Latest / The Adults In The AI Room / Stephanie Lemieux on Taxonomy & Information Architecture - Episode #13

Transcript

Stephanie Lemieux on Taxonomy & Information Architecture - Episode #13 · The Adults In The AI Room

• 0:36
  SPEAKER_05: Welcome to today's podcast.
• 0:38
  SPEAKER_05: I'm Seth Early.
• 0:39
  SPEAKER_05: And I'm Chris Fetterstone.
• 0:41
  SPEAKER_05: Before we get started, I'd like to give a shout out to our sponsors, Simpler Media and the Marketing AI Institute, and of course, my organization, Early Information Science.
• 0:51
  SPEAKER_05: Our guest today is a woman after my own heart.
• 0:54
  SPEAKER_05: She's an expert in taxonomy and unstructured content.
• 0:57
  SPEAKER_05: As president and principal consultant at Dovecot Studio, she consults on content modeling, information architecture, search, and more.
• 1:05
  SPEAKER_05: In her spare time, other than wondering why she has no spare time, she serves as chair of the taxonomy bootcamp conference, which is held in each November in Washington, D.C.
• 1:15
  SPEAKER_05: Please welcome Stephanie Lemieux.
• 1:18
  SPEAKER_05: Hi, Seth.
• 1:19
  SPEAKER_06: So at this point, at this point, Stephanie, what we should have is like that background applause, right?
• 1:24
  SPEAKER_06: Where it's like the roar of the crowd.
• 1:27
  SPEAKER_06: So that's right.
• 1:29
  SPEAKER_06: I think what I am not most interested in, Stephanie, but since you've known Seth, I want to know maybe some some background dirt on him.
• 1:40
  SPEAKER_06: Oh, wait a minute.
• 1:41
  SPEAKER_06: Wait a minute.
• 1:42
  SPEAKER_05: This is not in the script.
• 1:43
  SPEAKER_05: This is not in the script.
• 1:46
  SPEAKER_05: Well, background, you know, I mean, I there's probably a lot, but you know what?
• 1:51
  SPEAKER_05: I tell you, we have a mutually assured destruction clause.
• 1:55
  SPEAKER_05: So she tells me my background dirt, I get to tell hers.
• 1:59
  SPEAKER_05: So yeah, it's there's no back.
• 2:02
  SPEAKER_05: There's no dirt on Stephanie.
• 2:03
  SPEAKER_05: Sure, true.
• 2:04
  SPEAKER_01: Let's let's both back away from the buttons.
• 2:09
  SPEAKER_06: So, you know what?
• 2:11
  SPEAKER_06: You know, in all in all seriousness though, Stephanie, like, you know, Seth's um, you know, we've been doing um a lot of really fun work together, and I know his, you know, his MO, him loving to jump into, you know, the information architecture and things, and you know, you guys kind of being kind of at the same claw.
• 2:31
  SPEAKER_06: I would like to get a, you know, you know, what'd they say, birth of a feather, you know, flock together.
• 2:35
  SPEAKER_06: You know, what does that look like as you've worked together in the past and how where's that led you to and all those kind of things?
• 2:42
  SPEAKER_06: You be the dirt without the dirt, right?
• 2:44
  SPEAKER_01: He wants our meat cute.
• 2:48
  SPEAKER_03: Yeah, exactly.
• 2:50
  SPEAKER_01: I met Seth uh pretty much the second I was done my master's degree, and even just before I was done.
• 2:58
  SPEAKER_01: So I was uh in library school, uh trying to figure out what to do with my life.
• 3:04
  SPEAKER_01: Uh, and this particular library school had just kind of transitioned away from being pure library stuff and starting to focus on this new at the time concept of information architecture and taxonomies and all this stuff.
• 3:18
  SPEAKER_01: So uh I had kind of fallen into that and and found some interest.
• 3:22
  SPEAKER_01: And I had run into Seth uh uh through this community of practice that he was running at the time called the taxonomy community of practice.
• 3:32
  SPEAKER_01: And I just kind of reached out to say, hey, I'm really interested in this content and I I want to do some uh summaries of some of the key things that have come out of this group.
• 3:42
  SPEAKER_01: And I guess that impressed him because he immediately offered that maybe I could help him with his webinar uh that he was running at the time for that community of practice.
• 3:52
  SPEAKER_01: And then that snowballed pretty quickly into us working together.
• 3:57
  SPEAKER_01: And so I joined uh back then it was called Early in Associates, uh right out of school, pretty much.
• 4:05
  SPEAKER_05: How long ago is that, Stephanie?
• 4:08
  SPEAKER_01: Uh, like almost 15 years ago, I think, if not more.
• 4:12
  SPEAKER_05: Yeah.
• 4:13
  SPEAKER_05: Wow, yeah.
• 4:14
  SPEAKER_05: That's uh it's a long time in technology dog years.
• 4:18
  SPEAKER_01: It is.
• 4:19
  SPEAKER_01: Back in the day when taxonomy was uh a very triggering word and people didn't want you to use it in meetings so that you didn't scare people.
• 4:29
  SPEAKER_05: Right.
• 4:29
  SPEAKER_05: Yeah, it was so funny.
• 4:30
  SPEAKER_05: Yeah, no people really had no clue.
• 4:33
  SPEAKER_05: It was really interesting.
• 4:35
  SPEAKER_05: The level of immaturity in the marketplace was was pretty interesting, and there's a lot of opportunity to to really impact uh organizations.
• 4:44
  SPEAKER_05: And uh Stephanie worked on you know some great projects.
• 4:47
  SPEAKER_05: Best Buy was one.
• 4:48
  SPEAKER_05: Um we had some incredible success there.
• 4:51
  SPEAKER_05: Um uh Ford Foundation, Motorola.
• 4:55
  SPEAKER_05: Uh and when Ford Foundation, it's like you know, you're impacting the world, you know, you're you're impacting that impacts the world.
• 5:02
  SPEAKER_05: So it was very it was very gratifying, and it was very um a lot of fun and and great work.
• 5:07
  SPEAKER_05: We worked with Hasbro, we worked, well, I don't know, there's a bunch of companies that we worked with together.
• 5:12
  SPEAKER_05: So it was it was a lot of fun.
• 5:14
  SPEAKER_01: What I'm most grateful for is we were really there, and you took me under your wing at the moment where e-commerce was really booming.
• 5:23
  SPEAKER_01: And what is e-commerce without taxonomy and information architecture, right?
• 5:29
  SPEAKER_01: So we would be in these big meetings with uh large retailers trying to convince their executives that you need to care about product categories and and how you make data available on your site and you have to care about user terminology.
• 5:47
  SPEAKER_01: Uh, and it was a hard sell back then, but now it's a breeze.
• 5:51
  SPEAKER_01: But you know, it was it was so great to be there at the time when all of this was really growing.
• 5:58
  SPEAKER_05: But I'll tell you, Stephanie, I want to push back a little bit on that.
• 6:02
  SPEAKER_05: It actually still depends on the organization, but it is still many times a difficult sell, believe it or not.
