Skip to main content
December 11, 2023

ContentOps 2024 Boom or Bust? (webinar)

Scriptorium principals Sarah O’Keefe, Alan Pringle, and Bill Swallow have decades of experience in the content industry. In this webinar, they share their analysis of key content operations trends.

After watching, you’ll learn

  • Three key trends in content operations
  • The predicted impact of these trends in 2024
  • How your organization should adapt

Related links



Scott Abel: Hello, and welcome to Content Ops 2024: Boom or Bust? Welcome to our show. We’re going to have our host of Let’s Talk Content Operations Sarah O’Keefe lead a panel discussion where we’ll talk about the content operations topic and the trends for 2024. My name is Scott Abel, and I’ll be the host of today’s show for just a few moments until I transfer it over to Sarah. If you’re new to the BrightTALK platform, let me tell you a few things about participating today. One thing is a concern you don’t have to worry about because we don’t have access to your camera or your microphone, which means we cannot hear or see you during today’s broadcast. But if you’d like to be heard, you can ask a question at any time by using the Ask a Questions tab located underneath your webinar viewing panel.

You should also know that we’re recording today’s show and that you’ll be able to watch a recording of this show on demand anytime you’d like. After the show is over, you can use the same URL that you’re using today to watch the live show, to watch the recording, and you can share that link with others who might like to see the show after you’ve done it, and we encourage you to do so. There’s some content in the attachments section of your webinar viewing panel that could prove useful today. There’s contact information for some of our guests on the show today, as well as information from our sponsor and some resources that will prove handy if you are interested in the topic of content operations.

So definitely meander over to the Attachments tab today and see what’s available there for you to download. We’ll also be asking you several polls during today’s show. First, I’ll be launching the first poll right now. Our polls are super easy to participate. They are multiple choice questions. You simply navigate to the polling feature and click on the answer that is best representative of what you think, the answer you would like to give. That will be added to the poll and our presenters will then see the cumulative totals and be able to address your concerns. More specifically, knowing a little bit more about you, our first polling question is, Considering a content ops initiative for learning content. Are you? Yes or no? So take a moment to participate.

Also, at the end of the show, I’ll ask you to rate and provide some feedback. The show rating system is one through five stars, with five being an excellent rating and one being low. There’s a little field into which you can type some feedback, which will be shared with the panelists today. I know they’d appreciate hearing from you, so don’t be shy before you leave today. Please do take just a moment to give them a rating and provide some feedback. We’d also like to thank our sponsor today, Heretto. For those of you who are unaware, Heretto is an AI-enabled component content management platform designed to help you deploy documentation and development portals that will delight your customers. I’d like to give just a moment for a customer from Heretto to tell you a little bit more about that.

Video testimonial: … and sometimes the problems I didn’t even know that I had. It’s an entire package. It’s an entire solution. I have a CCMS that stores my content, and I have a portal that knows how to publish that content. It’s been a great relationship. We have become partners, and I’m looking forward to what we’re going to do next.

Scott Abel: All right, and today’s show is also brought to you by Scriptorium. I will let our guest host today tell you a little bit about Scriptorium. First, let’s join everyone on screen so you’ll magically see all of us on camera if the technology gods are working in our favor, and here we go. All right. Look, hey, step one.

Sarah O’Keefe: It’s us.

Scott Abel: Ta-da, we’re all here at one time. How did that happen? Sarah, thanks for joining us today. Can you tell us a little bit about today’s show?

Sarah O’Keefe: Yeah, so we are taking on with many thanks to Rahel Bailie who’s been running this Let’s Talk ContentOps and started the whole thing. We’re taking this on as a webinar series, and we’re going to be talking over the next year about some of the interesting things going on in content ops, some of the fun new developments that are out there. We will try to talk about something other than AI at least some of the time, and we’re excited to be here. Scott, thank you for organizing because there’s a lot of stuff going on behind the scenes here.

Scott Abel: Thank you for that. I appreciate it. Hey, and just as a quick aside, help our audience members who might not know what your company does, tell us a little bit about Scriptorium Publishing.

