Podcast: Play in new window | Download
Subscribe: Apple Podcasts | Google Podcasts | Spotify | Android | TuneIn | RSS
Seth EarleySeth Earley helps enterprises use ontologies to power the intelligent content experiences that create personalized content interactions and that run chatbots, voice assistants, and other new technologies.
An ontology is a business practice that helps you understand the knowledge in your business and connect and share it in new and powerful ways.
We talked about:
- his consultancy, Earley Information Science
- what an ontology is and how to use ontologies to describe a domain of knowledge
- how an ontology can help traverse the knowledge in an organization
- the difference between “is-ness” and “about-ness”
- the importance in ontology practice of starting with concepts that are important to the business
- how an ontology can help content strategy practice
- how these technologies and practices can help understand user intent to deliver the correct content
- the importance of structuring content to avoid TL;DR situations
- how ontologically organized structured content can help deliver personalized content
- the arrival of conversational cognitive assistants
- the role of knowledge practices to power new tools as the workforce changes
- the importance of understanding – and conveying to leadership – the importance of ontology work
Seth’s bio
An expert with 20+ years experience in Knowledge Strategy, Data and Information Architecture, Search-based Applications and Information Findability solutions. Seth Earley has worked with a diverse roster of Fortune 1000 companies helping them to achieve higher levels of operating performance by making information more findable, usable and valuable through integrated enterprise architectures supporting analytics, e-commerce and customer experience applications.
Seth is a sought-after speaker, writer, and influencer. He is the author of “The AI-Powered Enterprise” from LifeTree Media. In 2021, the book received the Axiom Business Book Silver Medal in the Artificial Intelligence / Robotics / Algorithms category.
His writing has appeared in IT Professional Magazine from the IEEE where, as former editor, he wrote a regular column on data analytics and information access issues and trends. He has also contributed to the Harvard Business Review, CMSWire, CEOWorld, TechTarget, eCommerce Times, Analytics Magazine. Journal of Applied Marketing Analytics, and he co-authored “Practical Knowledge Management” from IBM Press.
Connect with Seth online
Links and resources mentioned in the interview
- The AI-Powered Enterprise: Harness the Power of Ontologies to Make Your Business Smarter, Faster and More Profitable, Seth’s book
- EarleyAI, Seth’s podcast
- There’s No AI without IA article
Video
Here’s the video version of our conversation:
Podcast intro transcript
This is the Content Strategy Insights podcast, episode number 116. As companies create personalized interactions for their customers and as they implement new technologies like chatbots and voice assistants, they are discovering that they need new business practices to manage these experiences. One of the most powerful ways to organize and structure the content that drives these new interactions is an ontology. Seth Earley can help you understand what an ontology is and how it can help your company’s content efforts.
Interview transcript
Larry:
Hi, everyone. Welcome to episode number 116 of the Content Strategy Insights podcast. I’m really happy today to have with us, Seth Earley. Seth is the founder and CEO of I’m sorry, of Seth, of Earley Information Science and information architecture and other kind of consultancy. Welcome Seth, tell the folks a little bit more about what you do there at your consultancy.
Seth:
Well, thank you for having me. It’s great to be here. We’ve been around for 20 plus years, 25 plus years. I don’t want to date myself too badly, but been around for a long time. I started when I was 12 just so you know, just the 30 years in business doesn’t scare you.
Larry:
The math checks out.
Seth:
Actually, I like to say … Anyway, I’m much younger in base 12, whatever my age is.
Larry:
Well.
Seth:
It goes along with another joke, but that’s okay. We’ll skip that. I’ve been around for quite a while, been working with Fortune 1000 companies. Over the years, we have about 50 consultants and everything we do is about information, making it more findable, usable, and valuable. That includes doing things like working with large product catalogs and e-commerce sites it entails work around knowledge management, knowledge architecture, knowledge engineering. It entails work around content operations, especially componentized content and content reuse … Sorry, excuse me. It entails pretty much the range of things that you would do with content and data and information and making it more accessible to people and applicable to whatever problem you’re trying and solve.