• 6:08
  SPEAKER_05: We've had people who are in charge of e-commerce asking, why do I need uh information architecture?
• 6:14
  SPEAKER_05: Astoundingly enough.
• 6:15
  SPEAKER_05: I mean, it's it's crazy.
• 6:16
  SPEAKER_05: You know, maybe I I don't know, but but you know, right now we're with one organization and like, well, we have to convince our our leadership why this is important, even why their website is important, because they do a lot of work with their sales force direct to their customers.
• 6:31
  SPEAKER_05: And it's like that is not that's a bit dated, uh maybe, but um it's a model they use, but but the value of the content and the product information and the supporting information, all of that, the reference stuff, that is super important to salespeople, even if they're not doing e-commerce.
• 6:51
  SPEAKER_05: But but I still think it is interesting, you know.
• 6:53
  SPEAKER_05: I think companies nod their heads, they they understand it to some degree, but many times they're not necessarily making the investment um that they should.
• 7:02
  SPEAKER_05: And uh, and there's a lot of complexities behind that.
• 7:05
  SPEAKER_05: And then you still find a lot of silos.
• 7:07
  SPEAKER_05: So, yes, the the the early days were very interesting and uh and there was a lot that we could get done.
• 7:13
  SPEAKER_05: And today the the industry is much more mature, but there's still a lot of foundational um uh problems and issues.
• 7:22
  SPEAKER_01: Definitely, there's still a lot of foundational problems and issues, and I think that uh the complexity question has changed as well.
• 7:32
  SPEAKER_01: So, you know, back in the day of early e-commerce, you had to worry about the website and you had to worry about how you were going to get data on the website for your products, but you didn't have quite as many systems connecting as you do today.
• 7:46
  SPEAKER_01: So, as things have migrated over to omnichannel and you have all these data syndication services, and you have lots of different manufacturers that are producing their own data now, like it's it's a much more complicated question than it was 15 years ago.
• 8:04
  SPEAKER_05: That's a really good point.
• 8:05
  SPEAKER_05: And then that marketing technology ecosystem has gotten extremely complex.
• 8:10
  SPEAKER_05: We're actually our next webinar on um uh what is it next week, I believe, when we have a webinar on uh next Wednesday, I think.
• 8:18
  SPEAKER_03: Yep.
• 8:18
  SPEAKER_05: Actually on that topic, that very topic.
• 8:21
  SPEAKER_05: It's uh about how to get a handle and understand and uh and assess your marketing technology stack.
• 8:28
  SPEAKER_05: So for those of you listening, you can if it's after uh March 23rd, you can always get to the recording.
• 8:33
  SPEAKER_05: But uh, but yeah, it's a complex, complex ecosystem.
• 8:37
  SPEAKER_05: It's very fragmented.
• 8:38
  SPEAKER_05: There's something like 15,000 tools out there at last count, and many organizations will have a hundred, 150 technologies, and they are very rarely in alignment and harmonized and normalized.
• 8:51
  SPEAKER_05: And uh in order to get things like, you know, well, you talked, you can talk about this, but um, you know, in order to get that 360-degree view of the customer, in order to, you know, have that seamless customer experience, you really do have to get all those things uh aligned.
• 9:06
  SPEAKER_05: And what have you been seeing in the marketplace these days in in that regard?
• 9:12
  SPEAKER_05: You're seeing the complexity, but are you seeing organizations understand um the value, especially across that marketing ecosystem?
• 9:21
  SPEAKER_01: I I think the understanding is definitely increasing.
• 9:24
  SPEAKER_01: I'm having to do less and less kind of ROI and sales pitch about the the underlying value of having good data and and uh information architecture.
• 9:38
  SPEAKER_01: Uh, but what I am seeing that's shifting a little bit is this question of machine learning and AI kind of jumping into the fray and muddying the waters a little bit and saying, yes, we absolutely need information architecture, but let's let the machine do it.
• 9:55
  SPEAKER_01: Uh so we don't want to dedicate people and and consulting budget or you know, uh system development budget to solve these problems.
• 10:04
  SPEAKER_01: We just want to throw throw some machine learning at it and see what comes out the other side.
• 10:09
  SPEAKER_05: And how's that been working out for them?
• 10:12
  SPEAKER_05: It's been really challenging.
• 10:15
  SPEAKER_06: You know, I'll I'll tell you, you know, with with uh just the problem scape you guys outlined, um, which we know is is huge, um it there's that fallacy that, you know, like you said, Stephanie, if I just throw machine learning at it, that it'll solve all of it.
• 10:32
  SPEAKER_06: But but people don't understand that that you still have to um, you know, in a training scenario, you still have to get the data into some common, you know, canonical or you know, or normalized format for the machine to even know what the hell to look for in any kind of a pattern.
• 10:53
  SPEAKER_06: So what do you you know when you work with customers?
• 10:56
  SPEAKER_06: Because we've got, you know, and that's with probably structured data, but we also have now, you know, so much unstructured data, we've got rich media that needs to be dragged out of that, you know, and to turn it into data.
• 11:11
  SPEAKER_06: So when you're with when you're working with your customers and stuff, how do you help them over the humps of understanding that there's thousands of data types?
• 11:21
  SPEAKER_06: They don't have access to all of them, but they want to get access to all of them.
• 11:26
  SPEAKER_06: And how do you lead them and guide them through that?
• 11:29
  SPEAKER_01: Yeah, it's been uh it's definitely been a challenging transition, I think, over the last couple of years, uh, as uh clients are certainly getting more savvy about the question of uh integration and and this concept of data orchestration, certainly.
• 11:48
  SPEAKER_01: Um getting people away from this idea that there's this magical solution that you can just pop into place and it'll churn through all of your data and create an infrastructure for you.
• 12:03
  SPEAKER_01: Uh there's a lot of demystifying and education required up front in these projects.
• 12:10
  SPEAKER_01: Uh, I'm I'm right in the middle of a project right now where I'm doing um enterprise taxonomy strategy for a content-heavy organization who also has uh data as product.
• 12:24
  SPEAKER_01: And uh in the middle of these interviews, practically every person that I'm talking to who's talking to me about either content publishing, digital asset management, uh ad targeting, all of these things that are part of you know creating content at scale are all kind of convinced about the value of taxonomy and information and uh information architecture and data management.
• 12:49
  SPEAKER_01: But in the same breath, they'll say, oh, but we also want to get uh some machine learning in there and AI to see if we can be more agile and be more responsive to our users' needs.
• 13:03
  SPEAKER_01: So I think there's just uh a confusion about what the scope of these new tools are and what kinds of use cases they can solve well.
• 13:16
  SPEAKER_01: And that even if you do implement these kinds of uh new sophisticated tools, that they still depend on a core information architecture behind them.
• 13:29
  SPEAKER_03: Right.
• 13:29
  SPEAKER_01: That you can't just throw an AI at uh, you know, thousands and millions of pieces of content and not have any kind of structured metadata or controlled vocabulary behind it for which that AI can like sink its teeth into.
