Zohra: 0:02

Hello friends and welcome to season 8 of Inside Tech Comm with Zohra Mutabanna. This season features a collection of conversations that explore the ideas, challenges, and opportunities shaping technical communication today. From AI and content strategy to leadership, product thinking, and the future of our profession. Each episode offers a fresh perspective from people doing the work. Let's get started. Season 8 is proudly sponsored by LavaCon. Use discount code ITC26 to save $200 on registration for this year's conference. I'll share more about LavaCon later in this episode. Hello, listeners. Welcome to another episode of Inside Tech Comm. Today I have with me Elizabeth Favre. She and I are colleagues and friends, and she will be my LavaCon co-presenter. She goes by Liz. Liz and I have spent the past year experimenting with AI in real content work, building agents and workflows and skills and trying to figure out what holds up and occasionally discovering that we try to make one thing do far too much. And that is what we are going to try and unpack. Since we are going to be presenting at LavaCon, this conversation is going to be a teaser to what we intend to present to you at LavaCon. So let's get started. Hey Liz, welcome to my show.

Liz: 1:30

Hi, Zohra. Thank you so much for having me.

Zohra: 1:32

Yes, I'm excited to be chatting with you. Liz, give us a little background about yourself.

Liz: 1:38

I did not take the traditional route to becoming a technical writer. I do not have an English or communications background. I actually studied biology in college, and I spent about 18 years in biomedical research. And um, when I was ready for a change, I was really interested in some of the work we were doing with nonlinear thinking and oddly enough, AI, natural language processing, and with neural networks and principal components and all of the stuff that really requires AI. And so my interest in that made me think maybe I'd like to go into bioinformatics. I went to a coding boot camp and I discovered that I am not a developer. But with my newfound knowledge in the world of software, I was able to get a job writing requirements and security tabletop exercises and taking part in agile ceremonies and things like that for a startup. And um, when that startup restructured, I ended up getting a government contract job. And I was doing some um policies and procedures for the risk management framework for the defense health agency. And after that, I went to, you know, other government contracts. And then one day I was so thrilled because I had been trying for years and years and years to get a job at Blackbaud because I knew Blackbaud because my husband had worked at Blackbaud for many years. And I knew that they were a company that really cared about both their employees and their customers. And I was so excited to hear there was an opening for a writer there. I applied, I talked to the other writers, and luckily I got the position and I got to work with you. So a very roundabout way to getting to meet you and working in tech writing. That's pretty much my story.

Zohra: 3:31

If I may add, Liz, you are one of the nerdiest tech writers I have ever encountered.

Liz: 3:37

I wear that badge with pride.

Zohra: 3:39

And you should. I'm so glad to hear that. You should absolutely wear that with pride. Yes. And Liz, you and I have worked on some very interesting projects over the past few years, right? Even before AI became available in our corporate environment. And that's how I got to see the nerdy side of you, where you would create these interesting pivot tables and do all math stuff that is not something that comes easy to me. I'm not a number person. And to see a writer be comfortable with that was uh it was amazing. So yeah, Liz, it's been fun, and now we're gonna make that even more fun by presenting, co-presenting at LavaCon.

Liz: 4:19

That is gonna be so fun. I'm so excited.

Zohra: 4:22

Absolutely, me too. Me too. So let's it's a very different place we are in today.

Liz: 4:28

It's amazing how quickly everything is changing in the world of AI. It is.

Zohra: 4:32

We've been discussing how do we try to retrofit our content into the summary that we have. And I think that's the wrong question that I was asking. Instead, you pushed me to think, how would I pitch this to the audience? Right. And you started thinking about the audience. Now, as a tech writer, I should be thinking about the audience, but I was sort of framing it in a very different context and time.

Liz: 4:56

There's a really big difference between presenting at a conference and trying to do our tech, because all of our tech writing is about like how do we make our audience able to perform this one specific task. And with a conference, I don't know, this is new to me. We're gonna figure this out.