Larry:
Nice and I think that’s a swath of information that’s of great interest to my listeners. One thing, you wrote a book recently, it’s called, look at the title, the AI Powered Enterprise and the key to that is … Oh, there, thank, thank you. Yeah.
Seth:
Anybody not just listening, if they’re watching.
Larry:
Yeah, no, there will be a video so folks will get to see that. I’ll also put a link to the book in the chat. The first couple chapters of that book you talk about, well, first, like just the benefits of these kinds of technologies. The second chapter you talk about ontology and that’s what I really would like to focus the conversation on today. Because that’s such a powerful organizing scheme for modern information. Can you just describe for our folks what ontology is?
Seth:
Sure. An ontology describes domain of knowledge. If you’re in life sciences or pharmaceuticals, it consists of multiple taxonomies. Everybody’s familiar with taxonomy and controlled vocabulary, well, you might have a list of chemical compounds and you have another list of generic names of compounds. Maybe another list of brand names and commercial names. Well, maybe they’re all representing the same things so they’re actually non-preferred terms, you’re building up the thesaurus structure. Because you have preferred terms, non-preferred terms and so on, and alternate terms, temp, chemical terms, brand names, et cetera. Then, you also have things like mechanisms of action. You have drug targets, you have diseases, indications, treatments. You have everything for a life sciences firm. It’s also going to include, go to market strategies and regions in which they work and key opinion leaders, and practitioners specialists.
Seth:
All of these vocabularies, all of these organizing principles, all the buckets in which you would put information, you never create a single taxonomy. You always create multiple taxonomies when you’re doing any kind of information architecture or data strategy. Essentially what that means, is the multiple taxonomies, with the relationships between them. Because again, you can have a thesaurus structure, that says, here’s a preferred term and a non-preferred term, those are synonym relationships. You can also have a related term. You can say it’s kind of the “see also” term. You could have diseases and indications for those diseases. You can have diseases and treatments, right? Treatments for those diseases. You can have drug targets and mechanisms of action. It’s the relationships between those. Here’s the mechanism of action assigned with this drug target, which is related to this generic compound, which is related to this brand compound.
Seth:
If you want to gather information throughout the organization on say, performance of that product, or market share, or pricing, we have a way of connecting those different things together. We have one ontology that we did for a life sciences firm that talked about all of, taking a generic compound and then all of the different brand names, internationally and US, and then all the formulations. If you wanted to get, say the average pricing for a compound, you had to know all those things. You had to know what it was called in those different markets and those different regions. Then, you have to understand what the sales volume is. The point here of the ontology in that context is to understand, market share and revenue and pricing. Because you have to map all those things together. In any kind of an organization, we have those types of relationships.
Seth:
We may have products and we have services. We have services that go with that product. Well, we have problems and we have solutions and here’s the solution to that problem, right? Everything is related conceptually. Anything you can think of that’s related conceptually, can be in your ontology. That’s when you get you play the game of six degrees of separation from Kevin Bacon, with IMDB, the database, and you say, “Okay, what are the movies that Kevin Bacon has been in? What are the other actors that have been in those movies? Now, let’s find a connection between that actor and another actor and a movie that Kevin Bacon is in.” That’s an ontological relationship. It’s an ontological structure. You can think of it as all the sets of organizing principles to describe that domain of knowledge and the relationships between them.
Seth:
This is where you start to say, what is that thing? How do we define it? That is the knowledge infrastructure. That is the knowledge scaffolding of the organization. It’s all the buckets in which we can place our information and our content. You could say it’s a chart of accounts for knowledge. Once, we did a work of project that’s in the book, for applied material many years ago. The senior finance person said, “Well, why do we need ontologies and taxonomies? Why don’t we just get Google?” I said, “Well, do you have a chart of accounts for your finance organization?” He said, “Of course I do.” I said, “Well, why don’t you get rid of your chart of accounts and just get Google?” Because a taxonomy is a chart of accounts for knowledge. Again, in the case of applied materials, it was fabrication plans, and equipment types, and fabrication techniques, and regions. I’m just making these up right now, but they’re about 30 different sets of organizing principles that described the knowledge of that enterprise.