• 13:45
  SPEAKER_05: Right.
• 13:46
  SPEAKER_05: We need to tell the AI what's important.
• 13:48
  SPEAKER_05: We need to tell the names of our products, the names of our services, our processes, our content, our regions, our customers, our customer attributes.
• 13:57
  SPEAKER_05: And that's the key, the critical piece.
• 13:59
  SPEAKER_05: Yeah, AI can do uh certain things in machine learning, but it needs that reference architecture.
• 14:04
  SPEAKER_05: And that's what people don't necessarily understand.
• 14:07
  SPEAKER_05: And that taxonomy is foundational to that that reference architecture.
• 14:11
  SPEAKER_05: You know, when we when we build the the new the new buzzword du jour that scares executives is ontology, right?
• 14:18
  SPEAKER_05: And so when we build taxonomies, we build multiple taxonomies, and then when we have the relationships between them, it really is an ontology, like product, like services for a product, you know, solutions for a problem.
• 14:31
  SPEAKER_05: You know, each one of those may be a vocabulary, but when we relate them together, we build a triple, we build uh edges, we build uh associative relationships, and those become part of the ontology.
• 14:43
  SPEAKER_05: And so it it really is an area that's getting a lot of uh attention at some levels, but very, very uh deep technical attention.
• 14:52
  SPEAKER_05: But I think there's still uh people are still grappling with the use cases and the role of that.
• 14:58
  SPEAKER_05: Do you want to talk a little bit about what you've seen in either graph data or knowledge graphs?
• 15:03
  SPEAKER_05: Because I think what you said just now is very, very important that you know people are thinking, okay, but well wait, the the the the AI vendors say that's possible, so it must be possible, right?
• 15:13
  SPEAKER_05: It may be you know that you don't need any any architecture that the tool does it all.
• 15:18
  SPEAKER_05: Oh, wait a minute.
• 15:18
  SPEAKER_05: Yeah, I yeah, I don't know if that's true.
• 15:20
  SPEAKER_05: I don't know if I I I I know that sometimes that salespeople are sometimes not completely oh come on, what are you trying to say, man?
• 15:30
  SPEAKER_05: You know what I think uh uh you know the difference between you know the difference between a a used car salesperson and an AI software salesperson?
• 15:41
  SPEAKER_01: No way.
• 15:42
  SPEAKER_05: Well the AI software no the AI software uh the used car salesperson knows when they're lying.
• 15:52
  SPEAKER_05: Think about it, right?
• 15:54
  SPEAKER_05: Heads of engineers, yeah, right?
• 15:56
  SPEAKER_05: It gets translated into marketing, it gets translated into sales, and then they try to talk to the customer, and it and a loss is a lot is lost in translation, you know.
• 16:04
  SPEAKER_06: Well, I think part of it too is is you know, they just you know, as we talk about this, and I think after you know, you kind of dig into Seth's question there, you know, one the one thing that I'd love to kind of get your sense of is how many organizations don't really know and understand what their starting use case needs to be.
• 16:26
  SPEAKER_06: And you know, because they want to boil the ocean.
• 16:29
  SPEAKER_01: That's exactly what I'm seeing.
• 16:31
  SPEAKER_01: So uh knowledge graph, I agree, is the most popular buzzword in that sort of information management space, definitely in my world at least.
• 16:44
  SPEAKER_01: Uh, and everyone wants to do a proof of concept.
• 16:48
  SPEAKER_01: And the two key things that I'm seeing is they want to do a proof of concept, but they don't have a clue of what the proof of concept should do.
• 16:57
  SPEAKER_01: Uh so they know that they want to get into knowledge graphs, but they don't know what the graph can do, what they should point it at, uh, and what problem really they're trying to solve to your point.
• 17:08
  SPEAKER_01: Um, the other piece that I'm seeing is that, okay, if they do have a clear use case and they have the budget to move forward with a with a proof of concept, they hire a vendor and then they don't understand that they have all of this uh infrastructure that they have to build behind it for it to be successful.
• 17:26
  SPEAKER_01: So they have to get into data normalization, they have to get into taxonomy and ontology building, they have to get into a lot of framework building in order for that proof of concept to be successful.
• 17:40
  SPEAKER_01: So they can't just point the technology at it and then magic happens and you have a knowledge graph.
• 17:48
  SPEAKER_05: Right.
• 17:48
  SPEAKER_05: And what's interesting as well is many of the things that we used to call uh fasted search, for example, are showing up as knowledge graph applications, right?
• 17:59
  SPEAKER_05: So that is one of the when people talk about semantic search or graph search or powering um virtual assistants.
• 18:06
  SPEAKER_05: You know, that's stuff that we've done, but we didn't call it a knowledge graph.
• 18:11
  SPEAKER_05: We called it taxonomy, information architecture, ontology, um, all you know, all the things that we need to do in terms of building our content models, you know, normalizing the data, uh, making sure it's harmonized, making sure we understand data flows, ownership, provenance, quality, all of those things.
• 18:28
  SPEAKER_05: And and again, you know, it's many of the same, the same concepts are applied.
• 18:33
  SPEAKER_05: Uh, I'm working on a workshop that I'm delivering for computers and libraries um uh at the end of the month with a woman named Mara Mara Owens, and she's very knowledgeable and very technical.
• 18:45
  SPEAKER_05: And we've been able to harmonize our terminology, right?
• 18:48
  SPEAKER_05: Because part of it was, oh, you call it this, I call it this.
• 18:52
  SPEAKER_05: But really, you know, the use, the same use cases, the same derivation of terminology, derivation of relationships, understanding of nodes, understanding of of data flows, very, very analogous, very analogous, different tools, different applications, ultimately different experiences, and you can do a lot of other things, you know, with with graph data and knowledge graphs in terms of data governance and um and uh data, you know, data uh uh provenance, as they say, data quality, processes, uh, all sorts of things that relate to keeping data in good shape can be part of uh but but again, it's it's kind of a Swiss Army knife, but you really need to know what you're gonna do with that Swiss Army knife, right?
• 19:37
  SPEAKER_05: It's it's in and of itself, it's not gonna do much for you, but what are the problems you're trying to solve?
• 19:42
  SPEAKER_05: And that comes back, as I always say, to use cases.
• 19:45
  SPEAKER_05: Same thing with AI.
• 19:46
  SPEAKER_05: If we say, you know, we need a knowledge graph um as part of our AI solution, well, what is it doing, right?
• 19:54
  SPEAKER_05: And is it give if it's giving you reference uh data, that's great.
• 19:58
  SPEAKER_05: If it's giving you relationships that you can do inference around, that's great.
• 20:02
  SPEAKER_05: Um, but you know, at the end of the day, what's the problem your AI is solving, right?
• 20:06
  SPEAKER_05: And it really is not about you know an entire process, it's about usually a step in the process.
• 20:12
  SPEAKER_05: Maybe there's an end-to-end process, but we need to understand what our process is and who the who the ownership who owns it, um, and uh what the what the data source is, what the data use is, upstream, downstream.