Zohra: 5:13

We are gonna figure this out. And I think even by the time we get, we are in September, early September right now, as we are doing the interview, but as we get into October, things are gonna change. The tools we use, I mean, the tools are becoming more advanced and their applications are becoming more advanced. But we have to ground ourselves at some point and say, this is what we are gonna present. I thought, let's think about what the pitch is gonna be from my end, and then I am gonna invite you to give me your pitch because we are thinking about this independently. And I want us to kind of talk through where we are at and why we wanted to do this presentation and to really share what your nerdiness and my shadowing you, and me experimenting and breaking stuff, and then you joining me and challenging me on some of the things. So we've challenged each other over the course of the past year.

Liz: 6:10

I want your audience to know that Zohra is really selling herself short here. Zohra is the mastermind behind almost all of our experimentation, and she's the one who bravely led me into the world of AI and tech writing. She's the one who was like, hey, come join this group. She's kind of one of our brave leaders. So don't let her tell you otherwise.

Zohra: 6:33

Thank you for that. But we are gonna jump right into the pitch, and I want us to unpack this awesome exploration that we are on, right? So the pitch that I came up with was if you are experimenting with AI in content and wondering why some efforts take hold while others stall, we are gonna share what we built, what didn't work as expected, and what that taught us about everything that needs to exist around the technology. So Liz, with that pitch, what is your take?

Liz: 7:08

So to me, the biggest take and the thing that I would love to have our audience walk away with is it's okay to not get it right the first time. The worst thing that you can do is not try. Because, you know, like Wayne Gretzky said, you missed 100% of the shots you didn't take. The thing that I would like our audience to walk away from is that if you go in there and try, you've already succeeded because experimenting is the only way to learn. And like Thomas Edison said, it isn't 3,000 ways not to invent the light bulb. Okay. So yeah, I just want people to feel like I can go in there, I can play with this, I can mess it up, and it's gonna be okay because I will walk away with more knowledge than I had when I started.

Zohra: 7:53

Yeah, that is so true. We invite them to attend our presentation because it's not only about the success that we're gonna be talking about, it's the many attempts to succeed that helped us eventually.

Liz: 8:07

Absolutely. I mean, to me, see, this is a good thing about our partnership that we're bringing to this conference is that Zohra is very much, she's gonna tell you how this works and and and what the tools are. And she's gonna give you all of that fine, wonderful, detailed information, and I'm gonna be your cheerleader and say, you can do this.

Zohra: 8:28

We've truly challenged each other. And I think that is another thing that probably if you are scared, I think that's what I maybe want to also pitch for is if you are a solo writer, maybe you can partner with somebody outside your team. Uh if you have other writers on the team, maybe partner up with them. So that fear factor drops. As a team, you're able to build and challenge each other and learn from each other. And that's what has been our, I would say, our blessing.

Liz: 8:59

Yeah, and I feel like like when you're starting a new skill or learning a new tool, I think you really gain a lot by finding somebody else who also wants to learn it. And at first you might be thinking, oh, other people might know more than I do, or other people might not be interested. But you've got to ask, you've got to ask and go out there and find them. Because usually if one person has a question, more than one person has that same question. And that's how you're gonna make your network, that's how you're going to learn from each other. I mean, you never know. You might be the one who knows more about part A and the other person knows more about part B. And when you put those two parts together, you come up with something really special.

Zohra: 9:41

So, Liz, when you are imagining the person sitting in our audience, what do you think they are struggling with?

Liz: 9:48

My guess is that definitely the struggle that I dealt with the most is the feeling of am I qualified to do this? Am I smart enough to do this? This looks really complicated. Maybe I can't do this. Which is again where I say just go out there, try things, break them. The world's gonna keep spinning. It's gonna be okay. And I feel like I really want everyone to feel empowered to go in there and push some buttons. And these days, I want everyone to know that AI is not just there to make things, it'll also help you. You can ask it questions, even better, ask a person some questions. Because as much as AI is great, it's never going to replace human interaction, you know. And I would say that in terms of specific to our project, one of those main lessons learned for me was definitely if you're working with an agent or a skill, or if you're trying to design prompts and all that, don't try to make each individual piece do too much. That's one of my biggest pitfalls, is that I get excited and I'm like, oh, we can make it do this and we can make it do that, and we can add all of these different resources, and then we'll separate them out and we'll tell it when to use this and when to use that. And the fact of the matter is AI is similar to a person in the aspect that if you give it too much information at once, it just gets confused. You can give it all the information in the world and it will retain it, but it's not going to know what to do with it if you try to make it do a specific task with all the information. So that was one of my biggest lessons learned here is make small pieces and fit them together like a jigsaw puzzle. Really focus on what you want to do, do one thing really well, and then integrate the one things.

