authentication (verifying an agent is legitimate and was authorized by a real human, not a fraud bot), tokenization, and buyer intent metadata capture — ensuring what an agent purchases matches what the consumer actually wanted.
ARC's Jennifer Watkins made the airline-specific observation: "Airlines' primary concerns include disintermediation from the customer relationship, fraud exposure as payments teams navigate new flows, and readiness of third-party agency distribution channels."
The B2B first mover: Both Stripe and Visa identified B2B flows as the most immediate agentic commerce use case — high-volume, repetitive transactions such as OTA-to-hotel payouts or bulk flight bookings. Airlines should expect their OTA and TMC partners to be early adopters of ACP and TAP infrastructure, and should build toward compatibility with both standards now to avoid being the constraint in the agentic payment chain.
Finding 6: Personalization at Booking Time Is Architecturally Blocked — Post-Booking Is the Achievable Near-Term Opportunity
Evidence: Neil Geurin, VP Global Sales at American Airlines, offered one of the conference's most technically honest assessments of the personalization gap: "American Airlines has records on 40 million customers. The latency problem — parsing that data meaningfully within the sub-second window a consumer will tolerate before abandoning a search — is technically unsolved." His near-term solution is to concentrate personalization investment on the post-booking, pre-travel window, where milliseconds don't constrain.
Geurin's perspective was corroborated by Phocuswright research presented at the conference: 56% of leisure travelers take only one or two trips in 12 months, severely limiting the behavioral signal available to build personalization models. The data problem is structural, not merely technical.
The backend AI success case: Allianz Partners CEO Jeffrey Wright noted that 65% of Allianz's insurance claims now go through an AI model recommending a decision, with customers simply knowing they were paid in a day. This is precisely the model available to airlines in post-booking operations: AI acting invisibly in the backend, delivering customer-visible outcomes without requiring consumer-facing AI interactions.
The broader architecture implication: Amadeus's Buckman and Sabre's Ekert both identified offer and order transformation as the prerequisite for booking-time personalization. Until an airline's inventory, pricing, and ancillary logic are expressed in a modern offer/order architecture, there is no data pipeline capable of feeding real-time personalization at the booking interface regardless of AI capability. The sequence is inescapable: infrastructure first, personalization second.
Finding 7: Business Travel Distribution Is Restructuring Around AI Servicing
Evidence: The Business Travel panel (American Express GBT, Spotnana, Festive Road) delivered a consistent message: the corporate travel distribution stack is being reorganized around AI-native servicing. Steve Singh (Spotnana Executive Chairman) described three structural shifts within three to five years: search expanding to include conversational AI already accounting for a meaningful share of bookings; content distribution moving toward direct, personalized supplier-buyer relationships; and servicing evolving to "an agentic travel agent for life" — an AI that knows the traveler, their employer, and every city they visit.
Evan Konwiser (GBT Chief Product & Strategy Officer) reframed corporate travel policy as "the manifestation of the company's culture into what you can buy as a traveler," arguing AI-driven integrated solutions will surface the overlap between traveler preferences and corporate policy more effectively than current rule-based enforcement. He noted that consumer use of agentic AI for daily life is "organically training the entire ecosystem" — compressing corporate AI risk tolerance faster than any top-down mandate.
Caroline Strachan (Festive Road) introduced the "Disney effect" — corporate travel has two customers: the company that pays and the traveler who books. Airlines building NDC-based direct content distribution into corporate channels need to address both simultaneously. The GBT-Concur partnership was framed as a direct response to customer demand for integration — a signal that distribution deals at this scale are increasingly about workflow integration, not just content aggregation.
The embedded booking signal: Strachan specifically cited Amadeus's Microsoft Teams booking capability — "travelers want to interact from within Slack or Teams rather than switching to a separate booking tool" — as an example of the "be where I am" traveler principle. For airlines, this means NDC content must be accessible from wherever corporate travel decisions are made, not just from legacy OBT interfaces.
Finding 8: The Look-to-Book Crisis Will Accelerate — Caching Infrastructure Is a Strategic Asset
Evidence: Sabre's Ekert identified the look-to-book ratio explosion as one of the most underappreciated distribution challenges in the NDC-plus-agentic era. NDC already produces significantly more search queries per booking than EDIFACT-based distribution. Agentic AI, which runs multiple parallel searches to construct a comprehensive trip plan, will generate "logarithmically more search queries" than human users per transaction.
The implication: processing cost per booking under agentic AI could become punishing for airlines and their distribution partners without intelligent caching infrastructure. Ekert argued that petabyte-scale caching is "a critical capability — one only a few players in the world can provide," positioning Sabre as a cost-effective distribution layer that becomes more valuable, not less, under agentic load.