Seth:
Once we had those relationships, it was much easier to go in and say, use a part number to find all the conversations about problems with that part, or that maintenance. To enter in an image and find out what that image was, what type of part, what type of assembly, what type of subassembly, how it was used? What the level of inventory in stock was based on the ERP integration? The ontology gives you this mechanism to traverse the knowledge and information and data of the organization. We do it that way when we have that ontology, which think of that as that structure, that knowledge scaffolding. When we have the data and we can access the data, that’s actually a knowledge graph, right?
Seth:
Graph data consists of the ontology and an access mechanism to the data. Now, so what that does is it allows you to do faster reporting, to find relationships that were not readily available to do information queries, that were not anticipated. It allows you to leverage lots of different data sources that may have different descriptors because you’re mapping those together. It really gives you a tremendous amount of flexibility, agility, extensibility, when you build these things correctly. Correctly means looking at the different dimensions of your information. We’re doing a project right now for a financial services company, the content side. Essentially, we need to understand all the different types of thought leadership and all of the different themes within that thought leadership and the topics. Then, the experts who produce that, and then the audiences for that. Then, all of those relationships allow the ability to go in, do a search for a concept, and then use that concept as a jumping off point to get to other concepts and other information. You might hear a little bit of background noise, both of my cats decided that they were going to help today with the webinar, but they haven’t come by yet. They’re making-
Larry:
Well, I’ll just point out, it’s not an official internet event until a cat shows up so thanks.
Seth:
They should show themselves, they should make some noise. That’s right.
Larry:
Hey, Seth, you just said, there’s so much in everything you just said. Two things I want to come back to. One, is that you mentioned part of the process of building ontology, is what is that thing that you’re talking about? Google famously described, when they introduced their knowledge graph, this notion of things, not strings and the importance. I think a lot of content people are still strings people. We’re still thinking about the content the strings of characters that make up content. You’re talking about things. Can you talk a little bit about that, distinguish?
Seth:
Sure, sure, so when you think about a piece of information, you kind of have to say, well, what is it? What’s the “is-ness”? We were actually just awarded a patent for one of our approaches for leveraging ontologies and chat bots.
Larry:
Cool.
Seth:
Then, all these offers for like plaques and things like that. I’m like, “Okay, so what is this?” This thing is a brochure. What is it about? It is about an award for a patent, right? The is-ness is a brochure. Well, how do I tell a thousand of these brochures apart, a thousand of these things apart? I do that with about-ness. The is-ness is brochure. The about-ness is patent award, right? We can take anything like that. This is a book. It is about information management. It is about AI. It’s about ontologies. If we have a contract and we have a thousand contracts that we have to tell apart, how do we tell them apart? We tell them part by contract type, by customer name, by region, by all sorts of descriptors. Those descriptors become the about-ness. The about-ness, that’s describing the entities and the objects. Then, within those entities and objects, we need to define all those organizing principles. Now, people would say, “Well, doesn’t that sound like master data or isn’t that just content architecture? Why don’t we just start there?”
Seth:
My answer to that is, it’s not exactly the same, because if you start at the master data level, you’re diving right into the data. You’re saying, “Okay, this is an organizing principle. It’s a data field, or it’s a value within the field.” Excuse me, and this cat’s running about they haven’t made their appearance yet, but they’re making themselves known. Come here bud.
Larry:
Well, they have to come on the air at some point.
Seth:
You got to see them, but they’re busy torturing each other. Maybe if I get a laser pointer.