• 20:24
  SPEAKER_05: And again, that stuff is all all needs to be derived by humans.
• 20:28
  SPEAKER_05: The AI cannot do it itself.
• 20:30
  SPEAKER_05: And so, you know, when we start to see people talking about these things, you're right.
• 20:34
  SPEAKER_05: They you know, they're putting the cart before the horse.
• 20:36
  SPEAKER_05: They're talking about the technology.
• 20:37
  SPEAKER_05: We need graph data, we need a graph data pilot.
• 20:39
  SPEAKER_05: But it really is about where we need an AI product, we need machine learning.
• 20:43
  SPEAKER_05: It really is about going back.
• 20:44
  SPEAKER_05: You can't automate a mess, and you can't automate what you don't understand.
• 20:48
  SPEAKER_05: And so people need to come back and say, what is that fundamental problem and what is that fundamental process?
• 20:54
  SPEAKER_01: What one thing that I do find nice about those focus on the shiny new toy of you know, graphs and and AI is that it is bringing the light uh back onto this metadata debt that we've been accumulating over many years now.
• 21:12
  SPEAKER_01: Uh, you can't have a bunch of messy data underneath uh a knowledge graph.
• 21:18
  SPEAKER_01: And so I'm doing a knowledge graph proof concept right now.
• 21:21
  SPEAKER_01: And the idea was let's connect some uh data about our staff from Active Directory and the content that we have in our document management system and kind of build an expertise locator based on what topics people are authoring on, except that half the data in Active Directory is uh incomplete or incorrect.
• 21:44
  SPEAKER_01: And people have not wanted to tag in this document management system for years, and that's been an uphill battle.
• 21:51
  SPEAKER_01: And so the knowledge graph is not going to fill in any of that data or correct any of that data.
• 21:56
  SPEAKER_01: So we've had to do normalization, we have to go in and do a bunch of.
• 22:00
  SPEAKER_01: bulk tagging so there's just no way around dealing with that that metadata debt at some point yeah that's a great observation and I think yeah go ahead Chris I was just gonna say Stephanie when you go into these organizations which roles uh do you find are are the because I'm assuming you know that it's it'll be a lot of the same um which roles do you find are actually gonna help you propel these projects forward which ones are the naysayers which ones do you have to educate I'd love to know the dynamics of the of the different roles and profiles and personas that you have to go into and say okay that's gonna be our stakeholder but that's gonna be our promoter that's gonna be this and you know almost the dynamics of the people because really what you're talking about is the data is there yet we're limited by our own um naivety on how to get started and some of that naivety comes in the form of bias and everything else right so which roles do you find are the ones that are really accelerating and which ones do you find are actually blockers for you and how do you work through that but sometimes the data's not there as well but go ahead Stephanie take take take take that uh direction go go in that direction not there but it is I mean it's not clean but there yeah exactly yeah um interestingly enough I find that a lot of these interests in knowledge graphs and these solutions are coming from the business themselves uh who are interested in those use cases.
• 23:37
  SPEAKER_01: Uh they're coming sometimes from centralized knowledge management or information management shared service kinds of roles who are trying to enable digital transformation and are have been mandated by the organization to kind of solve these problems at large versus for a specific unit.
• 24:00
  SPEAKER_01: What I'm less uh faced with is where I would expect these questions to come up and which is like enterprise architecture.
• 24:08
  SPEAKER_01: Enterprise architects are definitely allies in these projects but I don't often see them as um as spurring these these interests or starting these projects.
• 24:20
  SPEAKER_01: So definitely more from the business side uh but in collaboration with folks over on the data governance or enterprise architecture side.
• 24:31
  SPEAKER_06: Is that what you're seeing by the way yeah I'm I think you're spot on yeah there I you know just to tie it off and then um it it it's always interesting because I there's a lot of folks on that listen to these that are looking for one you know because this is a this is we're we're you know if you think about it like the old OSI models we're getting to a the deep level of data prep and and data structure and data understanding and metadata and why it's all so important.
• 25:00
  SPEAKER_06: And part of that too is you know like you said we want to skip to solutioning it with a tool or something.
• 25:06
  SPEAKER_06: And there's a lot of folks that listen to this just to try to get an idea and a sense of of where to start but why why do I need to start there and why is that so you know why is that going to help me actually understand what I need to do you know or what it's gonna actually deliver in terms of true value based results as opposed to just some science projects.
• 25:27
  SPEAKER_06: So it's always interesting for them to see this right in terms of who's going to be the components and who's going to be the blockers.
• 25:34
  SPEAKER_04: We'll be right back.
• 25:35
  SPEAKER_04: We'd like to thank Marketing Artificial Intelligence Institute for their support of this podcast.
• 25:41
  SPEAKER_04: Marketing AI Institute is a media event and online education business that makes AI approachable action and accessible to marketing leaders around the world they run the world's leading marketing AI conference Make Off which brings together hundreds of marketing and business leaders to better understand pilots and adopt AI podcast is brought to a custom media group cover of customer experience technology podcasts covering the latest and customer experience markets covered the podcast design to deliver intact from data customer experience with digital touchpoints customer targetful data assets and voice of the customer big thanks to early information science to support the podcast.
• 26:27
  SPEAKER_02: Early information science to professional source forms proven methodologies design specifically to address product data content assets, customer data and corporate knowledge their information architecture effects that allowed clients from all over the world to officially organize access to have and utilize their data ultimate resulting in better business output.
• 26:46
  SPEAKER_02: They deliver customer driven solutions that scale and adapt that needs of a business as it evolved founded in 1994 early information science proudly delivered result driven solution for more than 100 brands across the top industries you can get more information about early information science at early.com.com working with customers in AI, the number one question that I get usually is where do I start?
• 27:14
  SPEAKER_02: What data do I need?
• 27:15
  SPEAKER_02: How do I gather it?
• 27:16
  SPEAKER_02: That's a great question.
• 27:18
  SPEAKER_02: Well if you want to catch the wave of AI and not be crushed by it, you really do need to read a book written by Seth Early called the AI Powered Enterprise.
• 27:26
  SPEAKER_02: It gives you the ability to look and understand how to harness the power of ontologies to make your business smarter faster and more profitable.
• 27:34
  SPEAKER_02: That goes into the intricate details of how to build an ontology develop and employ ontology as a secret weapon for the basis for how to actually utilize AI as a powerful competitive advantage.
• 27:46
  SPEAKER_02: That's book is really the first one to combine sophisticated explanation of how AI works with a practical approach of applying AI to the problems of the business.
• 27:55
  SPEAKER_02: And it's applied to you know very different cases from customer experiences to test operations to product development.
• 28:01
  SPEAKER_02: It's a powerful powerful tool in the arsenal of understanding how to deploy AI to make your business better.
• 28:07
  SPEAKER_02: The AI Powered Enterprise by Seth Early harness the power of ontology to make your business smarter faster more profitable out on Amazon.com today.
• 28:15
  SPEAKER_05: And now back to the show.