Zohra: 11:39

Yeah. So, Elizabeth, the case in point here is an agent that we built with Copilot Studio. At our company, we have this AI champions program that you can sign up to. And that's what we did. And we get access to tools before anybody at the company does. And we got super excited and we decided we are going to build an agent. This was very new. This was we're talking about late last year, right, Liz?

Liz: 12:04

Yes.

Zohra: 12:04

As part of that, we had six weeks, six to eight weeks to build it and demo it. And that is what this is referring to, and the learnings that came out of that, that later informed how we set up AI tech and how do we scaffold, right? So that really became the basis of how we build things up between the two of us and then beyond.

Liz: 12:29

That lesson for me has continued to be a hard lesson learned in my continuing experimentation with AI. Whether I'm making skills that are supposed to help with accessibility, or I'm trying to make an onboarding app to help train our new employees, or whether I'm working up an agent to help review content or review PRs, pull requests in the code, all of the time. I always have to rein myself back in after I do my first iteration and say, wait a minute, wait a minute. Why am I trying to make this one tool do 17 different things? We're not making a Swiss Army knife here. We're making specialized tools to do specific tasks. So yeah, I would say um, if you see somebody doing something super fancy that can do all of the things, I might caution you to ask them, um, so do you have a lot of parts each doing specific things? Or do you have one tool doing everything? And if they have that one tool doing everything, you don't need to be intimidated that and say, Oh my goodness, I can't make one tool that does everything. You can say, This is a lesson. This is a lesson. We all approach this differently. So maybe the way that you're doing something might be different, but it might be better. So don't be afraid to experiment and try new things and uh compare and and talk to those people, and maybe you'll learn something from each other.

Zohra: 13:59

We didn't begin with a grand idea about systems or processes or culture. We just began with a specific content problem. It's another problem that we try to over-engineer our solutions, right? But we were looking at you and I, when we partnered up on that agent development, we were thinking about at that time we didn't even know what problem we were trying to solve, right? I think now we have come far ahead where we think in terms of what problem do we want to solve, and then what do we build to solve that problem? Right. So looking at problems and then trying to solve that. And to some extent, when we started with the agent, our goal was what can this agent do for us? And we did actually think about the problems. First, we got excited about building the agent, and then we got excited about okay, what can it solve? And that is where we over-engineered it. We both came from different ideas. What were you trying to solve by building an agent? Not trying to talk about the multiple things, but also this was our one shot. So we had to do everything. This agent had to do everything for us. But where were you coming from, Liz? I've never asked to ask you that.

Liz: 15:06

I'm gonna be really honest with you here. I was very much relying on you to figure out the use case and the whys and the what are we trying to do. And I was just so this was my nerdy badge. I was so excited about pressing every button in that studio and saying, like, what does this one do? What does this how would I use this one? And that is a terrible way to approach a problem and engineer something. So I would say that my approach to it was not a practical one. I would say that I've definitely learned from that that even if you are excited about a new tool, you need to have a reason why you're using it in order to figure out how to use these things. They may be exciting new toys, but you still need to use them in the right way. So once I did start thinking about the actual focus of what we were trying to do, the problem we were trying to solve, I would say that this is, I think, is a common problem that I have seen in every place that I've worked. There is always information scattered. Some of it's on people's hard drives, some of it's in shared repositories, there's something in the SharePoint, there's something in the ADO repository, there's something else in GitHub, there's something else on Jack's computer. And not only do you sometimes have to guess, well, which resource do I need for this particular project, but sometimes you don't even know all of the resources that exist. Like there would be times when I would write into our Team Slack channel and say, hey guys, I think we need a rule about whether we use term A or term B. And I would get a response back within a couple of minutes saying, like, well, did you look at this page in the guide that was written four years ago? And I'd be like, I didn't even know we had a guide from four years ago. That's great to know. I wish I had known that because I could have used this three months ago. And so the problem that I was thinking we were trying to solve is how do you tell the agent, this is what I the information I have, and this is the problem I'm trying to solve. So what are the resources? Like, give me the answer from the right resource. Don't make me go hunt down that resource.