Larry:
Nice. We’ll go back. One thing I want to interject real quickly, when you talk about about-ness, that sounds like metadata. Is that?
Seth:
Yes.
Larry:
That’s the term that most of my folks-
Seth:
Yeah, it’s metadata. It is metadata. Let me give you another example.
Larry:
Yeah.
Seth:
Let me give you another example. The reason why we don’t want to start with metadata, or just taxonomy, or just master data, right? The reason we want to start at that level is, you want to start at the conceptual level, what’s important to the business? Then, where does that show up? Where does that concept show up? Concept can show up in lots of different ways. A lot of people used to think that taxonomy was the same as navigation, right?
Seth:
They’d say, “Oh, we’ve got to build the taxonomy. We think of a hierarchy.” That was a navigational hierarchy but it’s much more than that. Because if you think about something, a collaboration technology, like SharePoint. I know everybody’s into Teams, but SharePoint is a good example because if you’re a consulting firm like us, we have methodologies, right? A lot of firms have methodology, manufacturing methodologies, consulting methodologies, problem solution methodologies, troubleshooting method, whatever. The issue is a methodology is a concept. You could in SharePoint have a whole site collection called methodology. Within that, there could be other sites about different types of methodology. You could also have, I hope you can take those out in post-production, my coughs [my Fiverr helper did his best 🙂 ]. Then, you have the site itself, which could be called methodology. Within that, there could be libraries. You could have a library called methodology and within that, you could have artifacts. You could have a list within the library called methodology. You could have a content type called methodology. You could have a metadata field called me methodology, or you could have a term within a metadata field as a controlled vocabulary item called “methodology.” Methodology as a concept needs to be translated into lots of different structures that are not just navigation and not just classification. They can be workflows. They can be other organizing principles. They can be bigger constructs, they could be processes, right?
Seth:
There’s a lot of stuff – and that’s the cat trying to make background. We’ll cut that. We’ll take that out in post-production too. The point here, is that starting at that data level actually loses a lot of that content. What we want to do is start with concepts and then decide what those important concepts, get agreement on those. Then, start thinking about how those get designed into multiple downstream systems within your infrastructure. Does that make sense?
Larry:
That makes perfect sense. One thing as you talk and hearkening back to something you said earlier, I think, I gather, that a lot of your work has been in technical communication. You mentioned earlier componentized content and things like that. In that world, the DITA world, there’s the notion of topics. Then, in the ontological world, there’s the notion of concepts. How do stitch together?
Seth:
Well it’s interesting. We did a webinar recently about building componentized content for chat bots, to power chat bots. Topics and tasks fit very well into that. Because you can actually have a mapping of topics to the organizing principles that are important for your chat bot. Without getting into too much detail, the answer is it’s incredibly important and topics do map into taxonomic structures. We have other types of identifiers and we have other types of metadata that gets surfaced. Again, one of the webinars that we did, we did a deep dive into that around, again, componentizing content, and then automatically ingesting that into a chat bot. You still have to think about the additional metadata around topics. Because a topic in and of itself can have additional descriptors, right? Because a topic can also have the product that it’s associated with. Maybe there’s a troubleshooting code. Maybe there can be other metadata associated with that. To retrieve that topic in the right context. Usually it’s already in a hierarchy, but you want to be able to retrieve that topic separately and out of the context of the hierarchy. Excuse me.
Larry:
Yeah and that, I think, and what you just said also reminds me of another big concept that’s in the air these days, this notion of decoupled-ness. Separating even, it’s so meta because it applies at every level here, it’s like separating concepts from topics, from things and strings and even like the individual details in all the implementation of this, is that decoupled-ness, is that like even metadata to identify what a concept is and help you stitch it back to some other concept. It seems like a lot of this and the way other thing you’ve said is about how this spans, not like just content. Because that’s what my folks are most concerned about, but also the way you just effortlessly integrate all of that with like business objectives, and metrics, and analytics, and business goals, and domain knowledge and all that stuff. Tell me a little bit more about how an ontology, a little bit more into the mechanics of how the ontology stitches all that together in a way that helps you get better create and manage better content?