• 28:17
  SPEAKER_05: We're with Stephanie Lemieux who is president of DoveCon Studio so I think the question is really when when when organizations are looking at this stuff and that they think they're thinking about knowledge graphs, uh AI uh but but thinking about the aspect of this around um what's related to information architecture and they need to understand the nature of the problem but if they're trying to do a POC or or a pilot you know it take pick knowledge graphs.
• 28:45
  SPEAKER_05: Where where do they start and why do they start in a particular place?
• 28:50
  SPEAKER_05: What is your uh frame of reference on that?
• 28:54
  SPEAKER_01: Most of the time I'm getting uh a couple of flavors of knowledge graph uh graph interest uh one is more from a knowledge management perspective so wanting to be able to support knowledge workers in doing more with large uh swaths of unstructured content and connecting it up with data so giving a more uh robust experience for the knowledge worker in creating new connections creating new knowledge and being able to reuse and gain more value out of the content that already exists uh the other context that I'm seeing a lot of for graphs is uh definitely search search enhancement has is probably the one that I see the most proof of concepts around so being able to support more natural language questions and answers uh being able to provide more more robust search results that again span different collections different data sources um yeah those are those probably the two biggest areas that I'm seeing this interest in yeah another area that I'm seeing is data catalogs right just getting a handle on what they understanding data ownership understanding rights to that data the quality of that data who touches it you know what's that value chain but I agree that knowledge management and in a way you know it does harken back to the earlier days of knowledge management because we did do things like fast food search but we have the ability to now traverse the knowledge graph the way I think of it is you know the ontology is the scaffolding of knowledge for the organization all the taxonomy all the relationship between those taxonomies when we uh build a knowledge graph using that ontology we now have access points and retrieval points and what we're doing is we're we're using the knowledge graph to access the data so the ontology plus the data is the knowledge graph and that gives us great flexibility in terms of doing queries and in terms of traversing from one tool to another one system to another one content or document or asset type one expert to another one you know from from all those different angles.
• 31:05
  SPEAKER_05: So that's kind of how I'm also seeing it is you know looking at the structured data side and then and then you know as you say enhancing knowledge but very similar to what we've seen in the past but with some added tools functionality and and nuances.
• 31:21
  SPEAKER_01: Yeah uh definitely that continuation of this service oriented architecture that we were seeing so much of 10 years ago I think there's a lot of that just kind of rewarmed with different names like this data fabric concept which I was in a webinar about this literally yesterday uh I'm not a data expert but you know data fabric and headless CMSs, these are all ways of aggregating content and applying information architecture or structure to content from disparate sources to be able to have this service and uh serve up data or unstructured content from a centralized service layer or however you're architecting the the end result, but to be able to support lots of applications.
• 32:13
  SPEAKER_05: Yes, that's a great point.
• 32:14
  SPEAKER_05: And the headless CMS, you know, for those folks who may not be familiar with that term, it's like separating the content from the application, the downstream application, because the same piece of content can be used by many different systems.
• 32:28
  SPEAKER_05: And it's the create once publish everywhere model cope.
• 32:32
  SPEAKER_05: And one of our large um customers who has a hundred thousand employees and lots of technical content actually uses that to serve up four million knowledge transactions per day.
• 32:44
  SPEAKER_05: And at that volume all of that content comes through this service's architecture and this headless CMS but it gets surfaced in self-service in call centers in field service in marketing in distributors in bots and virtual assistants and in troubleshooting tools all sorts of applications be in embedded content within the equipment so that you know this that the equipment's going to be able to diagnose itself open a trouble ticket order the park it needs and when the service person gets there tell that person how to service it right so that's that we didn't see and the other thing that's interesting in my opinion is when we think about bots and virtual assistants you know we know those run on content right people say well we need more training data why why isn't it done well we're still training it right what do you train it with you train it with knowledge and content and data and so you know the organizations that are getting it that don't have their their handle on this the ones that are not paying attention not looking at information architecture not looking at headless CMS not looking at component architecture don't understand their knowledge have those silos have those messes they are going to be caught flat footed when these virtual assistants and bots get very functional highly functional there's that chapter in my book where I talk about that vision for the future and they're fluent and the only way to get there is through the data.
• 34:07
  SPEAKER_05: So what are your what are you seeing out there?
• 34:09
  SPEAKER_05: Do you do you do you agree with my with my uh forecast my my my prediction here that that they're going to be scrambling once they start realizing the value here what what you what do you think?
• 34:19
  SPEAKER_01: Absolutely I think there are some organizations that are further ahead on this journey than others.
• 34:25
  SPEAKER_01: And then there are some organizations that are just starting to think about getting started and they'll be able to learn from those organizations that have already started for a few years now and have fallen into all of these traps and learned the hard way that it's not a magic solution and that there's a lot of data architecture and content architecture that has to support these initiatives.
• 34:50
  SPEAKER_01: So they'll have a little bit of a leg up uh and there's certainly more maturity that is needed before anyone can fast track these kinds of projects um we have a client who's been working on graphs and ontologies as a gaming company and it's a a really large and complex system of uh game design so you have all these different studios that are developing games and they all have their own little silos you have the gaming platforms themselves you have product data you have uh customer or gamer data from their interactions with the actual games and then you have super omni-channel content provision that you're having to source not only your own uh platforms but all these within the game itself uh so trying to tie all of those data points together connect that with content and connect a gamer with the next best game or the next best quest within a game uh you know these are all super interesting data and content problems to solve but this has been a multi-year journey certainly it's been at least five six years that they've been working on components of this overall architecture and you know it's still challenging.
• 36:16
  SPEAKER_01: So it's definitely not anything that you can just spin up in a a six month uh project.
• 36:23
  SPEAKER_01: Right.
• 36:23
  SPEAKER_05: You're gonna have time and and that's exactly what this other organization they've been doing this for several years.
• 36:29
  SPEAKER_05: They they they worked on this project with based on the architecture that we developed and the processes and the and the taxonomies and then they've been running with it and they said it wouldn't have been impossible it would not have been possible without the work we did.
• 36:41
  SPEAKER_05: Well maybe they would have gotten there some other way but it was all based on that and they're saving hundreds of millions of dollars per year.
• 36:46
  SPEAKER_05: Now then the question though is with that many pieces of knowledge how do you still surface what's most relevant for an individual at the time that they need it right that's that holy grail of the right information for the right person at the right time right we've heard that for years for 30 years.
• 37:01
  SPEAKER_05: But the question is well how do you take other signals and use that to inform search right so you can have these components you can tag them with metadata but now even with that if there's that many pieces of uh uh content if there's that volume how do you ensure that an individual is getting the right information and it has to do with what are your thoughts on that in terms of leveraging weak signals of other types what else would you think would be solving helping to solve that problem?
