South Korean travel data company DealLabs has built a database of more than 280,000 verified local travel records covering seven Asian cities, aiming to improve the accuracy of information delivered by AI-powered travel services. The database covers Seoul, Busan, Tokyo, Osaka, Bangkok, Taipei and Singapore, with each record accompanied by its source and verification date. DealLabs says its data powers more than 23,000 guide pages and is available through APIs and AI-readable files as demand for its information from generative AI services grows.
DealLabs Builds Verified Travel Data Infrastructure for AI Search
AI-powered travel assistants can help travelers plan trips, find restaurants and navigate unfamiliar cities. But their answers are only as useful as the information behind them. Outdated business hours, incorrect train fares and inaccurate details about local attractions can turn a convenient recommendation into a practical problem.
DealLabs is building a data infrastructure layer intended to address those gaps. The Seoul-based company has assembled more than 280,000 local travel records across seven Asian cities, with each entry accompanied by a source and verification date. The database supports more than 23,000 guide pages and is published in formats designed for direct use by AI systems.
The company reported that generative AI search and answer services accessed its data more than 11,000 times per day in mid-October, based on a seven-day average. In late September, the comparable figure was approximately 600 daily accesses, representing nearly 20-fold growth in less than a month, according to DealLabs.
A Data Layer for the Details Travelers Need
DealLabs focuses on practical information that travelers need after arriving at a destination. Its database includes subway station exits, train timetables, first and last train times, fares, restaurants, tourist attractions, hospitals and local business schedules.
The company also collects city-specific information, such as closing days for Singapore’s hawker centers and business hours for Taipei’s night markets.
The scale varies by destination. Osaka accounts for more than 114,000 records, including over 42,000 medical institutions and pharmacies and more than 30,000 restaurant permits. Taipei’s collection contains more than 63,000 records, including 7,200 station-to-station fares.
This focus addresses a particular weakness of AI-generated travel answers: information can change faster than general-purpose models or static travel guides reflect those changes. Opening hours, temporary events, transit schedules and local operating rules may require current source material rather than information learned during model training.
DealLabs says it continuously updates records using official sources. It cites discrepancies in the announced end dates of three Seoul pop-up stores among 25 listings as an example of the problem, with differences of up to 10 days between public event listings and organizers’ notices.
The example illustrates why even apparently straightforward travel facts may require source checking. However, the company has not provided independent accuracy benchmarks showing how often its database resolves these discrepancies or how frequently records become outdated.
Making Local Information Accessible to AI Systems
DealLabs has created AI-readable llms.txt files and data APIs for all seven cities. Seoul also has a standard OpenAPI specification, which can help developers understand how to interact with the API.
The company says all seven destinations run on a common underlying engine. Adding another city therefore involves incorporating its local data and sources rather than building a separate system from scratch.
That architecture could make geographic expansion more manageable, although the quality and availability of official sources will vary across locations. Adding a city requires more than replicating software: the underlying records must be collected, verified, updated and maintained.
The company’s approach also reflects a broader development in AI search. As answer engines become more prominent in travel discovery, structured information sources can offer a way to ground answers in identifiable records rather than relying entirely on generated text.
For travel platforms and developers, such data could support itinerary planning, local recommendations and navigation-related queries. Its usefulness will depend on whether AI systems can retrieve the relevant records at the right time and preserve the source information when generating answers.
Testing AI-Generated Itineraries Against Official Sources
DealLabs is also piloting a service called Visit Plan Check, which checks AI-generated travel itineraries against official sources daily.
The concept targets a different problem from building a travel database alone. An itinerary can contain individually plausible recommendations while still including a closed attraction, an impractical transit connection or a visit scheduled outside operating hours.
Checking plans against current source material could help identify those inconsistencies before travelers rely on the itinerary. The announcement does not detail the service’s supported AI platforms, its correction process or measured performance in detecting itinerary errors.
Those details will matter as the service develops, particularly for itineraries that combine multiple destinations, transit connections and time-sensitive activities.
Expansion Plans Extend Beyond Asia
DealLabs plans to expand its coverage to 10 cities by the end of 2026 and 30 cities by 2028, including New York and Paris.
The expansion will test whether its shared data infrastructure can accommodate cities with different languages, public information systems, transit networks and business practices. Maintaining verification dates and reliable sources across a much larger collection could become as important as adding records.
For now, DealLabs is positioning verified local data as a supporting layer for AI-driven travel services. As travelers increasingly use AI to answer practical questions during their trips, the challenge is not simply generating a helpful response. It is ensuring that the underlying facts remain accurate for the place and time in question.
Market Landscape
AI-powered travel services depend on several interconnected capabilities:
- Structured local data: Organized records covering transport, attractions, businesses and essential services.
- Source verification: Tracking where information originated and when it was checked.
- AI-readable infrastructure: APIs and structured files that allow software systems to retrieve relevant information.
- Real-time relevance: Updating information that changes frequently, including operating hours, fares and event dates.
- Itinerary validation: Checking generated plans against authoritative information before users act on them.
DealLabs is targeting the data and verification layer rather than building only a consumer-facing travel assistant. Its ability to maintain accurate information across more cities will be central to the model’s scalability.
Top Insights
- 280,000-plus records: DealLabs’ verified travel database covers seven Asian cities.
- More than 23,000 guide pages: The data supports a substantial collection of destination-specific content.
- Nearly 20-fold growth: The company reports daily accesses from generative AI search and answer services rose from around 600 to more than 11,000 between late September and mid-October.
- AI-friendly access: All seven cities have
llms.txtfiles and data APIs, while Seoul also offers an OpenAPI specification. - Expansion plans: DealLabs aims to cover 10 cities by the end of 2026 and 30 by 2028, including New York and Paris.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
