MGS 2026 Recap: Live Intent Targeting and Brand Safety with Pyler

Are you still managing ad placement by adding more words to a keyword blocklist? With AI-generated video now arriving faster than any person can watch, that approach can't reliably tell the right placements from the risky ones.
Live intent targeting means serving ads based on what someone is watching right now, not on who they are. The daglo marketing team attended MGS 2026, and this post recaps "Surviving the AIGC Era with Live Intent," a session by Seeun Kim of Pyler. We transcribed the session with daglo and went back through the transcript to check the case study numbers.
Modern Growth Stack (MGS) 2026 is a growth marketing conference hosted by AB180, the company behind Airbridge. It ran July 21 to 24, 2026, at the Westin Seoul Parnas under the theme "Stack AI, Rewrite Everything." See the official MGS 2026 site.
What is live intent, and why is targeting changing in the AIGC era?

Live intent targets the intent revealed by the content a viewer is watching at this moment. As identity-based targeting keeps shrinking under regulation, the context of the moment becomes the stronger signal. That was the starting point of the talk.
The problem with AIGC (AI-generated content) isn't just volume. Videos built from synthetic AI voices, clickbait thumbnails, and context-free scripts are produced in bulk while we sleep.
Pyler uses multimodal "omni-understanding" AI to read contextual signals inside videos. It processes 2 million videos a day, more than 1 billion a year. That is 60 years of footage every day. The scale itself, the speaker said, is the reason this job belongs to AI.
Is this media moment we're about to target actually the right moment for our brand?
What to filter out: brand safety and the limits of keyword blocklists

Brand safety means keeping ads away from shocking or controversial content. According to the talk, the industry on average was serving 19.3% of its ads next to sensitive content.
With TV and out-of-home, channels and locations were limited, so advertisers could pick and review placements by hand. Digital is different: millions of videos are created every day and real-time bidding happens in under a millisecond, so there was no way to apply the same principle by hand.
Keyword blocklists, the common workaround, produce absurd mistakes.
- When Taylor Swift was named TIME Person of the Year, the interview was flagged as sensitive because it contained the word "war."
- An article about filming the sequel to The Devil Wears Prada was flagged because of the word "shoot."
- The keyword "Paris," added during the 2019 Notre-Dame fire, stayed on the list and ended up blocking coverage of the Paris Olympics.
The reason is simple, the speaker said. Teams hold meetings to add words to the list, but nobody ever schedules a meeting to take them off.
What to target: moments, not people

The same person watches content for hours a day, with a different intent at each moment. That is why "someone watching a video about muscle pain right now" is a stronger signal than "someone interested in health." This is the core idea of live intent.
The speaker also shared three signals that came up again and again at Cannes Lions.
- The AI conversation has shifted from "what should we create" to "how do we make better decisions." NVIDIA, Yahoo, Pinterest, and others unveiled agentic media infrastructure.
- Advertisers now choose partners on curation: can they combine AI intelligence with premium media signals?
- The best ads go beyond the topic of the content and tap into culture and the psychology and emotion of the moment.
Live intent targeting results in three case studies

The session covered three brands: Nongshim, Tylenol, and Samsung Electronics. One widened its interest targeting, one split audiences into moments, and one cut wasted placements.
| Brand | Approach | Result |
|---|---|---|
| Nongshim x KPop Demon Hunters collaboration | Targeted K-food, K-artist, and K-travel content in addition to fandom content | Core reach 216% of standard targeting, top 1% share of relevant placements |
| Tylenol | Targeted pain moments for family lovers, active seniors, and travel and leisure fans (wrist and back, knees, muscle pain) | Search volume up 29%, found a segment with 5.8x higher click engagement |
| Samsung Electronics | Applied brand safety filtering | Sensitive-content exposure cut from 10.93% to 2.55%, a 76.7% reduction |
Core reach measures whether ads actually reached the content you meant to target. What the speaker stressed about Samsung was that it didn't spend more: it freed up budget by running more efficiently within the targeting it already had.
Add a context layer to your growth stack: a practical checklist
Growth teams already manage attribution (what converted), audience data (who we reached), and creative (what we showed). The speaker proposed one more layer on top: context, meaning which content and which moment our brand appears in.
- Measure sensitive-placement exposure: Compare your campaigns with the 19.3% industry average to treat brand safety as a budget-waste metric.
- Prune your blocklist on a schedule: Check whether words that were added and never removed are blocking good placements.
- Redefine segments as moments: Try writing "people watching X right now" instead of "people interested in X."
- Find efficiency without more spend: Cutting wasted placements alone can free up effective budget.
- Report on exposure context: Alongside attribution, audience, and creative, review what context your ads actually ran in.
When you revisit a talk recording, knowing who said what matters. daglo's Speaker ID remembers voices and fills in real names instead of "Speaker 1" from the next recording on, so you can find a remark even when Q&A is mixed in. daglo is the first service in Korea to offer this to every user. Learn more about Speaker ID
For the rest of the sessions we attended, read our full MGS 2026 recap.
#MGS2026 #ModernGrowthStack #LiveIntent #BrandSafety #ContextualTargeting #AIGC #AdTargeting #daglo
Where talk becomes work.
Get started with 1,000 credits a month — no credit card required.