Sarah O’Keefe: So where Heretto is a CCMS, so it’s a software system, we’re a pure place services provider. We’re interested in the question of once you buy the software, how do you configure it? How do you get it up and running? How do you use it to its maximum potential? Most of the work that we do is in structured content and DITA, not exclusively, but certainly most of it. So we’re interested in questions around scalability and localization and content velocity and how you make your content more valuable when you’re investing all this money in creating it.

Scott Abel: Awesome. All right. Well, I am going to let you take over the driver’s seat now and host today’s show and tell our audience a little bit about who you brought with you.

Sarah O’Keefe: All right. Thanks, Scott. So we are the, I don’t know, The Three Musketeers, The Three Horsemen of the… Nope, that’s not right. So with me today are Alan Pringle and Bill Swallow. The three of us are the three principles at Scriptorium. So we are the chief troublemakers over here, and we are looking forward to sharing some of our, hopefully, insights and interests and concerns around what’s going on in the wonderful world of content ops. With that, I think I’m going to launch our slides and-

Scott Abel: Okay-

Sarah O’Keefe: … jump right in.

Scott Abel: … go ahead and do that. I’ll disappear into the background, but I’ll be watching from afar, and I’ll jump back in here in just a few minutes

Sarah O’Keefe: We will see you on the back end.

Scott Abel: Alrighty, thank you.

Sarah O’Keefe: Alrighty. Off we go. So here we are. Let’s talk contentops, and is it going to be a boom or a bust? We have themed this thing around three people, three trends. The number 3 will appear throughout, so we’ll see how that goes. So there are the three of us. We did some quick intros, and we’re going to talk about three different trends, and we will see where that takes us. So trend number one… oh, sorry, three trends, but infinite opinions. If you’ve met any of us, this will come as no surprise to you whatsoever. So I’m going to turn it over to… I’m not going to turn it over to Alan quite yet.

Alan Pringle: Not quite yet, no.

Sarah O’Keefe: I have to start with the AI disclaimer. AI is this super mega whatever trend and there’s just no getting away from it, but we really didn’t want to talk about just AI in this session. So we have basically said, “Okay, yes, AI is out there. It’s going to be a tool. It’s going to affect all the things that are going to be happening,” but we’re going to set that aside because I think that AI is going to become part of your groundwater in the same way that it wouldn’t occur to you, well, it wouldn’t occur to me to write a document without a spellchecker.

So AI is going to be a tool that you apply to various kinds of things, and I hope that people are going to focus on this to do patterns and ideas and drafts. Really, my big takeaway with AI is that it introduces huge governance challenges, huge questions around how are we going to do this? Can we keep it accurate? Can we control what AI is generating or modifying? I think it means that we’re going to have to do more investment in our content, not less. So I’ll let you think about that, and we’ll see where we land on that at the end of the session. But with that, I will turn it over to Alan to talk about our first trend, which may have been slightly telegraphed by the poll.

Alan Pringle: Yeah, just a little bit. Our first trend is learning content and better content operations for learning content. So content creators and the learning and training space, they have to deal with this matrix of requirements that gets complex really quickly and frankly, scary pretty quickly. They may have this core group of content that more or less stays the same, and then they need to adapt and modify that content to address, say, a different audience, a different version of the software that they’re training on or a particular location, all of those kinds of facets. Then they also have to account for all of these different ways to deliver training. You’ve got in-person versus online. You’ve got instructor led versus self-paced and on and on and on. When you look at that as something that you have to face and then you put a layer say, of localization requirements on top of that matrix, you can understand why people in learning and training want to look at improving their content operations.

