[Edaily Reporter Yun Junghoon ] “I only buy Coca-Cola when a 30-pack costs 13,000 won. If they advertise it at that price, I’d actually appreciate it.”

Song Hoon-hwan, head of the AI Marketing Team at SKTelecom(017670), made these remarks on the 15th at the “E-Marketing Insight Summit 2026” held at the FKI Tower in Yeouido, Seoul. Team Leader Song presented AI marketing strategies to boost ad efficiency and conversion rates under the theme “Hyper-Personalized Marketing Powered by Telecommunications Data.”

“Customers leave traces of their interests at every moment,” he explained. “Using real-time location, call, and web/app behavior data, we can precisely reach not only our own customers but also competitors’ customers and interest-based audiences.”

Team Leader Song analyzed recent consumer behavior patterns of ad rejection and emphasized the importance of personalized advertising.

He stressed, “A message that has nothing to do with me is essentially spam,” adding, “The problem isn’t targeting itself, but targeting that lacks context.” He explained that once a customer receives an irrelevant message, they will opt out, and the brand will lose even the opportunity to reconnect with them in the future.

He proposed personalization as the solution. According to surveys by McKinsey and others, companies that adopted AI-based personalized marketing saw their marketing return on investment (ROI) rise by an average of 25% and revenue increase by 20%. Customer engagement rates doubled, and marketing spend efficiency improved by 10–30%.

3.3 Petabytes of Data Accumulated Daily... Customer Segmentation Through Lifecycle Analysis

SKT’s key asset is its vast telecommunications-based data. As of December 2025, SKT holds the top spot in South Korea with a 38.78% share of mobile lines and serves approximately 23 million customers. The data accumulated each day alone amounts to 3.3 petabytes (PB). In 2021, the company participated in the first Experimental Statistics project (National Population Mobility Statistics) organized by Statistics Korea (National Data Agency), thereby validating its reliability.

Team Leader Song explained, “SKT segments customers based on their life cycle, lifestyle, and behavioral characteristics, and lays the foundation for customer-segment-based marketing strategies through precise profiling.” The company gains a multidimensional understanding of customers by using AI to analyze location data collected 24 hours a day, 365 days a year; actual places of residence and work; access logs from over 4,000 apps and websites; and call and text records from over 2.8 million main phone numbers.

The company also presented examples of creating specific personas. Whereas in the past, the customer group was simply defined as people in their 20s and 30s preparing for marriage, the company now targets “trend-conscious, AI- and interior design-loving prospective newlyweds in their 20s and 30s.”

By identifying customers who fall in the top 20% for fast fashion and AI service usage, the top 10% for interest in real estate and interior design, and the top 5% for searches related to wedding venues, wedding photography, and wedding planning services—while ranking in the bottom 5% for interest in elementary and secondary education—the customer who previously existed only in one’s imagination becomes a tangible reality based on data.

The key is the “dynamic profile,” or “living customer persona.” Whereas traditional customer relationship management (CRM) focused solely on who the customer is—such as “a 35-year-old woman living in Seoul with a family of three”—the dynamic profile tracks how customers are evolving in real time by analyzing changes in app usage habits and mobility patterns.

It automatically generates segments based on large language models (LLMs) and updates them monthly. For example, instead of sending blanket advertisements to “women in their 40s,” it targets “customers preparing for overseas travel” the very moment they are about to go on their trip.

The targets created in this way are connected to various channels, including Google, Meta, and Naver ads, as well as text messages, app push notifications, and Kakao Talk notifications. SKT also unveiled its Agent AI marketing agent, “T-Marketer.” It automates budget allocation across media channels, ad creative production, and 24-hour performance monitoring, significantly reducing repetitive tasks.

Data-driven survey services are also yielding positive results. Various clients, including CJ Olive Young and Garmin Watch, have selected precise customer data to increase their click-through rates (CTR). The time required to obtain this data has been drastically reduced to just one hour compared to traditional surveys.

Team Leader Song stated, “SKT uses AI modeling based on telecommunications data to meticulously identify customer characteristics and interests, and continuously optimizes these based on actual response data from ads, messaging, and surveys.” He added, “Going forward, we will move beyond the realm of wireless communications to precisely match customer databases with various marketing channels, and continue to expand our marketing role in a broader sense.”