Zohra: 17:22

And when we started building the agent, that was a problem we were trying to solve. So I will actually take back and say that no, we were not trying to solve any problem. We were actually trying to solve a problem as you and I started brainstorming, and then we had two other members on our team who came from different business units and they had their own problems, and we had to try and make this agent accommodate those use cases as well. I don't even remember now what were the use cases that the agent tried to solve?

Liz: 17:48

Well, blog article, HTML output, UX writing. Right, right. So we had internal UI text writing, we had help documentation, long-form help documentation, and then we had marketing, we had blogs, and we had uh education and training.

Zohra: 18:08

Yes, that's right. What do you remember about the moment when we realized that despite everything?

Liz: 18:15

I had no doubt that we were gonna get it to work because you were working on it. And let me tell you, Zohra, when she puts her mind to something, it's gonna get done. So I wasn't surprised that it worked, but I was really surprised how well it worked. So the way that our system was was set up, it started with an adaptive card that would ask you, like, okay, well, what are you trying to make? And then it would ask you, like, what stage are you at? So, like if you were just brainstorming still or if you had a draft review, but it would ask you what type of material. And I was so excited that when I put the same content in and said this is for a blog, and then I put in, and then the second time I ran it, same content, but I said this was for help documentation. And the blog spit out something that had our brand colors and it was friendly, and it was it was a conversation narrative. And I compare that to our instructions that it spit out, which was just, you know, it was a very consistent format that had the numbered steps and it was really clear. I was so excited that first of all, it had found the right guidance. Like we had set up these guardrails of if they say it's a blog, use these resources, and if they say it's instructions, use these other resources. I really thought it was going to be confused. I thought it was gonna make something ugly. It clearly needed human revision. It wasn't ready to publish, but just the fact that it could do that so quickly that I didn't spend three hours reading the manuals myself. I was really impressed and I was thinking, this is going to make talking to the other people on my team so much easier. Next time that our marketing person comes to me and says, Hey, I saw that you wrote up these instructions, but I need to know how to sell it. I could actually put my all of my source material into this agent, tell it, okay, now convert this to a blog, and it would spit something out that maybe it wasn't ready to publish, but I could say, Hey Jason, here you go. And that used to be like three days of work, and it had taken maybe 15, 20 minutes at with iterations. That was really exciting.

Zohra: 20:40

When we were designing the agent, we were thinking about how does the user select the option that they want, for example, UI, text, or blog or help. And then you designed, you made it look good and more user-friendly. So that was another angle that we tried to take with our experimentation, right? The more the thing was the more we engaged with it, the more comfortable we became with it. And then you pushed the boundary with we are gonna put adaptive cards and the user can pick it, it was just more uh clear, right? As you're interacting with the interface. So we started thinking about the interface, what it could do, and then now how do we invite users of these agents to interact with this agent in a more user-friendly way? And that's what the adaptive cards are. And I think that was a pivotal moment for me when you demoed it, because I wasn't getting it. And this is when your nerdiness showed through.

Liz: 21:34

Well, I appreciate that. I think for your audience, for members who haven't used Copilot Studio, the thing to know here is that normally when you're creating an agent with Copilot Studio, it defaults to starting off with just a chat and an input box. There you go. So like it will just start with the chat, you know, you put in "hi" or whatever, and then it will kick off your flow. But the thing that we discovered, what I say discovered from reading the docs. So somebody out there is reading these, all of your hard work. From reading the docs, we learned that if you didn't want it to start off with just a chat bot, you could add this thing called the adaptive card, which you could add multiple choice, you could add radio buttons, you could add all of those fun things so that the person didn't have to guess. Because that was one of the things we were wondering. Like, how do we tell the person? Like the first thing you have to do is. So that's kind of the nice thing about you're trying to get somebody else to use your AI workflow, and you just you need to kind of nudge them in the right direction to get started. It was a nice tool to be able to have that to set that in. Yeah. I like that.

Zohra: 22:53

And I think I had built something with very old style radio buttons. And then you come with these fancy adaptive cards, and it really changed the experience, right? Because we had one member on the team who was very resistant to the idea of even testing this as an agent.

Liz: 23:10

Oh, yes.