Several months ago, one of our clients, she leads a group of trainers who explain how to use software, said something that really resonated with me and the rest of the Scriptorium team. She said, quote, “We want to get off the hamster wheel,” end quote, of relying on copy-and-paste to maintain all of these versions and variants of their content. Every time that the software is updated, they have to do a new release of training, and so copy-and-paste, copy-and-paste, copy-and-paste. It’s not fun at all. So basically, they are looking at ways to eliminate that copy-and-paste. One way you can do this is to look at your body of content as individual small, basically, I will call them format-neutral components. Then when you have all your content broken up into these format-neutral components, you can mix and match them to create whatever it is that you need to create.

So if you have a case where you need to do, say for example, a printed study guide for an in-person course or you need to create an online course in a learning management system, you use the same source. You rely on the same source files, you just arrange them a little differently and then you process them with automated publishing processes that give you the various delivery formats that you need. Right now, a lot of people in the training and learning space are having to copy-and-paste from platform to platform to platform to do all of these different delivery targets, that’s going away when you break out of this copy-and-paste world. So basically, I see a whole lot of people breaking away from the copy-and-paste hamster wheel, jumping off of that and landing in better content operations to deal with these increasing requirements that trainers are facing with their content.

Sarah O’Keefe: Interestingly, if we look at the poll results from just now, it is, in fact, roughly a 50/50 split. It was like 52/47 or something like that. So people are definitely thinking about this and certainly more… I would say there’s no question that this is an increasing need, right? We’re hearing-

Alan Pringle: Absolutely.

Sarah O’Keefe: … about this more and more. Yeah.

Alan Pringle: Yeah, multiple clients, absolutely.

Sarah O’Keefe: Yeah. Okay, so that’s our first trend, and it looks as though the audience is at least halfway considering this as well. Hey, Bill, let’s take a look at the second one here.

Bill Swallow: That sounds good. Our second trend is, it’s actually not a new trend, but it is a trend that will continue going forward and that is replatforming. We’re seeing a lot of this over the past few years, and it seems to be increasing where a lot of companies maybe about five, 10 years ago invested a lot of time and a lot of money setting up documentation systems, CCMSs, publishing systems, web portals and what have you. Things are starting to, well, show their age because they are five, 10 years old, and your requirements then are not what your requirements are now, and they’re probably not going to be the same requirements that you have five years from now. So looking at the aging infrastructure, it’s time to start revisiting a lot of the decisions that were made. How are things working?

Do a retrospective on how the system has been performing, how content development, how it’s generally been going over the past X many years that you’ve been using that system. What works well, what doesn’t? It’s time to really assess all of that and get rid of what doesn’t work and look at future proofing going forward. It may mean shifting to something different. It may be just an upgrade and a re-tailoring of what you’re already using. But given that this is not necessarily a new trend, there are some helpful tips I think that we could probably share to ease the transition when you’re looking at a replatforming operation.

But first thing to consider is that even if you are moving from one system another that share the same type of source content, they may not be plug-and-play with your content. One system likely will interact with the content in a very different manner than another one. It’s something to be prepared for because even though your content may not change, that the source content structure, the source content format may never change, how the system interacts with it definitely will. Also, plan for a period of redundancy when you are going through a lot of these replatforming initiatives because you’re going to need to keep producing in your current system until your new one is fully set up, vetted, tested and ready to go live.

So you need to be able to figure out how long you’re going to need to maintain these systems. I would err on the side of caution and say longer and not shorter, but definitely take a look at that and try not to allow any type of a maintenance agreement tie you into when you’re going to switch those systems. You want to make sure that the new one you’re setting up is good to go. Another good tip is to start small and slowly gradually add more content into the system. You want to make sure that you have a solid pilot project in place so that you can not only prove that the new system will work and do what you need it to do, but that you understand exactly how that content is going to interact with each other, how the system is going to process all your various content and allow access for multiple users as you start adding more content in.

All that said, change management is critical on these things. You need to keep an eye toward the people using the system as well as what the system is actually doing, how it’s affecting other technologies that perhaps are in your tech stack, a myriad of things. But probably the biggest takeaway I can offer is when you are switching systems, avoid falling back into your comfort zone. You’re moving from one system to another probably for a reason. You’re getting rid of some old practices, establishing some new ones. It is critical not to fall back on those old practices and make sure that whatever it is that you may be getting rid of in the way you used to work that you are focused on not bringing that back in.