Seth:
Yeah. Well, it does get a little bit into the weeds and so a little bit hard to explain without diagrams and book charts and so on. At the end of the day, you want to think about a piece of knowledge or piece of information, not only in the context of the overall structure of a document, so that you can shuffle in and out pieces to do translations and updates and so on. You also want to think about the identifiers and the handles of that content so that you can pull it out into another context, right? There should be enough descriptors on that piece, on that component, so that I could say, “Okay, this component can answer this particular question, right?” People don’t want, when we were doing work for a large insurance company, in their call center, basically they did not want, when they did a search on anesthesia claims, they get a hundred results. One of those results would be a 300-page document, right? The retrieval of that, you didn’t want that 300-page document. You wanted to understand how to process that anesthesia claim. You needed to know, well, what type of anesthesia claim that was and what the coverage principles were. That was a little piece of content. That was a small paragraph, right? We need to think about the granularity of the context of that paragraph and be able to structure the metadata around that.
Seth:
Now, again, Darwin Information Typing Architecture data is extensible. It’s adaptable. You evolve it. That’s why it’s called Darwin information typing because you can adapt it. You can build custom metadata around that, for audience, for task, for troubleshooting code, for product, or whatever. What you’re trying to do is you’re trying to say, let’s give this piece of this component enough of the identifiers that we understand its context and the task it’s supporting and the problem it’s supporting, or the solution it’s supporting, and be able to pull that out by identifying an utterance and the intent from that utterance. Think about it, when you work with virtual assistants and chat bots, what you’re trying to just take phrase variations of how people would describe their problem. You’re trying to classify that to an intent, right? Utterances, or like, “Geez, I can’t, my password isn’t working or my ID is locked out, or I forgot my password, or my computer’s mad at me.” I’ve actually did a workshop and I said, “Think of all the different ways you can phrase, I can’t get in my computer, you know?” It was crazy. All those variances, all those different utterances boiled down to an intent of change my password. It’s much more complex than that because intents are multidimensional, right?
If you just try to classify an utterance to a single thing, you end up with like a pre-coordinated taxonomy. You’re saying, “Well, what does this thing mean?” “Well, it means this one thing.” Well, it means more than one thing, because if I say I need claims coverage, I need coverage or conditions for a type of business, an employment agency in Massachusetts, covering liability for employee actions, right? There’s about four different entities in there. There’s a type of business. There’s a type of claim. There’s a type of coverage. There’s a state, right? All of those become facets in which you can retrieve that very specific piece of content. That’s how we think need to think about component authoring, not just the context of that document for translation and localization, but the context of that document for standalone retrieval in the context of a specific problem that you can describe across multiple dimensions. That you can describe in a way that gets you zeroed into that specific piece of content out of that ocean of information. All of these content objects. Now, the old days, it was monolithic documents. You’d have to tag across multiple dimensions.
Seth:
Well, that doesn’t work. I don’t want a 50 page document. I don’t want a hundred page manual. The saying is TL didn’t DR. Right too long didn’t read, right? You people say RTF, I read the freaking manual and people’s response is TL:DR too long, didn’t read. RTFM TLDR. You’ve got to keep those two things in mind and say, “What is exactly that information I need to get to this user in this context or this purpose?” That is what is good about AI driven content. You get organizations that are starting to say, “Oh, we have an AI content group.” No, it should just be a content group. It’s not AI content. What is valuable about that, is the fact that they’re thinking, very specifically, about the user, the use case, the task and the context, and building just that one piece of content to answer that question.
