Tmap Mobility, which ended the taxi-hailing service it had operated jointly with Uber last year, is strengthening its artificial intelligence (AI)-based platform strategy. The plan is to evolve beyond a simple navigation app into an "integrated mobility agent" that combines data and AI.
On the 18th, at SK Telecom's pressroom in Jung-gu, Seoul, Tmap Mobility held a press briefing and presented a vision: "We will combine more than 7.4 billion travel data points with AI to advance into an integrated mobility agent that goes beyond navigation." As Tmap Mobility posted its first-ever quarterly profit on an EBITDA (earnings before interest, taxes, depreciation and amortization) basis in the second quarter, it plans to accelerate business expansion by adding AI to its data-centered revenue model.
The core point Tmap Mobility emphasized that day was its data competitiveness. Based on four pillars — ▲ map and road network data accumulated over the past 20 years ▲ more than 26 million travel routes collected daily ▲ lifestyle behavior data such as driving scores, insurance, parking, charging, and designated driver services ▲ location visit data for 30 million people secured through the "Where to go" service — the company laid out its foundation. Park Seo-ha, head of D&I (Data & Innovation) and vice president, said, "The success or failure of an AI agent ultimately depends on the volume and quality of data," adding, "Tmap holds a differentiated advantage in navigation and place data."
The most eye-catching case of AI integration is the "conversational navigation." With this update, the existing NUGU-based voice guidance is replaced by SK Telecom's AI service, adot. Thanks to this, drivers can request routes in natural, everyday language rather than fixed commands. For example, if you say, "Let's stop by a nearby gas station and go home," it recognizes both the destination and the waypoint and suggests the optimal route. Even if you say an incorrect place name, a correction function kicks in to guide you to the right candidates. It also supports themed searches like "a restaurant with a nice vibe" or "a place good to go with kids," personalized responses that reflect favorites and driving history, and summaries of reviews, business hours, and parking information.
On site, an "IU demo video" was shown, drawing attention. In a scenario where a "reporter has an interview scheduled with singer IU," when the driver said, "Find me a tea house near Cheongdam Station," adot recommended a destination and explained details including business hours and menu. Follow-up questions like "Is parking available?" and "How much is it?" continued the conversation. It even handled the request, "Send a text to IU's manager," showcasing multi-turn dialogue that had not been realized in the existing NUGU-based system. Jeon Chang-geun, head of product and vice president, said, "With this revamp, we have opened a conversational mobility AI experience that surpasses the usability of existing navigation."
Tmap Mobility also emphasized its differentiation from global general-purpose AI models. While it leverages multiple LLMs (large language models), it says it is difficult to apply general-purpose models like OpenAI's ChatGPT or Google's Gemini as is to the in-vehicle environment. Vice President Jeon said, "While driving, AI must minimize response latency, and general-purpose models are not optimized for navigation and place search," adding, "Tmap can deliver a differentiated experience with domain-specific data." Vice President Park also said, "Going forward, domain agents specialized for specific areas will see higher usage than general models," and explained, "Tmap, which has the most navigation and place data, will gain an edge in that market."
Tmap Mobility presented not only AI use cases but also plans to diversify its revenue models. In the B2C space, it secured 99% coverage through an insurance partnership based on driving scores and recently launched a driver accident insurance plan that can be purchased for 500 won per day. It plans to expand to products such as golf and travel insurance to offer a lineup that covers the entire car life. In the B2B space, it is using Tmap data for oil company sales forecasting, logistics and parcel ETA calculation, and festival visitor analysis, and it also disclosed a case that improved modeling accuracy for an oil company by 34%.
Alongside this, the company also unveiled a mid- to long-term roadmap to implement an "integrated mobility AI agent" spanning mobile, automobile, and chatbot by 2026. The goal is to provide a multimodal experience that continues context before and after driving by linking an in-vehicle conversational voice agent with a mobile chat-based agent. For example, when you arrive in Gangneung, it automatically recommends nearby restaurants, or it lets you continue reading in the chatbot the news you listened to while driving after you get out.
Vice President Park said, "Based on Korea's largest mobility data, Tmap will evolve into an agent that supports the entire spectrum of mobility life beyond navigation," adding, "We will deliver an AI experience that naturally permeates customers' daily lives."