Sarah O’Keefe: It looks as though about three-quarters of our audience is happy with where they are, but the other quarter is definitely thinking about replatforming in 2024. So one in four, which implies that there’s a decent bit of, if not dissatisfaction, interest in making a change out there.

Bill Swallow: It may not be dissatisfaction so much as you can’t get where you need to go with the tools you have now.

Sarah O’Keefe: Right. Interestingly, that ties us right into the trend that I wanted to talk about, which is content as a service. Now the learning content trend is really a category of content that previously has not really been focused on in terms of content operations and in terms of structure. Replatforming arguably is a software tooling like, “What system should I pick?” Kind of decision. Content as a service is a change in how your publishing actually works. Actually, arguably it means that you no longer have publishing. So if you think about structured content for a second, we talk about how you separate the authoring process and the formatting process, you author the content and then you layer on formatting and you package it up and you deliver it.

With content as a service, you take that a step further and you separate the authoring, the filtering and the formatting processes, and you end up in this situation where you’ve completely fragmented what you’re doing. So what we’re talking about in content as a service is a scenario where the authoring that you’re doing in your CCMS like something like Heretto allows you to create topics or even smaller fragments that are inside that. But historically, I hesitate to use the word traditionally, but with something like DITA, you are going to then have a map file that assembles everything and you use the map to generate your HTML, your website, your output, your PDF, your whatever. When we talk about content as a service, instead of saying, “Package this up and deliver it,” what we actually say is, “Don’t package it at all, just make it available.”

Then the website or the endpoint consumer, the app, the software that needs that content reaches into your content database and grabs what it needs and then assembles it in whatever appropriate ways. This opens up some really, really interesting possibilities so that for example, I could have a service management system that needs certain kinds of procedures and instead of delivering the five procedure variants like beginner, intermediate, advanced, super user and internal expert who knows all the secret tricks, at the point where the content as a service reaches in to grab that information, it could say, “Oh, this person’s only been working here a week, so they get all the information, they get all the details because they don’t know anything. But the next time they get that procedure they get less information because the assumption is that they now know how to do some of these things.”

So I think this is a next gen, this is what content delivery is going to look like going forward, and it requires collaboration and integration and cross-pollination way beyond just the content development group. I think this is maybe the key thing to realize about content as a service is that it is no longer, “Hey, I’m in tech comm and I can just go and write my topics, put them in a map and render that map into HTML, PDF, whatever, and then I’m done.” You have these other contributors, maybe you’re also sourcing a product database content in addition to the tech comm content and then integrating them at the point of the website. There’s some really interesting stuff you can do there, but the problem is, of course, that you have to cross collaborate and step outside of that departmental role. So I think it’s going to be actually quite challenging and I’m very curious to see what happens there. Bill, what do we see in the polling there?

Bill Swallow: We’ve got about 50/50 on this.

Sarah O’Keefe: How interesting, so maybe not. We did not give you an, “I’m not sure,” option. We thought about it but we thought that was too easy. All right. So having said all of that, and I think now is the point where you might want to start thinking about putting questions in the Ask a Question tab if you’re interested in getting us to touch on some of the things that are out there. We come to the core of this whole thing, which is, is content ops going to be important going forward into 2024? I do expect that we are going to get robots with attitude. These guys are clearly headed for the disco.

So as we move into this AI world, I think that we are quickly going to reach a point where content ops is not or are not, I’m not actually sure which one, optional because in order to deliver the automation and to support the patterns that AI needs and/or expects, you have to have good content ops. You have to have good content, you have to have tagged content, semantically-useful content. You have to have all that automation so that you can drive the AI piece. I think, Alan, if you wanted to touch on the learning content and what it looks like over there with content ops.