Seth:
Now, I was trying to activate a credit card and I pulled off the sticker and threw it away before I did it. I thought I’d find it someplace else. I’m looking and looking, looking, I’m getting pages and pages and pages of content about credit card activation and all this nonsense about it. All about credit card this, and credit card that, and how we’re protected and how do I activate the fricking thing? Where’s the number for that? I could not find it in 10 pages of content. I forgot where I found it, but it was like horrendously difficult. I searched for activation. I searched for, I forgot what I searched. I did all sorts of searches and I couldn’t get that number to activate my credit card. Now, what is the purpose of all that other content that is associated with credit card activation, if it doesn’t tell me how to do that? That is the value of AI power of AI driven content. Because we’re trying to take a bot, we’re trying to use that bot for a very specific audiences, specific set of use cases. That’s where we need to think about content in a much more precise way.
Larry:
Yeah. The way that whole series you just set out, that alignment of the user, the use case, the task, and the content. That sort of gets at like what you were just talking about, the specific example of chat bots, but also generically, any kind of personalization, delivering personalized content.
Seth:
Absolutely.
Larry:
Yeah.
Seth:
It’s no different, it’s no different. When you think about personalization and you think about a chat bot, they’re the same mechanism, right? You’re taking a signal. The signal in the case of a chat bot, is an utterance. You’re interpreting that and you’re responding. You’re predicting the answer. You’re making a prediction, you’re personalizing, right? You’re saying it’s a signal and a response. The richer the signal, the more details about the user that I have, understanding their use case, their persona, their task, their objective, I can predict the content that they need. Well, that doesn’t matter, whether it’s on a webpage, because I’m getting signals, I’m getting their digital body language, I know who they are. If they’re authenticated, something about what they have, that’s why first party data is so important.
Seth:
I collect those attributes and those identifiers and that data and that digital exhaust from all those other systems that they interact with. I consolidate with the customer data platform and I use that to inform the content, the products, the next best offer the next best action for that user in that context. That can be the same thing as answering a question, right? If they’re doing a search, if they’re navigating, if they’re downloading a white paper, if they’re looking at certain product, those are all signals. Then, I want to respond to those signals. Those signals are metadata, right? We respond by reading that metadata and aligning that metadata with the content, with metadata on our content, right? It’s a matter of processing those signals and aligning those with the content.
Seth:
Now, there’s also, there’s work that we had done for a large global technology firm. They handle over 10 million knowledge transactions per day. In other words, they’re answering questions. They’re serving content once. I’m sorry, publishing content once, consuming it everywhere.They are very advanced with component authoring. They’ve doing this about 10 years, based on work that we started with them almost a decade or ago. They said, They have operationalized this and they do this without armies and armies or content creators and content managers, website managers. They have saved hundreds of millions of dollars per year on content operations by doing this.
Seth:
This also sets them up perfectly for being able to build cognitive assistance, high functionality, cognitive assistants. Those cognitive assistants are going to be the way organizations do their business. There’s no other way, because as we look at lower, we’re constantly reducing costs and we’re trying to provide better low as a customer service. We can’t afford to scale the human expertise. We have to capture, codify, automate, right? That’s just what’s going to happen. In the chapter in my book, where I talked about Alan Perkins and how he goes about his day, interacting with virtual assistants all day, every day, that’s what our world will be, right? They’re going to be conversational and they’re going to work really well. Right now they suck. Between where we are today, where they are really bad, and where we’ll be, we know in the next several years, that’s the world. The world will be a world of conversational assistants, intelligent virtual assistants that are very much analogous to talking to people. They’re not going to be the same. They’re not going to think, but they will model that. They will replicate that. They will synthesize that. They will simulate that.