• 37:32
  SPEAKER_01: Yeah depending on the context you certainly have uh user information that you can take uh you know maybe they're logged in and you have some concept of what profile they have what products they own what they've called about in the past uh you have behavior data what have they most recently been looking at so you can certainly take cues and signals from uh activity uh profile data location data all kinds of other potential sources that you have certainly and what that means is we have to be as intentional about our customer data models as we are about our content models and our product data models but a lot of times we don't see that right we see pre-coordinated development right I want it I want a CIO in the Midwest in the market for a router and it's you know I want I want um working mothers who like yoga who eat sushi on Tuesdays you know I mean those should be those should be separate you know uh separate facets they should be orthogonal vocabularies that describe different dimensions of that customer that's the way we line it up with attributes and with uh product attributes and content attributes and so on so it is interesting and I think you know we're starting just starting to see organizations begin to become aware of that and you know that's the that's the future state and um and and it does require harmonization of that customer data harvesting those signals from lots of different systems so you know once we do all of that once we find the right content and the right product information and we pull it from all these different systems and we get all that customer data you know there's a there's a very significant opportunity to truly truly personalize that experience but it requires much more shopping cart analysis or lookalike audiences it requires the integration and leveraging of all of these different signals.
• 39:24
  SPEAKER_01: So you know how far away do you see organizations being from that and if they do want to go down that path you know what are the things that you think they should be uh trying to do looking at doing um preparing for what are your thoughts that's uh also a very crunchy problem to solve uh you know and I see some organizations that have a uh one small use case that they're trying to apply uh behavioral analytics to or customer data to and that's a little bit of an easier problem to solve if you're being targeted.
• 39:59
  SPEAKER_01: Uh but kind of trying to to create a whole platform for that can certainly be a little bit overwhelming.
• 40:07
  SPEAKER_01: I was I'm talking with one client right now or we're working through Discovery and one of their use cases is being able to better connect their content with their users' interests.
• 40:17
  SPEAKER_01: And one of the things that's changed again in the last 15 years is people's expectations.
• 40:24
  SPEAKER_01: I no longer expect as a user of any kind of content platform to have to go in and manually tell you what I'm interested in.
• 40:32
  SPEAKER_01: I have been trained very well by Google and my feed on my device, my mobile device that it should know what I like, what I read most often and what I want to open app wise in the morning when I wake up my device has learned things about me through my interactions with it.
• 40:50
  SPEAKER_01: And that trains me to expect that from other sources in my life.
• 40:55
  SPEAKER_01: And so I think people are becoming more and more expectant that their experiences with different vendors that they work with, products that then brands that they want to interact with or even at work that their work tools should start to be moving in that direction.
• 41:16
  SPEAKER_01: And so we have to start solutioning those problems in different ways than saying you need to log into your profile and go choose your 17 interest categories from our taxonomy of 500 categories that's that's not you can't do that anymore.
• 41:33
  SPEAKER_06: Right, right well how do you how do you then you know help them help them work through though because one of the big mystical myths too is that it's one and done and it's never done so how do you help them work through understanding that it's not just a a point solution but this is actually a full-on journey that needs care and feeding forever really if they want to continue this.
• 42:01
  SPEAKER_06: You get you get the sense that they that they think that oh this is going to solve all the problems but it's never really done ever.
• 42:08
  SPEAKER_06: Right?
• 42:09
  SPEAKER_01: Yeah well I mean thankfully I'm not in involved in offering point solutions.
• 42:15
  SPEAKER_01: I I am doing that sort of infrastructure component so yep I get called in for point solution stuff and then it turns into all right you you wanted a taxonomy or metadata schema for this DAM or for this headless CMS or for whatever fill in the blank application.
• 42:36
  SPEAKER_01: But really what you need is solid enterprise level uh information architecture that spans your whole Martech ecosystem or that spans your whole product data lifecycle that spans your whole customer journey and that's what we have to tackle.
• 42:57
  SPEAKER_01: So the the piece that I'm working on most of the time is that foundational layer.
• 43:04
  SPEAKER_01: So the use case or the excuse to start these metadata and information architecture the excuse case yes the excuse case yeah the excuse case can be a point solution but the solution itself has to be something more enterprise wide and more foundational at that kind of service layer or um backend architecture layer.
• 43:31
  SPEAKER_06: Yeah it always it's always bigger than the sum of these parts that's what that's what it you know especially with AI because there's training and tuning and then there's retraining and tuning and oh yeah then there's retraining and tuning and um you know depending on what the outcome needs to be and you're spot on.
• 43:50
  SPEAKER_06: I think that's one of the common pitfalls that people don't think about is that you know it's one thing to in my mind to get a good taxonomy and an ontology about What your data is data is where it is, it's another thing to then put the pipelines together to even be able to get it together.
• 44:07
  SPEAKER_06: It's another thing to be able to train and tune on top of it, and then actually let them derive the outcomes that are changing all the time, right?
• 44:16
  SPEAKER_06: And I don't think people put that all together to say, oh, this is kind of a living breeding system that needs care and feeding, and we need to continue to help create the right input path to create the right output, you know, or outcomes, regardless of you know what those are, because it just generally, you know, it goes huge in terms of the thing, and then it comes down to a simple goal use case they want to solve, but then they don't think about the best specific iceberg on all the infrastructure that needs to be put in place to actually get it done.
• 44:46
  SPEAKER_06: I think that's is one of the common misconceptions that people have.
• 44:51
  unknown: Yeah.
• 44:52
  SPEAKER_01: So coming back to you know what we were talking about at the beginning, that the overall maturity of the market has increased.
• 45:00
  SPEAKER_01: It's still hard to sell these kinds of projects.
• 45:04
  SPEAKER_01: Um, but I think it's less hard than it used to be.
• 45:07
  SPEAKER_01: So people are interested now more and more so in building or fixing uh that technical debt and then creating that data and content infrastructure in a smarter way than they used to be, relying less on point solutions and more on uh common or info integrated frameworks to be able to support uh much more sophisticated use cases.
• 45:34
  SPEAKER_01: And when they don't have the appetite for that, they often find their way towards it through some kind of proof of concept or some kind of smaller specific use case project that ends up snowballing and turning into tackling some of these problems in order to make the original use case work.
• 45:53
  SPEAKER_01: So that kind of snowball effect, or you know, if you build it, they will come, you start uh being able to address some of those long-standing problems because you have this.
• 46:05
  SPEAKER_01: I like that excuse case.
• 46:09
  SPEAKER_06: Do you step any industries that are ahead of the curve?
• 46:15
  SPEAKER_01: Uh it depends on what you're talking about, right?
• 46:17
  SPEAKER_01: So if you're talking about chat bots and virtual assistants, I mean definitely customer service is leaps and bounds ahead of most organizations on that.
• 46:26
  SPEAKER_06: Uh if you're talking I'm just saying in just in terms of data and ontologies and stuff and getting it.
• 46:32
  SPEAKER_01: Uh Financial Services has been doing this a long time.
• 46:36
  SPEAKER_01: There are a lot of very standard ontologies for anything having to do with money transactions, certainly.
• 46:42
  SPEAKER_01: Um, we definitely see a lot of this in pharmaceuticals also and kind of health, healthcare provision, uh, you know, a lot of things around medical ontology.