Alan Pringle: Sure. I was at an event a few weeks ago, a training event, and I was talking to instructional designers and trainers. When you mentioned the whole concept of content operations having a single source of truth for a particular piece of content, instead of 14, 15 versions and copies of that, you should have seen their faces light up. This is something that really resonated with the people that I was talking to. The ability to do automated publishing where you’re not having to dump your content into a bunch of different platforms to get hit all these delivery endpoints, these people, they really have their hands full, and they need a break. I think better content operations will give them the opportunity to do what they do best and that’s creating content for the people that they’re trying to educate, to train, instead of spending time on this busy work that happens over and over and over again and is a never-ending cycle for them.

Sarah O’Keefe: Bill, what about on the replatforming side of things?

Bill Swallow: I think with the AI question, it really comes down to if you have a directive to incorporate AI into your work, whether it be from an end user point of view or from a source author point of view, does your system allow for that, or is it something that you have to try stapling onto the side and hoping a strong breeze doesn’t happen? There are a lot of tools that are starting to adopt it and a lot of tools that are starting to look at different ways that it can be used rather than the typical means that we see with ChatGPT and all. So it’s looking at the replatforming, AI is not a reason to jump, but you may be limited in what you already have, in which case you either have to work in elaborate workaround in place, or look at switching it to a system that will get you to where you need to be quicker.

Sarah O’Keefe: Yeah, and I think I feel the same way about content as a service that if you have that requirement for additional fragmentation, then you have to make sure that your tool stack and your systems and all the rest of it will support it. We’ve got some interesting questions around that coming in which we are not ignoring and we will get to. So I think some of our listeners are also concerned about those kinds of issues. So I will encourage you, again, to go ahead and start putting your questions in. I’ve got a couple of slides here that I need to show you that are related to resources, and I wanted to ask you, you the audience about your content ops prediction for 2024. Basically, this is technically a poll but not really, because really what we want you to do is just pop it into the Ask a Question.

Where do you think this is going? What do you think is going to happen? We’ve got a couple of really interesting comments on that already, and we’ll touch on those in a second. So with that, we will take your questions. Alan and Bill, this is your 10-second warning that I’m about to turn the slides off, which means we’re about to be on live video. Again, the attachments, there are a whole bunch of resources in the attachments. Additionally, you can use this QR code that Christine put together for us and reach, I think, a landing page on our website that has a lot of the same things in it. So with that, I’m going to skip over here to my other screen and hit this button that says End Screen Share with fear and trepidation. Hey-

Scott Abel: Yeah, you did very well.

Bill Swallow: You made it.

Scott Abel: If you would’ve clicked End Talk, that would’ve been disastrous.

Sarah O’Keefe: I am familiar with the-

Alan Pringle: Or not.

Sarah O’Keefe: … End Talk clicking button, and I don’t want to talk about it. Okay, so a couple of things here that have come in and yeah, this is really the key thing. Somebody else named Sarah has left a comment that says that generative AI will continue to be the buzz phrase for executives, and I have to agree with other Sarah. Of course, you’re exactly right, and part of this is that I think sitting in technical content, especially if you’re sitting in content that is regulated or affects life and safety, it is really, really hard to take generative AI seriously because it’s going to write a procedure that if applied to, “How do I use this pacemaker?” Would kill somebody.

So this is concerning and we don’t like it, but it is the catchphrase, buzzword, whatever, and so we can’t just ignore it, unfortunately. Okay, now Bill, I think this is to you, there’s a question here from Michael asking about busting content silos and how to unify siloed content ops. How are we going to pull that off? He says, “There’s a lack of operations, integration, technology and automation to help weave siloed content groups together so that they can collaborate across the entire customer journey.” Your thoughts?

Bill Swallow: I will use the standard answer to begin with, it depends, and then elaborate from there. So it is a tough problem, and it depends on how siloed these groups are and how siloed they need to be. We are seeing a lot more groups, for example, the learning and training groups and the technical content groups starting to come together more because there is that collaboration there on content. The training group may have insight that they need to bring back to the technical documentation group and help them re-tailor how information is being presented, how it’s being written, elaborating, and so forth. The training group may also say, “It would be much easier to just have a poll of this information as you update it so that we don’t have to go through and update our 15 different training guides and our slides and our instructor manuals and our quizzes and everything else with this new content all over again.”

