Jet2 uses AI to learn why customers call and signs an AI deal with Adobe
Jet2 runs a language model, trained further on its own data, on Snowflake to analyse the intent and sentiment of about 2,000 contact centre calls a day, by Snowflake's account. In September 2026 it signed an agentic AI partnership lasting several years with Adobe to personalise holidays for its 10 million myJet2 customers. No launch dates have been given for the Adobe work.
The operating picture
Listening to the contact centre first
G-GDFF Jet2 Boeing 737-800 2. Photograph: ahisgett / CC BY 4.0 (cropped to fit). Illustrates the airline, not an AI deployment.
Jet2 is both a UK airline and the UK's largest tour operator, flying from 14 UK airports. Its customers call before they travel about hotels and tours as well as flights.
Its published AI work starts with those calls and moves on to marketing and personalisation through the Adobe agreement.
Selected implementations
The work behind the evidence
01 / Deployed
Why customers call
The operational problem
A contact centre taking thousands of calls a day, some lasting 40 minutes, cannot listen to them all, so the reasons customers call and how they feel stay largely unknown.
What the evidence reports
Snowflake says in an undated case study that Jet2 uses a Mistral model, adapted to its own calls, in Snowflake Cortex AI to analyse customer intent and sentiment in call transcripts. It covers about 2,000 calls a day, 25% of calls to the line customers use before they travel, and went from no language model experience to production in three months. Jet2 combines the model with manual classification to control cost, and tracks the model's accuracy over time. Mark Atkinson, head of data science, said that without Cortex, Jet2 could not have used a language model "without the costs outweighing the benefits". Jet2 has identified calls about basic hotel and tour information that could move to an online portal where customers find answers themselves.
Operational significance
Knowing why people call tells Jet2 which questions could be answered online, which could reduce the load on the contact centre.
Evidence still needed
The account is Snowflake's and undated. Jet2 has not published accuracy figures or said how many calls have moved online as a result.
A source can verify an announcement without verifying its projected outcome.
02 / Reported · unverified
Personalised holidays with Adobe
The operational problem
A holiday company with 10 million account holders wants offers and messages that match each customer at each stage of a holiday.
What the evidence reports
Adobe said on 22 September 2026 that Jet2 will use its agentic AI tools to help staff deliver personalised experiences, a live customer data platform for a single view of each myJet2 customer, and Journey Optimizer for offers and messages at key moments. Firefly Foundry will build custom AI models on Jet2's brand assets. A joint Customer Experience Lab will build and test services such as better matching of customers to holidays. David Hills, chief customer officer, said AI would make "every interaction more personal, more relevant, more timely and always helpful".
Operational significance
The deal moves Jet2's AI from understanding customers to acting on what it knows about them.
Evidence still needed
Adobe gave no launch dates or milestones. Nothing in the announcement is described as live.
A source can verify an announcement without verifying its projected outcome.
The wider register
Other tracked activity
Other records are shown separately, including non-AI technology and programmes at different stages of deployment. A register entry does not establish a measured AI benefit.
No other publicly listed records.
Editorial assessment
Jet2's AI in production is an analysis tool for its contact centre, and the Adobe agreement is its larger bet.
Whether the personalisation reaches myJet2 customers, and what it does to bookings, will show if the second step pays off.
Live data from the Aerostack register. Status reflects published evidence only. See Our approach for how records are graded.
Aerostack register · 2 tracked deployments
Aerostack tracks AI deployments across airports and airlines, with the published evidence behind each record. Photography is real aviation imagery. Record thumbnails may show the named airport or airline; other photographs do not.