The Google framing: Google's James Byers described Google's competitive advantage in AI Mode as rooted not just in Gemini (the LLM) but in the grounding data behind it — live flight and hotel pricing feeds, Maps data, and the Knowledge Graph. He advised travel brands to maintain strong feed data (prices, availability) as a non-negotiable foundation. Airlines that maintain high-quality, low-latency pricing feeds will see those feeds serve as the data substrate for AI travel recommendations; airlines with degraded or intermittent feed quality will increasingly be invisible in agentic results.
Strategic Implications
The agentic distribution window is 12–24 months. Google AI Mode's hotel booking partnerships are live. Booking.com and Expedia are transactable inside ChatGPT. Airlines are not. The infrastructure gap (orders, MCP connectivity, agentic payments compatibility) is addressable in this timeframe for airlines that treat it as a priority. For airlines that don't, the risk is not that agentic AI fails to scale — it is that it scales without them, concentrating the agentic booking channel around hotel and vacation rental supply that has moved faster.
Offer/order transformation is no longer optional for distribution relevance. The IATA 2030 timeline being "unlikely to be met" does not reduce urgency — it reframes the competitive dynamic. Airlines that complete order transitions first will have a structural distribution advantage in the agentic channel, not merely a compliance posture. The airlines most likely to be accessible to AI agents in 2026 are those that have already committed to an order transition roadmap.
NDC's operational cost must be addressed alongside its distribution benefits. The 20–30% increase in call center handling time attributable to NDC complexity is a direct tax on NDC adoption — for airlines, agencies, and TMCs. Airlines that deploy AI post-booking automation reduce their own costs, reduce their channel partners' costs, and accelerate their partners' willingness to invest in NDC workflows. The causal chain runs in both directions: better post-booking AI reduces the resistance to NDC adoption.
GDS modernization is real but the strategic question has shifted. The relevant question for airline distribution leaders is no longer "are GDSs modernizing?" (they are) but "which distribution architecture is most compatible with agentic AI at scale?" Sabre's caching thesis and Amadeus's MCP alignment represent two different answers to that question — both legitimate, both worth engaging with directly rather than assuming legacy procurement relationships will hold.
Payments and authentication for agentic transactions must be designed now, not retrofitted. Stripe's ACP and Visa's TAP are open standards being defined in 2025–2026. Airlines that engage with these standards during the design phase — ensuring merchant-of-record preservation, pricing control, and fraud detection compatibility — will have materially better agentic transaction economics than those that retrofit existing payment flows to agentic requirements.
The corporate travel restructuring creates both a risk and an opportunity. Spotnana's Singh described content distribution moving toward direct, personalized supplier-buyer relationships within three to five years. Airlines with direct NDC connections to corporate buyers will be beneficiaries; airlines distributing corporate content solely through legacy OBT channels will be disintermediated. The GBT-Concur partnership is the largest corporate travel distribution event of the past several years and is explicitly about workflow integration — airlines need to understand which side of that integration architecture their content sits on.
Action Items
Priority 1 — Immediate (Q1–Q2 2026)
Commission an agentic distribution audit. Map which AI platform surfaces (Google AI Mode, ChatGPT, Perplexity) currently have access to your airline's inventory, pricing, and ancillary content, and in what form. Identify the specific technical gaps (MCP connectivity, offer/order readiness, content depth) preventing agentic bookability. Assign an owner with cross-functional authority across distribution, product, and technology.
Engage Visa TAP and Stripe ACP directly. Assign a payments technology representative to track the development of both agentic payment standards. The merchant-of-record preservation clause in ACP is directly relevant to airline commercial policy; participate in the standard's evolution rather than adapting to it after the fact. ARC is the appropriate domestic aviation liaison for this engagement.
Deploy AI post-booking automation. The ROI case for AI in disruption management, exchange processing, and refund automation is the cleanest at the conference — demonstrated 60–93% handling time reductions, sub-three-month sales cycles, and directly quantifiable cost savings. Given NDC has already increased handling times 20–30%, this is the highest-urgency near-term investment. Evaluate Acai Travel, Wenrix/DeepFlow, and comparable specialist platforms on the basis of NDC channel coverage and GDS/LCC interoperability.
Priority 2 — Near-term (Q2–Q3 2026)
Accelerate offer/order transition commitment. Even if full IATA 2030 compliance is not achievable, publishing a formal order transition roadmap signals to AI platform partners and distribution technology vendors that your inventory will be agentic-compatible within a defined timeframe. This is a commercial positioning decision as much as a technical one — DirectBooker's Sanjay Vakil signed six of the top ten hotel chains on the basis of their willingness to make content AI-accessible, not on offer/order completion.
Audit NDC content depth for AI platforms. Google's James Byers was explicit: "AI grounding requires live prices and inventory — LLMs without real-world data produce trip plans that don't hold up in reality." He advised brands to surface "niche, human, and factual content that can now match complex contextual queries — content that previously never surfaced in traditional keyword searches." Airlines should evaluate the granularity of their pricing feeds, schedule data, and ancillary content against what AI platforms need to generate accurate responses, and invest in content enrichment where gaps exist.
