YT YouTube2Text

YouTube2Text → Transcripts → Beyond the Artifact Mindset

Beyond the Artifact Mindset — Transcript

by Martech World Forum · 2,814 words · 429 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00Throughout
0:05the enterprise martekch outlook research
0:07briefing uh one of the strongest themes
0:09has been that AI readiness is ultimately
0:12an operating model question. The
0:14organizations getting further are not
0:16just adding generative tools uh
0:19willy-nilly. You know they have stronger
0:21governance uh stronger metadata
0:23workflows uh their content operations
0:25tend to be stronger too. So I wanted to
0:28bring in a practical perspective from
0:30someone that sits right at the point
0:31where AI ambition meets the reality of
0:34what brands can safely create, manage
0:37and deliver at scale. Jeff Eaton joins
0:39us from Contentful, a platform closely
0:41associated with this space with
0:43structured composable content and the
0:45operating models behind digital
0:46experiences. Jeff is the senior director
0:49of content systems and is a digital
0:51strategist with more than 25 years of
0:53experience in content architecture,
0:55editorial workflow, and digital
0:57publishing. A fun tidbit about Jeff is
1:00he started publishing a neighborhood
1:02newspaper on a FisherPrice printing
1:04press as a child and then later became a
1:07founding partner at Autogram before
1:09joining.
1:09>> My pleasure. It's great to be here.
1:10>> Jeff, [snorts] thanks for joining us.
1:12>> My pleasure. It's great to be here. So
1:14let's start with the content foundation
1:17because it is becoming a much bigger
1:19part of the AI readiness conversation.
1:21Our research suggests that uh customerf
1:23facing AI readiness depends a lot on um
1:28access to not a lot more than just
1:29access to generative tools. So, you
1:32know, workflow governance, metadata,
1:34approvals, brand controls, measurement,
1:37all these fairly unsexy things are
1:39really becoming a an increasingly
1:41important part of the content
1:43foundation. [snorts] Why are these
1:45things so important?
1:46>> Generative AI has been an eat your
1:48vegetables moment for a lot of
1:49organizations. Those, you know,
1:51unglamorous aspects of the communication
1:55stack have suddenly become a lot more
1:59critical. And I think I mean it's ironic
2:01in a way because the sort of the demo
2:05shine of a generative AI solution is you
2:08roll up, you tell it what you want it,
2:10bam, it spits it out and it's it's
2:12amazing. It took just seconds to do what
2:14would have taken us 30 minutes. And to a
2:18certain degree that's true, but once you
2:21get into a scenario where you're trying
2:22to do that at scale or you're trying to
2:24do that within the constraints of a
2:27particular business or a particular
2:29campaign or regulated industry or a
2:32particular audience, you're not talking
2:34about, hey, can it make something that
2:36sounds good? you're talking about being
2:38able to consistently and reliably and
2:40reproducibly build things that work
2:43together consistently within um a wide
2:47range of different kinds of needs across
2:50an organization. Um and I think these
2:54are challenges that a lot of really
2:56large enterprises have, you know, had to
2:58deal with for a long time. You know, any
3:00anybody who works in a, you know,
3:02Fortune 500 company knows somebody out
3:04there cares about taxonomy very deep in
3:07their heart and they, you know, their
3:08job is to make sure everything works
3:10and, you know, to, you know, push
3:12Sisphus's stone up the hill every
3:14morning and make sure everything's
3:16tagged well. Um, but for a lot of
3:18organizations, unless you were working
3:20at that scale,
3:23felt like a a nice to have. you know, it
3:25was like, oh, well, there's certain
3:26things we could do with it, but we don't
3:29we don't we're not really that big yet.
