About this report:
Field report from DMS 2026 (Digital Marketing Summit), held March 24, 2026 in Seoul. The session was presented by Kim Dong-wook, SVP of Chai Communication — the agency behind Gmarket's iconic campaigns and author of three books on strategic planning.
1. The irony that opened the talk
Kim Dong-wook opened DMS 2026 with a confession: he used AI extensively to prepare this very lecture — a lecture about doing better planning yourself. It was a deliberate frame. He wasn’t there to present AI as a shortcut. He was there to argue that the right use of AI in marketing is as a tool that makes you think sharper, not one that thinks for you.
2. Three chronic failures in marketing planning
Before getting to AI, Kim diagnosed the structural problems that repeatedly undermine marketing planning — patterns that extend well beyond Korea.
2-1. Target definition chaos
The brief says one thing, but every team member carries a different picture of the actual target. Decisions get made on misaligned foundations.
2-2. The KPI-strategy disconnect
KPIs are set. Strategies are built. The two rarely connect. Most KPIs in Korean marketing teams, Kim observed, are designed to be reachable — not directional.
2-3. Channel fragmentation without a center
With hundreds of media touchpoints, campaigns spread thin. Each channel runs on its own logic. Kim described the result as “a strange fish assembled from individually beautiful parts. Nobody wants to buy it.”
3. The MIT research that reframes AI
Kim cited a MIT Sloan field experiment comparing generative AI users vs. non-AI users on marketing tasks. The results cut against the standard productivity narrative.
- High performers with AI ~0% : No meaningful performance gain
- Low performers / new hires with AI +35% : Significant performance improvement
AI doesn’t make star performers brighter. It elevates the organizational floor. He paired this with a Go analogy: after AlphaGo, top players don’t copy the AI’s highest-probability move — they select within the viable option space based on their own judgment. AI sets the coordinates. The human makes the move.
4. AI Set: Chai Communication’s planning tool
Kim introduced AI Set, a proprietary AI planning workflow built internally at Chai Communication — originally to help junior planners produce higher-quality work. It operates across four functions.
(Function 1) Planning coordinate generator
Upload an RFP or brief. AI Set returns core message candidates, target segment analysis, brand identity diagnosis, and media mapping — a shared reference the whole team reviews and challenges together, replacing assumption-heavy briefs.
(Function 2) A/B/C multi-hypothesis strategy options
Three distinct strategic directions, each with rationale, expected outcomes, risk factors, and validation methods. Teams stop debating preferences and start evaluating hypotheses — what Kim called shifting from “taste battles” to “risk-weighted decision meetings.”
(Function 3) KPI-strategy linkage engine
Given target values from the brief, AI Set recommends KPIs calibrated to the strategic direction, builds the logical argument connecting strategy to metric, and maps channel-by-channel milestones with measurement risks.
(Function 4) Concept-to-creative continuity
Confirmed concepts generate video scenario scripts, scene images, compiled video (with BGM), and in-line image edits via prompt. Seeing a near-final creative before production briefing changes the quality of team alignment and client conversations.
5. The closing argument
Kim closed with a provocation: even if AI could think better than humans in every way, human thinking must still be protected. The ability to make meaning, exercise judgment, and take moral responsibility belongs to people. And the desire to grow and build mastery — that drive is irreducibly human. Organizations that outsource thinking to AI risk losing the very capability needed to use AI well.
“Do you want AI that thinks for you? Or AI that makes you think better than you ever have before?”
5. Key takeaways
- Shared context before more data.
The primary planning failure is misalignment, not lack of information. AI’s first job is creating a common frame. - Multiple hypotheses over single recommendations.
AI that generates structured alternatives with trade-offs builds team judgment; single-answer AI reduces it. - KPI logic, not just KPI numbers.
The causal argument between strategy and metric matters more than the target figure itself. - Creative preview as a planning tool.
Visualizing a concept before production briefing improves alignment and client communication. - AI lifts the floor, not the ceiling.
The biggest ROI comes from elevating underperformers — critical for fast-scaling retail teams with high junior headcount.
Field report based on direct attendance at DMS 2026, Seoul, March 24, 2026.
**Session: “AI, the Game Changer of Marketing Planning,” presented by Kim Dong-wook, SVP, Chai Communication.