I hate to use it depends, but I use it as a joke mostly, but it is true. I think that also we’re seeing an alignment more on the goals of various content development groups within a company. So as long as you can align those goals in the same direction, you can start getting people starting to think in the same direction, “Hey, I don’t have to write and rewrite this stuff all the time. There’s one central place that I need to go and I know exactly where to find the information, and I can get it and do my job with it.”

Sarah O’Keefe: Alan, did you want to weigh in on that?

Alan Pringle: I agree with it depends. I think you can make a case sometimes siloing is necessary and it’s not undoable. I think there can be a business case for it sometimes, but I do see the overlap that Bill’s talking about. I would even pull marketing into that as well-

Bill Swallow: Yes.

Alan Pringle: … because if you’re talking about product specifications for example, wouldn’t it behoove a company to have one version of those and every single department used them instead of different copies of that which will get changed to be wrong immediately, three, four different departments? So cuts both ways.

Sarah O’Keefe: Yeah, I think it’s interesting because I think that I’ve actually more or less given up on the concept of unified content, unified content strategy and unified content ops in general. I think there are specific instances where I can see it happening and typically, it’s things like overarching tools, enterprise taxonomy, enterprise-level terminology, style guides, those kinds of things. But I think I take this view that at this point there’s a reason that tech comm wants a certain kind of content management system and marcom wants a different kind of content management system.

So if we can get some unification on the critical stuff, which is to say the taxonomy and the terminology and the data sources, to Alan’s point, you should not run around having different height, weight specifications for a given product. That should be sourced from one place, and it is probably your PLM, your product lifecycle management system. I think those things are important, but I’m not so sure it’s important to have unified content authoring. It’s more this team owns this chunk and this team owns this chunk and this team owns this chunk, but we have come to some sort of agreement on these overarching concepts or these overarching, let’s say, taxonomy layer where we do need to be consistent.

Alan Pringle: I do think some of the tool vendors are becoming wise to what you just described in creating tools that play well together so they can give you that infrastructure so people are paying attention to what you just said, how you can still have separate groups yet still be at the enterprise level.

Sarah O’Keefe: So there’s a question here about content as a service and context. So if you are delivering content as a service, then how do you deal with this question of a chunk and whether or not it can stand on its own? So this is from the question, “How do we manage fragmentation to facilitate reuse and not lose sight of the importance of context that is all the blood, sweat and tears that goes into creating and managing books, maps, book maps and deliverables?” That’s a really, really interesting question, and do either of you want to weigh in on that or should I jump in? That was code for, “Give me 10 seconds to think about it.”

Scott Abel: That’s right.

Alan Pringle: Well, you and everybody else were thinking, it is a very, very good question. It’s a balancing act, and the question already implies that. That’s how I see it. You’ve got to find that sweet spot where componentization becomes too much versus where things are too big. There’s that Goldilocks place somewhere in the middle there, and finding that, that can be a challenge. It can be.

Sarah O’Keefe: Yeah, and I think sequencing and hierarchy. So if you think about a series of steps, let’s say you have a five-step procedure and you have somebody trying to do this five-step procedure using a mobile device as their help access, so you can’t put all five steps on a single screen, they won’t fit. What you’re actually going to do is put up step one and then they’re going to maybe swipe, and you’ll get step two and then you get step three and step four and step five and maybe there’s some cool images in there. But at the end of the day your system, whether it’s content as a service or anything else has to maintain that sequence. It has to know that one comes before two comes before… did I do that right? Yes. So that it’s 1, 2, 3, 4, 5 and not 1, 3, 5, 7, 9, 2.