Zohra: 23:11

Right. But then with the design approach that you took, they became much more comfortable. And we were actually able to articulate to them this is what we are trying to say. So I was too much in the weeds, and you kind of pulled the team out of that with your adaptive card presentation in the flow.

Liz: 23:29

Yeah. And and then a team member, she had made some really good points. She because she was asking questions like, Well, why is this any better than just putting, you know, just putting a request in AI in the regular Copilot chat and saying, Hey, make me this material. And so, like, I think that being pushed that way, being challenged that way was really helpful for us because we had to really think about, like, oh yeah, why is this? It got us thinking about like, well, we have these guardrails. Are we using the guardrails right? Is this actually helping? And yeah, it did turn out to be a better experience. And again, yeah, like you said, like for people who are resistant, for people who are like, I don't know where to start with this tool that you're giving me. Like, this isn't something that I can look up on YouTube and watch somebody else do a walkthrough. It's like all you have to do is just talk to it and like being able to pinpoint what that starting point is and to push somebody, I think you're much more likely to get interaction, even in a conversation at like a networking event or something like that, you're much more likely to get inter interaction if you can say to somebody, do you do A or B or C? Do you prefer lemons or limes than if you walk up to somebody and say, Hi, tell me about you.

Zohra: 24:44

And that conversation with this teammate actually it makes me think about something else. If you are one of those who thinks and has these problems, maybe you can approach another team member who might be a little more conversant with AI and ask them, hey, can you solve this problem for me? That is also a way to contribute because you are trying to solve a problem. So partnering doesn't mean that you have to uh bring hard skills to the table. But if you're thinking about a problem and you think there's a solution with AI, partner out. And it's amazing what you can build and what you can have from that simple conversation.

Liz: 25:18

It really is. I know I've relied very heavily on you for the use cases and like what is the actual like defining what it is we're trying to do. I love playing with those, with those buttons, but you can play with the buttons all day, but you're not going to make anything useful if you don't have a purpose and a definition.

Zohra: 25:34

The episode is sponsored by LavaCon, the content strategy conference taking place this October in Charlotte, North Carolina. With more than 70 sessions and workshops featuring speakers from companies like Salesforce, TikTok, T-Mobile, and Amazon, LavaCon brings together content strategists, documentation leaders, and content professionals to share practical ideas you can apply right away. And yes, it's known as the Fun Conference, too, with networking events, therapy dogs, comfort llamas, storytelling, and karaoke. Use discount code ITC26 to save $200 on registration. Do you want to talk about a favorite experiment of yours after that list? Because there are so many. And I didn't want to pick and say, hey, let's talk about this next.

Liz: 26:15

I think right now, the thing that jumps to mind is the most recent experiment. Do you want just an AI experiment or one that we did together?

Zohra: 26:25

Anything is fine. It doesn't have to be AI. I think everything informs what we end up doing later.

Liz: 26:30

Okay. Well, I think to me, the thing that's really fun right now is playing with vibe coding. I am really digging the vibe coding. And I'm sure that there are people out there who are going to cringe and say, no, vibe coding isn't real coding, and you're just going to make terrible things. But I'm having a lot of fun with it. That's my nerd badge again. It's a fun toy. And the use case that I have found for it was onboarding apps for our new employees. And one of the main things that I've learned doing this is that sometimes I feel like this is going right back to that lesson of don't ask AI to do too much. I find that if I ask AI to put in specific components or connections, if I tell AI I want to try to, in the end, I want this type of thing. Don't build it for me, but talk to me about what my options are. Going in and really using, in case people don't know, your AI has different modes. It has planning mode and an agent mode and the different ones, different tools have different options. But putting it in that planning mode and being able to ask questions, I found that if I just tell AI, I want an app and I want it to do this, the AI will, within a matter of minutes, make that app for me. But it's not going to be exactly what I wanted. It's not going to do it the way I wanted it to do it. For example, like I was I was trying to add a clock to this app that the people could type in where their coworker was and it would tell them what time it is in that time zone because you know you don't want to be bothering people in the middle of the night. And so I asked the AI to do it, and it did. It made a clock, and that clock you could set to certain time zones and stuff. I was like, oh, but it doesn't work the way I wanted it to work. I can't type in the name of this city in Russia and have it tell me what time it is there. So I found that giving small instructions and small pieces, being really specific about what you want is very important. If you have style guides that you want to apply to your components, I found that just pointing it at the style guides doesn't help. You have to be very specific of I want this component and I want you to follow this set of rules. But I think that the part of it that's been really fun is just in addition to getting the endpoint that I wanted it at, but by dealing with it by piece by piece, I'm learning more about the individual pieces that make up the app that I wanted to make. And I'm finding that by learning all of those different pieces, I'm learning all kinds of new things that I can actually apply even to my docs and to any other content that I would create. Like if you have a help page that you're working on, the pieces that you learn from whatever experiment you're doing with your vibe coding, whether you're making a resource center or an application or whatever you choose to make, the little pieces that you learn about, you can then apply across your work, if that makes sense.