• 46:54
  SPEAKER_01: Uh, I think where it's harder to do is kind of more bespoke business areas.
• 47:00
  SPEAKER_01: Um international organizations have a hard time with this as well.
• 47:04
  SPEAKER_01: Um, yeah, retail.
• 47:07
  SPEAKER_06: I think I mean, I mean, I've been in AI for a while, and I think one of the one of the things that I need to check myself on, just having listened to you guys talk through this too, is you know, is that there's I I haven't done and given it enough respect and or probably done myself some disservices in scenarios working with customers to say, Oh yeah, let's drive to the outcome, let's figure out what that is, let's work backwards, let's put these pieces in place, let's train and tune your data, let's you know, put all the models together and and you know, do some custom classifiers and all that kind of stuff, and kind of diminish the fact that part of that project is, oh yeah, why don't you guys just go get your you know, go do your data prep and clean and get it to a state where we can actually do this and and diminish the fact that that's probably that's the most important thing because everything else downstream of it will break.
• 48:02
  SPEAKER_06: And you know, that's where I think a lot of these pieces that I'm you know mentally now checking in my mind, especially having words set, where it's it's you know, I don't ever want to diminish the most important piece of actually putting it in the right structure and able to get what I need.
• 48:19
  SPEAKER_06: And I'm sure there's a lot of folks that are actually gonna be building that.
• 48:23
  SPEAKER_05: It's a good observation because again, if you start with a pilot where you have to put a lot of energy into massaging and manipulating and cleansing and filling the data, well, that's not gonna be available to you in production, right?
• 48:36
  SPEAKER_05: So we need to fix the production data in order for us to go from pilot to operationalization.
• 48:43
  SPEAKER_05: And that's where a lot of these things fall down.
• 48:45
  SPEAKER_05: I had a quick question, um, Stephanie, about your excuse cases, right?
• 48:50
  SPEAKER_05: And where are we seeing things come uh about?
• 48:54
  SPEAKER_05: What what types of things have you seen as an entry point that then has gotten people's attention and led to more um appetite?
• 49:04
  SPEAKER_01: Uh I'm definitely in a very specific area of work.
• 49:10
  SPEAKER_01: So I would say for my clients, it's mostly search, I would say.
• 49:16
  SPEAKER_01: Uh whether that's yeah, external search, uh customer-facing search of content or data or products, or internal knowledge worker search.
• 49:26
  SPEAKER_01: So that's definitely been uh the the main use case.
• 49:30
  SPEAKER_05: Everybody has problems with search, everybody says search sucks.
• 49:34
  SPEAKER_05: You know, yeah, uh remember the client that had it was the random document generator?
• 49:38
  SPEAKER_01: Yeah, exactly.
• 49:39
  SPEAKER_01: Remember the idea.
• 49:43
  SPEAKER_06: So um you're gonna you're gonna have to get on and actually um put a definition of excuse case either on urban dictionary or as a link right below our podcast.
• 49:53
  SPEAKER_06: Yeah.
• 49:54
  SPEAKER_06: You know what, or maybe the title of this podcast should be you know how Stephanie helps dispel what a you know the excuse case in any organization or something.
• 50:06
  SPEAKER_05: Find it the right excuse case.
• 50:12
  SPEAKER_05: So um yeah, interesting.
• 50:14
  SPEAKER_05: You know, it it it is such an uh a fascinating space.
• 50:18
  SPEAKER_05: There's huge need, but there's so much, you know, confusion, there's so much noise.
• 50:22
  SPEAKER_05: You know, I think uh a lot of companies try to go with the big vendors that um, you know, the typical suspects, usual suspects, right?
• 50:31
  SPEAKER_05: The Deloitte digitals and the IBM Global Services and the PwCs, and they certainly have uh the accentures of the world, they certainly have a lot of resources, but sometimes it's not the right resource for the project.
• 50:44
  SPEAKER_05: Do you have any thoughts about um the value of a large global integrator or vendor versus you know companies like yours and mine, you know, where we're smaller but uh very focused and specialized.
• 50:57
  SPEAKER_05: Uh what are your what are your thoughts about how we how we are different, how organizations like ours are different?
• 51:03
  SPEAKER_01: I'm sure my thoughts are extremely biased.
• 51:06
  SPEAKER_01: Um, but I think um kind of more niche uh or focused consulting firms such as yours and mine, I think uh put a lot of attention on uh knowledge, knowledge modeling and understanding expertise and understanding knowledge workers and their needs and user experience and uh those kinds of deep concerns about how people think and how people work.
• 51:37
  SPEAKER_01: Uh, what I do appreciate about larger organizations like your IBMs and and Beloites is certainly the very strong technical expertise and of course the sheer amount of uh person power that they can throw at large, large problems.
• 51:54
  SPEAKER_01: Yeah, exactly.
• 51:55
  SPEAKER_01: So I often work in tandem with these organizations on larger deployment projects where we're zoomed in on metadata and taxonomy and usability kinds of problems.
• 52:10
  SPEAKER_01: And those feed into these larger uh technical solutions, certainly.
• 52:16
  SPEAKER_05: Uh I agree with you, Stephanie, and you know, this deep specialization and expertise in the data, metadata, information architecture, usable usability, you know, how to drive those things dynamically and how to use data to support your decisions all along the way through use cases and iterative testing and uh and uh you know conference room pilots and tree jack testing and information prototype.
• 52:44
  SPEAKER_05: That's the stuff many times when people look at these messes, they don't want that anyway.
• 52:48
  SPEAKER_05: You know, it's like, okay, yeah, uh, yeah, yes, you handle that, and then we'll, you know, we'll do the we'll we'll build based on your blueprints.
• 52:56
  SPEAKER_05: So, you know, that's been a critical role that we've played for large, large uh projects like that.
• 53:00
  SPEAKER_05: So it sounds like we're definitely in alignment with that.
• 53:04
  SPEAKER_01: Uh, we often come ahead of those larger projects uh because an organization has correctly pinpointed that oh, we should probably get our metadata act together before we start this big rollout of a global dam or a global this or that platform.
• 53:20
  SPEAKER_01: And so we will often have a little bit of a head start, do this very focused work, and then hand that into one of these larger uh tech projects.
• 53:30
  SPEAKER_01: Absolutely.
• 53:32
  SPEAKER_06: One quick thing I want, yeah, I wanted to ask you your opinion on, especially um you know before we end and stuff.
• 53:38
  SPEAKER_06: So, you know, there's there's a lot of emphasis.
• 53:42
  SPEAKER_06: I think the question is going to stem from first, you know, how in this world where we need more comma, better skilled people thinking about these types of problems and projects and things.
• 53:59
  SPEAKER_06: And so the first part of the question is, you know, if somebody was interested in in how to get started in, you know, in you know, as their role to build ontologies, to help organizations with data preference stuff.
• 54:12
  SPEAKER_06: One, what would you recommend?