Seth:
There’s got to be human in the loop, but between where we are today and where that future will be in just a few years and it’s going to accelerate, is going to make some organizations obsolete, just like dot-com boom, and the internet vaporized. It’s my favorite term around that by a book by a colleague of mine, excuse me, called Vaporized, which talked about how the internet has swallowed up all these organizations. You see the same thing with cognitive assistants, right? Because you won’t be able to afford all those humans and those highly trained humans, and a lot of the expertise is aging out of the workforce. You don’t get people coming up as field engineers and factory maintenance guys, and factory engineers, and production engineers over 30 years, that knowledge is not being created in the ranks. Organizations are going to face this. We’re going to have this deluge of need, tsunami of need around knowledge architecture, knowledge codification, knowledge capture componentization so that we can power these tools. That is why component authoring and content is so critical in thinking about this in that ontological framework, that ontological construct.
Seth:
In the future, we’re going to have that capability and the companies that don’t have it and the ones that are not starting today or not don’t have a handle on it, or the knowledge is out of control. They haven’t even thought about it, it’s not on their radar, they are going to be flatfooted. They are going to be severely … still motor on because the size, and market reach, and distribution channels of brand strength, but they’ll erode. They’re going to erode. Their costs are going to go up. Their revenue’s going to go down, and eventually they’ll go away. There’ll be lower cost, more agile, born digital producers who understand these principles, and apply them, and implement, and operationalize and execute just like this large global organization. Boy, I almost saying their name. We got to cut that right out.
Larry:
No. I will. Hey Seth, I can’t believe this. We’re already coming up on time. You’ve said three things in the last five minutes that I want to do a whole other episode about, but is there anything last today? Is there anything, little last, little quick tidbit you’d like to leave the listeners with before we wrap up?
Seth:
Well, it’s a journey. I think if you’re not on this journey, you’ve got to start. If your clients or organizations you work for, don’t understand this stuff, invite me in to do executive briefings. I’d be happy to do that. It’s really important to get this message across to the C-suite. It’s really important to get this message to leadership because it is going to be existential. There’s a duty that the people on this webcast, this podcast have, as experts as professionals in this industry, you have a duty to get this on the radar of your leadership. It’s not a matter of, don’t be timid. You’ve got to be aggressive. You got to say, “Look, this is important.” If they don’t understand it, find the things that they need to see and learn, and read, and hear to make them understand it.
Because the people on this call are going to have jobs, right? The executives that don’t take this seriously will not have jobs. I have seen it. I have seen it multiple … I’ve seen it happen where organizations did not … I have a story about a publisher that lost their entire K through 12 textbook market because with another competitor using componentized content, guess what? That was like 10 years ago. Okay. The publisher had to get there. They couldn’t get there in time and they lost that market. That’s going to happen over and over and over again, in the industries that we see today. As these tools accelerate and component content, content management is more critical than ever. And taxonomy and information architecture, despite people saying, “You don’t need it, AI will take care of it.” That is the farthest thing from the truth. Read my article. There’s No AI without IA. You do a search for that and find it, of course, buy my book and read that too.
Larry:
I’ll link to both of those in the show notes. Hey, one very last thing, Seth, what’s the best way for folks to stay in touch to follow you on social media or online?
Seth:
Oh sure, so you can go to the website and subscribe to a newsletter www.earley.com. Now don’t forget the E before the Y. E-A-R-L-E-Y.com. You can also connect with me on LinkedIn. I’m Seth Earley, S-E-T-H E-A-R-L-E-Y and I’m also on Twitter @SethEarley. Just first name, last name. My email is Seth at earley dot com. Just my first name and last name. Just don’t forget the E before the Y.
Larry:
Got you.
Seth:
I have a podcast called EarleyAI, E-A-R-L-E-Y. We have some great guests from the world of machine learning and artificial intelligence. That’s a great show that I do with Chris Featherstone and so that’s another place too.
Larry:
Yeah, I’ve listened to several episodes. I love that podcast. Well, thanks so much, Seth, really enjoyed the conversation.
Seth:
Thank you. Thank you for having me. Apologies that the kitties made noise, but didn’t show their faces.
Larry:
No, we love our cats on the internet.
Seth:
All right, thanks Larry.
Larry:
Thanks Seth.
Seth:
Okay. Bye now.
Leave a Reply