Zohra: 29:29

Yeah, yeah.

Liz: 29:30

So that's the experiment I've been working on.

Zohra: 29:33

And it's been a fun experiment. And you invited me to join you on that experiment.

Liz: 29:39

Absolutely.

Zohra: 29:41

So what Liz was trying to build was an onboarding app. And then she wanted a version of it that complied with our branding. So she invited me to see if I could do that. And like Liz said, it ended up becoming a separate project in a sense. So I went off and started doing my thing and I over-engineered it. But again, having access to the branding through SharePoint, having access to the components in the library that we needed to build, it allowed me to write code. The idea was to with the collaboration, I wanted to build something and I was able to build something that worked just like you were able to. So we had two versions of the same application. But what brought us to that point was the fact that when you started, there was a problem that you were trying to solve. So again, re-reiterating why were we? It's not about the tools, it's not about the cool stuff that AI can do, but thinking about actual problems. And now you're at a point where we may want to actually build a tool, an onboarding tool, and actually use it, which is fantastic. Can you imagine a year ago technical writers even being part of that initiative?

Liz: 30:58

It's fascinating how quickly our skill set is growing and also how much the lines are blurring between some of these different roles as a result. And I definitely like two years ago, I never would have thought about designing an app. I never would have thought about coding an app. But you know, the part, the thing that I think is interesting about it is that I'm taking knowledge management skills because the idea behind this onboarding app was again, we have all of this information in all of these different places. And somebody who's new to the company isn't going to know what to look for or where to look for it. There are certain skills. We want some two-way communication between this new employee and their mentor. So let's give them a place to do that and to say, these are the questions I have and these are the things I need to be able to do. Where are the resources? And just have it right there at their fingertips. It might not be changing where those actual resources are located, not moving them out of the places where people know to look for them who are not new, but putting all the links to them, putting a list of them in one place for them to just have a little library of their own that they can refer to and get out there and actually know what to look for and links to where they are. And so I think what's really interesting is that we're taking the skills that we've developed as information architects, as knowledge managers, as writers, and we're able to extend that on projects that normally we wouldn't have greenlit. Because if I had gone to my product team and said, hey, engineers, I have this idea. We really need to be able to share this information faster. This is going to cut down at least in half the time of onboarding. Engineers would have said, Cool story, but no, no, we have a product to build. We don't have time for that. We can't build that. So, like, it's giving us the green light to move into areas that aren't going to be funded, aren't going to be resourced, aren't going to be staffed, and just get things done that just we know that this is a win. We know that this is a win for the company and for us. This is going to give us bandwidth to get back to the things that we need to do. And I see that as being that AI collaboration that where we're not giving up our job to it, we're not losing any of our autonomy. We're not saying you can replace me with this, but we are saying my time is valuable and I can get some of it back by using this tool in this way.

Zohra: 33:36

Beautifully said, Liz. Many, many years ago, about four or five years ago, I was asked by someone, generative AI was not even a thing at the time, or it was in development, but the world didn't know about it. This is before ChatGPT. And I was asked, what will happen if your job is taken away? And I'm like, I had never interacted with AI, but I said, when it becomes available, whatever that tech is, I'm gonna jump right in and have fun with it. And I think that attitude, that mindset has helped me despite the fear. And I still have that fear, no doubt about that. But the more opportunities that, like, the lines that blur and you're able to give yourself the permission to solve that problem, but that comes from a place of being open-minded and experimenting with it. So great points, there, Liz. So is that why our presentation became less about choosing an AI tool and more about examining what needs to exist around it? I think that is the part that I want to dig into with you. So as we've talked, looked at the different experimentations that we've done, it's it's really not about the tool, right?