• 54:14
  SPEAKER_06: And two, you know, part of this for me is I really, really love the the women in tech type of perspective, and I'm so grateful for your perspective because I also have daughters that I'm trying to help, you know, get into technology and help them understand that there's so many thousands of things that they can do.
• 54:32
  SPEAKER_06: And you know, Seth's daughter is you know one of my colleagues, um, you know, with Salesforce and things like that.
• 54:38
  SPEAKER_06: And it's just, you know, there's so many really brilliant opportunities.
• 54:42
  SPEAKER_06: And so, one, you're you know, I love that your bias is what it is.
• 54:47
  SPEAKER_06: And two, you know, how how would you rec or what would you recommend somebody you know do to get started and things like this because it's such a critical role.
• 54:54
  SPEAKER_06: Definitely there's a lot intact in there.
• 54:55
  SPEAKER_06: So I apologize, but you know, but it's still something, you know what I mean?
• 55:00
  SPEAKER_01: So yeah, absolutely.
• 55:01
  SPEAKER_01: And I have a a bit of an interesting path into that because I come from library school of all places, which you know, no fault of its own, but it's a very female dominated uh, or at least it was at the time, 15 years ago, uh, was a very female dominated space.
• 55:18
  SPEAKER_01: Uh, and as it was transitioning over to information science outside of library cataloging, I think it had to do a lot of introspection about what is the role of uh the classical librarian uh and information specialist in in the new era.
• 55:37
  SPEAKER_01: And, you know, I can't say it was the perfect education 15 years ago, but those schools have very much modernized.
• 55:46
  SPEAKER_01: And there are now new types of schools that are more information science focused.
• 55:52
  SPEAKER_01: Um, so if you're young or uh looking to go back to school, there's certainly a lot more programs.
• 56:00
  SPEAKER_01: There's actually programs specialized in digital transformation in asset management and all kinds of new disciplines that didn't exist 15 years ago.
• 56:09
  SPEAKER_01: Uh if you're not interested in going back to school, however, there are so many self-learn opportunities.
• 56:15
  SPEAKER_01: Uh, I mean, Seth's book is a great way of understanding and teaching yourself strategy stuff if you're at a more senior level of the organization.
• 56:24
  SPEAKER_01: But if you're interested in learning knowledge modeling or RDF, any of those kind of more technical capabilities, there are tons of little certifications or training courses that you can take that are available online and kind of build up your own capabilities.
• 56:40
  SPEAKER_01: And there's tons of open source uh tech that you can use to work on those uh on the side as well.
• 56:49
  SPEAKER_05: Absolutely.
• 56:50
  SPEAKER_05: And by the way, um by the way, Stephanie, uh Rachel now works for Deloitte.
• 56:56
  SPEAKER_01: Oh, nice, nice, yeah.
• 56:59
  SPEAKER_05: She took her Salesforce uh learnings uh and taught herself uh lots of different things, got to work worked uh in financial services for a while, and and now she's uh Deloitte and um specializes in Salesforce.
• 57:12
  SPEAKER_05: So I'm very proud of this wonderful job.
• 57:14
  SPEAKER_03: Fantastic.
• 57:15
  SPEAKER_05: If anybody wants to connect her, she's Rachel Benson at uh Deloitte.
• 57:19
  SPEAKER_06: So uh reach out and connect with her and yeah, so hey, uh so Stephanie, what uh what do you what are you looking forward to in 2022 for yourself and for you know any goals or anything?
• 57:34
  SPEAKER_01: I personally or professionally, it doesn't matter.
• 57:36
  SPEAKER_01: Yeah, I actually miss airports.
• 57:39
  SPEAKER_01: I haven't been in an airport.
• 57:40
  SPEAKER_01: I know you're in one right now, but I haven't been in an airport since the very beginning of 2020.
• 57:45
  SPEAKER_01: And I have airport dreams, so I definitely missed it.
• 57:48
  SPEAKER_06: I'm gonna tell you.
• 57:50
  SPEAKER_06: It's it's uh it's it's probably gonna go back, you know, like um, you know, sometimes I feel like uh, you know, rolling through the airport with all the stress, I mean, would give me facial ticks.
• 58:00
  SPEAKER_06: Um, and it may, you know, that experience living this dream, it's it's it can be super frustrating, right?
• 58:06
  SPEAKER_06: I'm hoping that we lift the mask mandates and stuff, but what else in terms of besides travel?
• 58:13
  SPEAKER_06: That's a big one for me, too.
• 58:14
  SPEAKER_01: Yeah, besides travel, I don't know, just uh being able to uh enjoy my work.
• 58:19
  SPEAKER_01: I think the pandemic really changed the relationship a lot of people have with their work.
• 58:25
  SPEAKER_01: Uh and while I miss traveling, traveling and seeing my clients face to face, uh, I've had a number of really long-term projects thanks to the pandemic where I've been able to feel a little bit like part of the staff.
• 58:37
  SPEAKER_01: So that's been really nice.
• 58:39
  SPEAKER_01: And I hope to continue that vibe in 2022, even though I may go back to a little bit more of the chaotic consultant life.
• 58:48
  SPEAKER_01: Um, I've definitely learned to appreciate uh some of the changes to the way work is being conceptualized.
• 58:54
  SPEAKER_05: Yeah, I'm I'm looking forward to more things to face groups on at with the clients and working sessions and collaborations and all of that.
• 59:02
  SPEAKER_05: It's hard to do on Zoom sometimes.
• 59:05
  SPEAKER_05: But hey, this has been uh we've really come to the uh end of our time.
• 59:09
  SPEAKER_05: It's been a real pleasure talking with you today, Stephanie.
• 59:12
  SPEAKER_05: It's it's I I think back and uh miss uh our work together, and I have such fond memories of having you as part of the uh the early team.
• 59:21
  SPEAKER_05: And uh before I close, I just want to remind everyone our sponsors, uh Simpler Media Um and the Marketing AI Institute and uh Early Information Science.
• 59:32
  SPEAKER_05: And I want to encourage um, oh, do you have anything that you want to share, Stephanie?
• 59:36
  SPEAKER_05: Anything coming up, any events, any anything uh you want to share with folks?
• 59:41
  SPEAKER_01: Uh yeah, for those who are those who are in the space, uh we have the taxonomy boot camp call for proposals that just went up.
• 59:50
  SPEAKER_01: Uh, we are hoping to have an in-person conference this year.
• 59:53
  SPEAKER_01: So if you're interested in taxonomies and graphs, uh we do a ton of that uh and that's in November in DC every year.
• 1:00:00
  SPEAKER_06: Wonderful.
• 1:00:01
  SPEAKER_06: Thank you.
• 1:00:01
  SPEAKER_06: We really enjoyed it.
• 1:00:02
  SPEAKER_06: All right, thanks for the time and and all the great wisdom.
• 1:00:04
  SPEAKER_06: Appreciate it.
• 1:00:05
  SPEAKER_06: Have a great day.
• 1:00:06
  SPEAKER_01: Have a great day.
• 1:00:07
  SPEAKER_01: Bye.