Assess look-to-book infrastructure scalability. Engage Sabre and Amadeus on agentic AI caching architecture specifically — not as a generic capacity discussion but as a scenario-planning exercise against the look-to-book multiplier that agentic AI is expected to introduce. Price the cost of absorbing agentic query volume under current architecture versus under an intelligent caching layer, and build that delta into distribution technology investment cases.
Priority 3 — Strategic (H2 2026 and beyond)
Build a direct AI platform distribution strategy. Booking.com's ChatGPT app, DirectBooker's hotel chain API connections to ChatGPT and Gemini, and Google AI Mode's booking partnerships represent three different models for distributing inventory into AI platforms. Airlines should develop an explicit point of view on which model(s) best preserve commercial relationships, merchant-of-record status, customer data, and loyalty crediting — and begin piloting at least one model with a single AI platform in 2026.
Integrate NDC corporate content into workforce-native interfaces. The "be where I am" principle from the Business Travel panel — embodied by Amadeus's Teams booking capability — applies directly to airline NDC content in corporate channels. Identify the top five corporate booking interfaces (OBTs, Teams/Slack integrations, TMC portals) where your NDC content is either absent or not fully functional, and prioritize integration with those tools as a corporate revenue recovery measure.
Sessions to Watch
1. "Competing in Travel's Next Era with Sabre's CEO Kurt Ekert"
The most operationally grounded session on GDS modernization, NDC realism, and the look-to-book problem under agentic AI. Ekert's acknowledgment that the IATA 2030 mandate will not be met is the most important single statement from a GDS CEO in several years. The Concierge IQ product launch with Virgin Australia is the first production-grade agentic distribution product from a major GDS. Essential for GDS strategy and NDC planning teams.
2. "AI Agents, Real Transactions — The New Travel Economy"
The definitive session on agentic payments infrastructure for 2025–2026. Clara Liang (Stripe/ACP) and Gloria Colgan (Visa/TAP) detail the open standards being built to govern agent-initiated transactions, with ARC's Jennifer Watkins providing the airline-specific lens on fraud, disintermediation, and channel readiness. Any airline or distribution technology leader making payments decisions in 2026 should treat this as required viewing.
3. "AI, Innovation and the Next Era of Distribution" (Amadeus fireside)
Robert Buckman's session is the most useful companion to the Sabre session — covering offer/order transformation as a prerequisite for personalization, Amadeus's MCP alignment with Anthropic, and a live demonstration of an agentic voice agent changing a reservation in two languages. The investment in Acai Travel and the description of disruption management as the "next major frontier" provide a clear view of where Amadeus is placing its near-term distribution bets.
4. "Acai Travel — People's Choice Award" (Innovation Launch)
Riccardo Vittoria's pitch is the single most directly actionable session for airline operations and distribution teams. The quantification of NDC's impact on call center handling times (20–30% increase), the $200 billion post-booking cost figure, and the three-agent AI architecture (supervisor, front-office, back-office) with demonstrated 60% handling time reductions provide both the business case and the product blueprint for airline post-booking AI investment. Amadeus's minority investment signals this is not a niche startup pitch.
5. "Wenrix/DeepFlow — Runner-Up Travel Innovator of the Year"
Amir Balaish's presentation complements Acai by going deeper on the execution layer — specifically the 82% agent-time consumption by complex edge cases, the 93% automation benchmark, and the technical architecture for GDS/NDC/LCC cross-channel policy automation. The eight-year data foundation (50 billion real-world servicing data points) and one-month integration timeline are the key differentiators for airline and TMC evaluation.
6. "Business Travel's Next Move"
The GBT/Spotnana/Festive Road panel surfaces the three structural shifts Singh predicts in corporate distribution within three to five years, plus the operational reality of the GBT-Concur partnership for mid-market TMCs. Essential for airline corporate sales and NDC direct-connect teams assessing which corporate distribution channels are worth investing in over the 2026–2028 horizon.
7. "Travel Search in the AI Era According to Google"
James Byers' session is the definitive briefing on Google AI Mode's architecture, the agentic hotel booking partnership announcement, and Google's specific advice to travel brands on feed quality and content strategy. The absence of airlines from the agentic booking partnership list is the most important thing that was not said in this session. Distribution and e-commerce teams should watch this to understand the technical and commercial requirements for airline inclusion.
8. "From Discovery to Departure — Designing Effortless Experiences" (American Airlines/Allianz/Sutherland)
Neil Geurin's session is the most candid airline-perspective statement on personalization's actual architecture constraints: 40 million customer records, a sub-second latency barrier that is technically unsolved, and a strategic pivot to post-booking as the achievable near-term opportunity. The 65% AI claims adjudication figure from Allianz Partners provides a benchmark for what backend AI looks like when it is working at scale. Valuable for airline product, technology, and loyalty strategy teams.
*This brief was prepared from session transcripts and summaries from The Phocuswright Conference 2025. All data points, statistics, and quotations are sourced from the conference corpus. Session citations refer to Phocuswright 2025 programming.*