3:31And I think the reason those processes
3:34were always really critical inside of
3:36large enterprises and large orgs was
3:39that their ways of making sure things
3:41work even if everybody's not in the same
3:43room and on the same page. you know,
3:45with a small team of, you know, like
3:47four, you know, even, you know, a couple
3:49dozen people, you know, scrappy working
3:51on a project. You know, everybody gets
3:53in sync and you can kind of, you know,
3:55you can be firing on all cylinders and
3:58producing things even if you're not
4:01explicit and, you know, spelling
4:04everything out whenever you're talking
4:05to each other. In a large org, you got
4:09to figure out how to make sure those
4:10things survive the handoff to the next
4:13business unit or, you know, the go they
4:16make it into the brief for the, you
4:17know, for the agency that's going to
4:19help. What's interesting to me is that
4:22generative AI
4:24brings that problem to the team of two.
4:29You know, as soon as you bring AI in,
4:32you're not just dealing with, oh, can it
4:34accelerate how quickly can we can
4:36produce copy? You're now having you now
4:38have to figure out, well, how do we make
4:41this machine get the stuff that we've
4:45intuitively gotten for years in our
4:49business, in our job, even in our
4:51function inside of this, you know,
4:52inside of a particular, you know, org.
4:54And I think to me at least that's where
4:58that process stuff fits. The taxonomy,
5:01the metadata, the good process, the
5:04having a good solid understanding of
5:07what your existing library of content is
5:09to draw on. So you're not just creating
5:11something new and tossing it out there
5:13every single time. You know, I I'll you
5:15can stage hook me at any time. This is,
5:17you know, this is near and dear to my
5:18heart, but um I I feel like the key for
5:23the key to actually being able to
5:25leverage generative AI effectively at
5:28scale is building those kinds of
5:30infrastructure tools that ironically
5:32we've always needed in large
5:35organizations. But no one realized quite
5:38how critical they were until we had
5:41these awesome, fantastic tools that
5:43suddenly need them even more than we do.
5:46Once that foundation and um you know
5:49that that uh context that you really
5:51need exists um you know the next
5:54challenge uh for for big enterprises is
5:56about scaling relevance without scaling
5:59complexity at the same rate. Enterprises
6:01of course want more contextual and
6:03responsive customer experiences. But you
6:06know fragmented workflows, limited
6:08resources, uh you know tough
6:10coordination can make personalization at
6:13scale you know really expensive to
6:15actually go and create and operate. You
6:17know the risk is that every new audience
6:19uh you know moment or campaign or
6:21channel creates another set of content
6:23variants and approvals and things for
6:25people to check. So how can enterprises
6:29increase personalization and speed
6:31without creating you know this
6:33unmanageable explosion of uh content
6:36variance approvals and you know channel
6:38specific work.
6:41Um, so that, you know, that that's the
6:44that's the $64,000 question. I think for
6:46a lot of organizations that are they've
6:48bought into the value of being able to
6:51target an audience effectively, but now
6:54they're sort of, you know, it's Monday
6:56morning and they realize, ah, that means
6:59we have to make 12 times as much content
7:01and we were already drowning in it to
7:03begin with. And you know, obviously
7:05that's that's not really sustainable.
