Latest / The Yield Doctor with James Deaker / RampUp 2026: Three Signals About Where AI in Advertising Is Really Heading

Transcript

RampUp 2026: Three Signals About Where AI in Advertising Is Really Heading · The Yield Doctor with James Deaker

• 0:00
  I just got back from the ramp up
• 0:03
  conference where AI was clearly the
• 0:05
  dominant topic in almost every session
• 0:09
  and every conversation.
• 0:11
  But what struck me the most wasn't the
• 0:14
  technology itself. It was how early the
• 0:17
  industry is in trying to figure out how
• 0:19
  AI will actually be implemented. So
• 0:22
  after two days at the conference, here
• 0:24
  are the three takeaways that stood out
• 0:26
  to me. The biggest thing is that the
• 0:30
  standards are missing from AI and
• 0:32
  advertising right now. Scott How the CEO
• 0:36
  of live ramp gave a keynote where he
• 0:38
  talked about various waves of technology
• 0:41
  adoption over time. And one of the
• 0:44
  things he focused on was the railway in
• 0:47
  the US in the 1800s and how the things
• 0:50
  that really made it succeed as a network
• 0:54
  were things like standardization of the
• 0:56
  time zones and standardization of the
• 0:59
  gauge of the railway tracks. Without
• 1:02
  these, the rail would not have become a
• 1:05
  real network and would not have
• 1:07
  succeeded in the way that it did. And as
• 1:09
  I sat through the sessions at ramp up, I
• 1:12
  was struck by how we don't have those
• 1:15
  standards right now in advertising
• 1:18
  technology for AI. The protocols are
• 1:22
  either still being developed or still at
• 1:25
  the proposal stage. And in some cases,
• 1:28
  there are even competing standards. For
• 1:30
  example, there's ADC CP, which is
• 1:33
  sponsored by companies like Yahoo, Scope
• 1:36
  3, and PubMatic. And there is the AMP
• 1:39
  protocol which is being sponsored by
• 1:41
  companies like Live Ramp and industry
• 1:44
  body the IAB. To state the obvious,
• 1:48
  Americans don't really like standards.
• 1:50
  We adopt them reluctantly. And the first
• 1:54
  version is probably going to be
• 1:55
  different than the standard from the
• 1:57
  rest of the world. And so my takeaway is
• 2:00
  that it isn't the speed of technology
• 2:03
  that is going to slow down the rise of
• 2:06
  AI agents. Instead, it will be the
• 2:09
  adoption of the protocols around AI. The
• 2:14
  second main thing that struck me is that
• 2:16
  if I compare this wave of innovation to
• 2:19
  others that I've seen in adtech in every
• 2:23
  other group or every other wave of
• 2:25
  innovation, there's been a group pushing
• 2:28
  back. Uh right, a good example is
• 2:32
  programmatic buying. There was a camp
• 2:35
  pushing for programmatic buying. They
• 2:37
  were very vocal and saying that it would
• 2:39
  lead to more efficiency and demand
• 2:42
  consolidation.
• 2:43
  And there was another group ping pushing
• 2:46
  back equally hard mostly the agencies
• 2:50
  and the publishers and they were arguing
• 2:52
  that it would lead to lower prices and a
• 2:55
  worse experience for the advertiser. And
• 2:58
  of course there were people who worried
• 3:00
  about their jobs and the revenue. But
• 3:03
  I'm not seeing the same dynamic yet with
• 3:06
  AI. Maybe people have learned from the
• 3:10
  programmatic transition that you can't
• 3:12
  really hold back the tide or maybe they
• 3:15
  just haven't organized. But right now my
• 3:18
  takeaway is that this wave of innovation
• 3:20
  there is almost no organized resistance.
• 3:24
  The final thing is and again this is
• 3:27
  less surprising but that the actual
• 3:30
  implementation of AI is way behind the
• 3:33
  conversation. Yes, almost every post in
• 3:36
  the industry is about AI. Yes, almost
• 3:38
  every conference session is about AI but
• 3:41
  there is a massive gap between the talk
• 3:43
  and the action.
• 3:45
  In one session, the moderator asked how
• 3:47
  many companies in the room had actually
• 3:49
  run campaigns through Agentic Workflows
• 3:52
  and only about 20 to 25% of the hands
• 3:55
  went up. And based on my conversations
• 3:58
  in the hallways, I suspect the number is
• 4:01
  actually lower.
• 4:03
  Even when the moderators asked the
• 4:05
  panelists for specific examples of how
• 4:08
  they were using AI, the examples were
• 4:10
  very limited. So again, my takeaway is
• 4:13
  that the industry conversation around AI
• 4:16
  is way behind the operational reality.
• 4:21
  One closing thought, most of the
• 4:23
  sessions that I attended were focused on
• 4:26
  AI, but I also sat in one that was
• 4:29
  focused on privacy and the patchwork of
• 4:31
  laws that were being created by
• 4:33
  different states in the US. And it is a
• 4:36
  patchwork, but it's a patchwork that is
• 4:39
  becoming increasingly impactful and
• 4:42
  assertive.
• 4:44
  So the AI acceleration execution is
• 4:47
  occurring at exactly the same moment
• 4:50
  that privacy laws are becoming
• 4:52
  increasingly constraining which means
• 4:55
  that the companies that are going to
• 4:56
  succeed in the future are going to be
• 4:58
  the ones that are best at governance,
• 5:01
  control, and structuring their data. I
• 5:04
  am James Deika. I am the yield doctor.
• 5:07
  Thanks for watching.