So somewhere you have to preserve that context or that information about sequencing and hierarchy is similar. What is the parent of the thing that I’m dealing with right now, and how do you address that? Now some things are context free or can be, like a tool tip. If you’re just explaining what a specific button does, you probably don’t have a whole lot of context around that, which makes it a little bit easier to deal with. But I think that it is important maybe to look at it not as an either/or, but rather as content as a service is a way of delivering or making available content that the endpoint requires that you cannot pre-package or pre-render for whatever reason.

There’s lots of reasons why you might need to consider that. I think that’s the best I can do. It’s not the best answer necessarily. Okay. We have a question about learning and training content and reducing copy-and-paste. Now this is specifically a tool question, but I think, Alan, this is going to be to you, so, “What options exist to reduce copy-and-paste?” Asks Marie. “We use Articulate 360 to create learning content for our LMS. Articulate does not support Reuse as far as the version I have, but we use XML and Reuse for technical content. We don’t have a separate learning group and tech comm group, which I think means the same people are working in both tools probably.” So what’s your take on that?

Alan Pringle: I don’t have any specific recommendations, but I’ll give a big picture answer. I have noticed that there are a lot of tools in the training space that I call closed. They do not play well with others. They would not get a good report card on how they behave on the playground. They do exactly what this person is asking about. They force you to do copy-and-paste, which is just simply not sustainable, and it’s not helping anybody to do that. I realize you need that end product, but what I see a lot of training groups realizing is they are going to have to really… this gets to what Bill was talking about, replatforming. They are going to have to take a hard look at the tools they are using. If you are forcing people to copy-and-paste into your platform to get at a certain delivery endpoint, that’s a huge red flag, and it’s time to look at ways to really, is there another way that you can get that delivery endpoint that has more automated transformation, can somehow use your existing content as it stands?

To me, I would love to be able to give an answer that is very specific and to solve this person’s problem, but I think this is time where you need to do some assessment and reflection. Basically, do a little strategy project and say, “This is where we are, these are the pain points, this is the end point. How can we get there and avoid the pain points?” That’s the kind of thinking that you need to do in regard to these tools. The answers to those questions may lead you to not only just replatforming but jettisoning and completely replacing a platform, so yeah.

Bill Swallow: Yeah. There may be options to push content from one system to another, but in the case of a lot of these learning and training tools, as Alan mentioned, they’re kind of a closed box. So once you push content there, it’s stuck there. If someone modifies and continues to improve the content in one location, now you have the problem with one group chasing the other as far as making updates in two separate environments. So it doesn’t really solve the problem. It might cure some headaches initially, but you’re going to end up with the same problem where you have two completely different content sets that maybe one gets updated with new content from the other every so often.

Alan Pringle: And there goes your single source of truth. Bye.

Bill Swallow: Yep.

Alan Pringle: Bye-bye.

Sarah O’Keefe: I think, ultimately, the question is how badly do you want to get off the hamster wheel?

Alan Pringle: Yes.

Sarah O’Keefe: We’re going to use hamster wheel forever, so thank you to the person-

Alan Pringle: We have to acknowledge-

Sarah O’Keefe: … who produced that. Yeah.

Alan Pringle: Yeah, the person who said that we thank you times 1000 because we love it. Thank you.

Sarah O’Keefe: It’s the best.

Scott Abel: Maybe we should investigate getting your show sponsored by Habitrail. We could have the hamster wheel of death. Yeah.

Sarah O’Keefe: All right. I’ve got another terrifying question over here that I would love to dump on somebody else. “What role do you think the content teams and the content play in RAG, retrieve augmented generation for generative AI?”

Scott Abel: That is super nerdy.

Sarah O’Keefe: That’s going to be me, isn’t it? Okay. It depends, but okay, first of all, we have no idea, because this stuff is like eight minutes old. But beyond that, so for those of you who are not familiar with it, I’m going to define retrieval augmented generation and the people who actually understand it well are going to cry. Retrieval augmented generation refers to the process of using a generative AI system such as ChatGPT but extending it with a database essentially or a knowledge graph of known good facts. So you imagine that you have a database with historical information, the date a certain war started and stopped, that type of thing. Think Wikipedia, but structured. So when you go in and say, “Hey, ChatGPT or whoever, generate an article for me about X, Y, Z topic,” it doesn’t just do what amounts to auto generation and free association.