7:07Even
7:09even if generative AI is capable of
7:11producing the raw materials for
7:14campaigns at that scale, the process of
7:17managing and monitoring them and making
7:19sure they stay up to date and all that,
7:22even just the oversight and governance
7:24work to go along with that becomes
7:27unsustainable. And what all of our
7:30customers who do this work at scale have
7:34discovered is they're not
7:36they have to move beyond an
7:38artifactoriented
7:40way of planning and producing and
7:42towards a system way of a systems way of
7:45planning and producing what they're what
7:48they're creating. um you know whether
7:50it's with you know a pile of
7:53spreadsheets and a bunch of Google docs
7:55and wishes and prayers or a tool like
7:57Contentful um what they're creating is a
8:02system that turns the stuff they want to
8:05communicate to their audiences into the
8:08final deliverables that work in a
8:10particular context. And if they're used
8:14to thinking about every different
8:16combination of audience, moment, you
8:20know, context or place, channel,
8:22whatever, as being a new artifact we
8:27have to go produce,
8:30that's how you get buried. And that
8:32artifactoriented way of thinking about
8:34things, it I mean it goes really deep in
8:36lots and lots of organizations. And
8:37we've even talked to Contentful
8:40customers who've, you know, done the,
8:42you know, the classic lift and shift
8:44migration and, you know, pulled all
8:45those stuff over. And then on the other
8:47side, what they realized was, oh man,
8:50you know, we've got a new platform, but
8:53we also really needed to think about how
8:56we were going to be working on that
8:58platform, not just what the feature list
9:01was going to give us. And I think that's
9:02where the concept of composability not
9:07just as a technical architecture but as
9:10an approach to content itself becomes
9:12very important. The idea that you're
9:15pulling together reusable pieces or
9:19tailoring just the just the elements
9:21that need to be created just for a
9:24particular audience or a campaign. and
9:27you have not just a giant Z drive full
9:31of mystery meat that you know you can
9:33dig around in and see if you've done
9:35something like that before, but pieces
9:38that are built to the right size and the
9:42right shape to be assembled on the fly
9:45to meet those new communication needs.
9:47And that's a combination of both
9:49technical capabilities in platforms and
9:52the the game plan that you know that
9:54customers bring to the table. And you
9:56know we we're I think as Contentful has
9:59grown one of the challenges for us has
10:02been moving beyond just providing
10:05technical infrastructure and also
10:07providing you know increasingly more and
10:09more tools to support that kind of
10:12process. Not just what they need to
10:14store but what they need to do with
10:16their stuff and what they need to be
10:17aware of and what kinds of processes
10:20they need to be engaging in on an
10:22ongoing basis. And it's um interesting
10:24because through our research we found
10:26that so we we basically asked a question
10:29about um you know how do you see
10:31yourself as a a martekch organization
10:34you know are you high performing low
10:35performing somewhere in the middle and
10:37there was a a a strong link between
10:40folks that were high performing and um
10:42also being high performing on content
10:44supply chain so their ability to
10:46actually manage the the whole content
10:48supply chain. So what we've found is
10:50these these MK advantage organizations
10:53that that we call them um they they tend
10:56to be a lot stronger on content. So if
10:57they can manage content, it's a sign
10:59that they're actually much more likely
11:01to provide a better customer experience,
11:03which we thought was a really amazing
11:04finding and it it really backs up um
11:06what what you were just saying there. Um
11:08[snorts] and
11:10>> I I love to hear stories like that.
11:12[laughter]
11:13>> Um but you you also mentioned the
11:16architecture question. So you um you
11:18know you were talking about um how
11:20architecture plays into this
11:22flexibility. Um of course there's
11:24operating model um but technical
11:26architecture is is a big part of it. So
11:29one of the other findings in the report
11:32uh was that most big enterprises uh you
11:35know they're not going for purely sweet
11:38or purely best of breed/composible.
11:41they're sort of picking and choosing and
11:42and almost all of them said that or they
11:44described their suite as a hybrid. Um
11:47you know the real question was about how
11:49much control they need um you know how
11:51much integration complexity and um you
11:53know operations that they could actually
11:55take on.
11:56>> So in in your view what what
11:58distinguishes that sort of pragmatic
12:01composability um from an architecture
12:03that is you know technically flexible
12:05but operationally difficult to govern
12:08and scale. that's a been an active
12:10conversation in like the the CMS nerds
12:13community for a number of years now. Um
12:16you know it's like there's different you
12:18know small events around the country
12:19that we sort of you know folks sort of
12:20you know congregate at and you know have
12:22their have our super nerdy conversations
12:24about how stuff's shaping up. And I
12:25think one of the themes that's been
12:27persistent over the last few years is
12:30that
12:31>> um that hybrid approach you're
12:33describing is something that everyone is
12:35arriving at. pure composable vendors
12:38like um you know like us Contentful um
12:41we aren't staying in the realm of like
12:45pure you know pure
12:48one pure slice of the tech stack in part
12:51because
12:52>> over the years we realized that there's
12:54a lot of adjacent needs that just
12:57[snorts] need to be solved and often
12:59they need to be solved where the content
13:01is being produced and managed. Sometimes
13:03they don't for certain organizations and
13:05you know they're content to you know
13:08content to use other tools for that but
13:09oftent times you know it's stuff that
13:13we're we're we're not precious about the
13:15idea that we're one little tiny you know
13:18piece in an ocean of you know millions
13:20of different composable pieces but we
13:23want to make sure that we play well with
13:26whatever systems need to be integrated.
