Lotte Duty Free has moved conversational AI from a service experiment into the product-discovery layer of travel retail. Its new ChatGPT-based shopping service lets travellers describe what they need, compare options and reach the retailer's shopping platform through a conversation rather than beginning with menus and category filters.
- Lotte describes the launch as the first generative-AI shopping service of its kind in Korean travel retail.
- The commercial shift is from catalogue navigation to intent-led discovery: gift, budget, destination, product type or occasion.
- The opportunity is meaningful only when the assistant is connected to accurate product data, live availability, commercial rules and credible impact measurement.
What Lotte Duty Free has launched
On 16 July 2026, Lotte Duty Free announced a shopping service integrated with ChatGPT. According to The Moodie Davitt Report and Korean business press, customers can find the Lotte Duty Free service inside ChatGPT, start a conversation and ask for products, bestsellers or promotions. The experience can direct the user to Lotte's online shopping environment to continue the purchase.
The distinction matters. This is not simply a customer-service bot placed on an existing website. The conversation can become the front door to the catalogue. A traveller might ask for a fragrance gift below a certain budget, a Korean beauty set for a long-haul trip or a whisky with a specific flavour profile. The system can turn an open-ended request into a smaller set of options.
Why conversational commerce matters now
Search boxes are poor at understanding a travel mission
Traditional e-commerce assumes the customer knows the category, brand or product name. Travel shopping is often less precise. The customer may know only that they need a gift, something suitable for a humid destination or a compact product that can be collected before departure.
Conversation is well suited to these incomplete requests. It can ask a follow-up question, explain trade-offs and combine several constraints. OpenAI's own product-discovery work describes a model in which richer merchant feeds provide up-to-date attributes, price and availability inside ChatGPT. The commercial value therefore depends on structured data rather than fluent text alone.
Discovery can begin outside the retailer's app
A retailer normally spends heavily to attract customers to its site, app or store. A service inside ChatGPT changes that acquisition path. The customer can begin in a general AI environment and reach the retailer only after the need has been clarified.
Travel retail has unusually complex fulfilment rules
Duty-free shopping is not ordinary e-commerce. Eligibility can depend on an international journey. The order may require passport and flight details, and the traveller must collect it at a defined point before boarding. Products can also be affected by liquid restrictions, destination allowances and transfer rules.
An assistant that recommends well but ignores collection time, stock location or regulatory limits creates frustration rather than convenience. The strongest systems will combine conversational simplicity with operational precision.
Catalogue
Clear attributes, claims, images and product taxonomy.
Inventory
Current price, stock, fulfilment point and cutoff time.
Context
Flight, destination, mission, language and eligibility.
Rules
Compliance, restrictions, disclosure and escalation.
Measurement
Incremental sales, margin, pickup and unsupported answers.
The conversation is only the visible layer. Product data, inventory, journey context, rules and measurement determine whether the experience is commercially reliable.
What this means for travel retail and tourism
Product discovery becomes more personalised and more competitive. A physical shelf gives visibility to every product placed on it. A conversational answer may present only a few options. Recommendation logic therefore becomes a new form of merchandising.
Where the commercial opportunities appear
Open a new front door
Use conversation to reduce catalogue friction, qualify intent and guide customers towards products that can actually be fulfilled for their journey.
Make products understandable to AI
Improve attributes, claims, usage occasions, pack size, destination relevance and comparison data so recommendations are accurate.
Connect digital discovery with pickup
Integrate terminal, collection point, operating hours, security time and walking distance into the commercial journey.
Add journey context
Use consented flight and loyalty data to support relevant recommendations during booking, check-in and disruption.
Build trip-led bundles
Connect local products, weather, events, gifting and in-destination services with the passenger's stated purpose.
Build the decision layer
Connect product feeds, inventory, identity, translation, rules, analytics and handoff to the transaction platform.
Move from impressions to assisted discovery
Test sponsored or promoted visibility with clear disclosure and measure whether it adds incremental consideration and sales.
Can your product and inventory data support a trustworthy recommendation in real time?
Risks and practical barriers
- Incorrect or outdated recommendations. A fluent answer can still contain the wrong price, claim, pack size or availability. Product feeds need ownership, validation and refresh rules.
- Opaque visibility. Brands and regulators will need clarity on why one item is recommended, whether promotion influenced the answer and how commercial relationships are disclosed.
- Inventory and fulfilment failure. The assistant must consider collection airport, departure time and stock at the correct fulfilment point.
- Privacy and consent. Flight, loyalty and purchase data can improve relevance, but organisations need a lawful purpose, minimisation and clear customer control.
How Marksyte can help
Marksyte can help retailers, airports and brands turn a conversational interface into a measurable commercial system.
Product data readiness
Audit attributes, taxonomy, claims, images, pricing and availability to identify where the assistant lacks reliable information.
Traveller-intent analysis
Group real questions into missions such as gift, self-use, destination need, last-minute problem or category exploration.
Recommendation design
Build rules and models that balance relevance, availability, margin, diversity, compliance and customer value.
Demand and inventory forecasting
Use conversational signals alongside flights and sales to forecast demand by product, pickup point and time window.
Retail media measurement
Separate organic recommendation, paid visibility and promotional effects, then measure conversion and incremental margin.
Quality monitoring
Detect incorrect answers, dead links, stock mismatches, unsupported claims and questions the system cannot resolve.
The aim is not to automate every decision. It is to make discovery faster while preserving commercial control, transparency and a clear path to human support.
A practical 90-day agenda
- Collect the real questions. Analyse search terms, customer-service contacts and store enquiries to identify the most valuable conversational missions.
- Choose a controlled category. Start with products that have clear attributes, reliable stock and manageable regulatory complexity.
- Connect fulfilment before launch. Validate price, availability, departure eligibility, collection location and cutoff time in every recommendation.
Frequently asked questions
Does the Lotte service complete the whole purchase inside ChatGPT?
The launch information describes product discovery and access to Lotte Duty Free through ChatGPT, with links into the retailer's shopping platform.
Will conversational AI replace airport stores or sales staff?
It is more likely to change discovery and preparation. Stores and staff remain important for experience, fulfilment and reassurance.
How should an AI shopping assistant be measured?
Measure successful discovery, click-through, product availability, conversion, basket value, margin, cancellations, pickup completion, satisfaction and unsupported answers.
Sources
- The Moodie Davitt Report, Lotte Duty Free launches ChatGPT-based shopping service, 16 July 2026.
- Asia Business Daily, Lotte Duty Free launches ChatGPT Shopping, 16 July 2026.
- Lotte Group, Lotte Duty Free corporate profile and multilingual digital channels.
- Lotte Duty Free, same-day airport pickup information.
- OpenAI, Powering Product Discovery in ChatGPT, March 2026.