It also uses these facts that live in the background or that live in that database to give it some guardrails to keep it from making stuff up. Now, I do want to point out, this is my favorite example ever. I asked ChatGPT for my bio, and it informed me that I had a PhD, which I think is awesome because that is the best and most efficient way to get a PhD ever. In a retrieval augmented generation scenario, presumably that type of data, the biographical data would be stored somewhere, which would prevent the generative AI from going off the rails and inventing things. That’s the concept. Now, what role do the tech comm content teams potentially play in that? Well, the job of tech comm is to provide enabling content, which is how to successfully use this product, which is or should be fact-based.

Provided that your content is sufficiently well-structured, you could then have the ability to make that available as a source of validation, that that would be one of the places where the language model is looking to figure out how it responds to the query that the person has put in. I will say a couple of things about this. One is that we have to be really careful with this because I’ve seen a lot of, lot of, lot of really bad technical content. So we have to be careful about assuming that the technical content is known good. We’re assuming technical content is known good, and therefore, it can be the foundational underpinnings of whatever we’re doing here. That’s really step one is make sure it’s good. So that actually really concerns me because I think there’s an enormous amount of stuff out there that’s not actually good and that’s a pre-req.

If the bios you’re writing or if the data that you’re embedding in your tech comm content isn’t very good or isn’t very accurate, we’re going to have some big issues down the road. But setting that aside for the moment, at least hypothetically, we should be able to provide structured, tagged, marked up metadata-enriched content that can then serve as a source of guardrails for the generative AI. Stay within this box and don’t get too crazy would be more or less where I think that would go. Well, the AI experts will cry when they hear that answer, so we’ll just leave it there. Okay, I think that’s it. I want to thank everybody. There’s some really interesting questions in there that made us think and/or squirm and/or run away down the Habitrail. Scott, I think I’m going to throw it back to you to wrap things up, and thank you.

Scott Abel: All right, great. So don’t forget, you can learn more about Scriptorium by visiting them on the web or you can check out some of the links they provided for you in the Attachment section. In fact, if you click through there before you leave, you can get access to a couple of books and some information about contacting Alan, Sarah and Bill if you’d like to to follow up on questions you may have about content operations. Also, I’d like to invite you to join Patrick Bosek and I on November the 16th. So just next week sometime for our Coffee and Content with Laura Vass. She’s going to be talking to us about API documentation, about developer experiences and about the dev portal award. So if you’re involved in software documentation, this will be a great show for you. Several hundred people already signed up. It’ll be a great conversation with somebody who’s deeply involved in the mix there.

I’d like to thank Heretto for being our sponsor. Heretto, once again, is an AI-enabled component content management system platform that can help you deploy documentation and developer portals that’ll delight your customers. You can learn more about them at If you would be so kind as to just give us a rating on the way out the door using the Rate This tab underneath your webinar viewing panel, you can do so by clicking one through five stars. One is a low rating, five is high, and you’re asked to rate the quality of the information that was provided today, and we’d really appreciate it. There’s a field into which you can type some feedback which will be shared with the panelists, and I know that they’d appreciate that. So thanks for joining us and for Sarah O’Keefe and her team today from Scriptorium to talk about Content Operations in 2024: Boom or Bust?

We appreciate you being here. As always, we’d like to thank you for participating in The Content Wrangler Webinar Series. Be safe, be well. Until next time, keep doing great work. Watch out for Sarah’s next show, which is coming up in January. There’ll be some publicity coming out very soon about that, and I know she’s got some great guests lined up. So definitely make time on your calendar to attend. Usually, it’ll be about every other month, so you’ll probably get six different opportunities to see Sarah next year talk about content operations on this platform and I’d encourage you to monitor that. She’s got great guests and topics coming up. So thanks so much everybody. Until next time, be safe, be well, and have a great day. Thanks for joining us. Bye-bye.