13:28We're never going to be, you know,
13:31solving every end toend problem of a
13:33customer's data, communications,
13:35delivery, stuff like that. You know,
13:36that's that's not where we live. But we
13:40need to be ready to fill certain
13:43critical gaps in you know in a
13:45customer's toolbox and bring a solid
13:48opinionated perspective on what
13:50approaches are going to work when they
13:53aren't coming with you know the entire
13:56stack ready for Contentful just to slot
13:58in. And I think ironically even the
14:01large um you know what you know
14:03classically we would always call the
14:04monolithic vendors have been moving in
14:07that direction too where they're more
14:09open to the idea of interoperation and
14:11swapping in different components to you
14:13know to replace something that might
14:15have been out of box functionality with
14:16their platform. So I think you know to a
14:19large extent uh you know the entire
14:22field is sort of converging on the idea
14:24that no matter what label we put on it
14:27the ability to work that way and the
14:29ability to right size what you're going
14:32to pick out from you know what you're
14:34going to pick out because it's just what
14:35you need what's the stuff that's going
14:37to be you know it's it's good enough
14:39it's what we've got you know we're going
14:40to work with it you know it all works
14:42together already there's no need to
14:44touch it what do we need to change I I
14:46think everyone one everyone has come to
14:49a place where they realize that yeah
14:50that that that's how the projects go in
14:52in the real world and um you know as
14:55somebody who's you know over my career
14:56I've worked at you know monolithic
14:58headless um terrifying PHP artififices
15:02you know the the whole deal um and I
15:05think it's really encouraging to see
15:07that pragmatic approach being uh sort of
15:10turning into the the shared
15:11understanding that everyone kind of you
15:13know uses regardless of what direction
15:15we're coming at it from. Um but I do
15:17think it it
15:19that said it also means that the
15:22organizations that are doing the
15:24shopping and doing the you know
15:26decision-m about that um figuring out
15:30what their future state needs to look
15:32like and having you know open eyes about
15:37the way that they work with the tools on
15:39a day-to-day basis potentially needing
15:41to change. you know, in those areas that
15:42we were talking about like dealing with
15:45the operational complexity of, you know,
15:47a glorious world of onetoone
15:49personalization for everyone, it's like,
15:51yeah,
15:51>> well, you know, that that raises cho
15:53that raises questions that you're going
15:55to have to grapple with about how you
15:57work every day. And sometimes that will
16:00that will change the questions you're
16:02asking about what tools fit best.
16:04>> Amazing. So, Jeff, I think that's a a
16:06great place to leave it. um thank you
16:08for helping us move the the conversation
16:10around AI and composability you know
16:13into the operating reality behind it. I
16:16think the important point is that better
16:18customerf facing AI is not just a a
16:20model problem. You know, it really
16:21depends on on structured context, you
16:24know, metadata, governance, all these
16:26unsexy things that we've uh that we've
16:28been talking about. Um, you know, these
16:30are the things that let teams
16:31personalize and and reuse content
16:33without, you know, uh, uncontrollable
16:35complexity. So, Jeff, thanks again for
16:37joining us and a massive thank you to
16:39Contentful for partnering with us on the
16:41enterprise MKH outlook for 2026. Thank
16:44you. Thank you.

About this transcript

This page contains the full transcript of Beyond the Artifact Mindset by Martech World Forum, generated from the public captions YouTube serves with the video. The transcript has 2,814 words across 429 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.