Liz: 34:40

Because you can we have a bunch of tools that are kind of redundant and do a lot of the same stuff. They might do it a little bit differently and they might do it to different levels, but you can't use any of them if you don't have your resources set up in a way that AI in general can access it. I feel like that a lot, like I feel about things like Flare and FileMaker or FrameMaker, sorry. Flare and FrameMaker and you know some of the other publishing tools. Like if you can work in one of them, you can probably work in the others. Some of them are better at certain features, and some of them there might be a best option, but if you had to, you could make it work with one of the others. But you're not going to be able to work in any of them if you don't have your resources set up.

Zohra: 35:29

Right. And the very first agent that we built, Liz, I think you mentioned we used Copilot Studio. But with uh the onboarding tool, we are not using Copilot Studio anymore because it's not easy to build things in that.

Liz: 35:41

No.

Zohra: 35:41

We are using GitHub Copilot and now Claude Code.

Liz: 35:46

Claude Code. Yep, that's my latest go-to is Claude Code.

Zohra: 35:50

Right. And so being tool-agnostic is, I think, another angle that we are looking at. Don't get married to the tools because the tools will come and go and they'll evolve.

Liz: 36:02

And we can all see that they are changing every day. You might have access to one one day, and then the next day you find out, oh, wait, they had a security breach, or they got bought by another company, and that's changing faster than we can keep up.

Zohra: 36:15

Right. But the interesting thing here, Liz, I feel is as much as the tools are changing on us fast, I'm amazed how quickly humans are also adapting with those tools, right? Nobody talks about that. Everybody talks about, oh my god, AI is doing these amazing things. But can you look at the pace at which humans are also aligning with that? We are not giving ourselves enough credit there.

Liz: 36:39

That's true. That's true. I mean, and especially if you think about just the attitude towards AI in January compared to now, there are a lot of, I mean, we still have a lot of the same ethical concerns that we had. But just in terms of the fear of getting started, just the fear that this is something that's going to replace me rather than this is something that I can work with, the fear that I'm not smart enough to do this or this is too hard. I think that the resilience we see is the same resilience we see with every major technological advancement. You know, I mean, you have the printing press, you have the telephone, you have the television, you have the computer, the internet. Yeah.

Zohra: 37:22

Right. Starting with the calculator, right? Right. We are going to unpack a lot more as we get to the final presentation. But what will we be able to unpack at LavaCon as we get there that we've only begun to explore here, you think?

Liz: 37:40

I think one of the things that we're going to be unpacking with a little more time and something that might be of interest to our audience is this idea of agents, MCPs, skills, apps, when to use them, which one understanding which one you need, which one you have time for, which one is going to meet the requirements that you have, and importantly, what your resources are able to interact with. Yeah. One of the things I really hadn't thought about going into all of the AI experimentation was how the way that we have been doing things for the last several years are going to affect our ability to do things in these new tools. For example, some of our content is held in repositories that you can't access with AI directly. Like AI is not allowed to go in there. And so, what do you do at that point when your style guide is behind the firewall? Let's say you have a great library that's really human readable, but you haven't put any metadata on any of those files and it goes back for years. AI is going to have a really hard time parsing out, you know, what what you need and when. So I think that between the two things that are mostly of interest to me that I think will be of interest to other people that will actually help them as they're going on the journey are one, setting up your resources appropriately, and two, using the right tool for the right job. And I'm not talking about tool in terms of, are you using Grok or Copilot or Claude. I'm talking about, do you need an agent? Do you need an application? Do you need a skill? And these are our bigger concepts that are still evolving. I'm certain that they're going to be adding to that list of those tools before the end of October. Yeah, I think that we can help people think about it in a way. And I think we'll also get some great input from our audience that will help us think about it in a way. That you don't end up reinventing the wheel, you don't end up over-engineering, and you just have the right tool for the right time.

Zohra: 39:54

We are almost at time here, Liz. Who do you particularly want to see in that room? And what do you hope they leave questioning about their own content operation once we have everything nicely tied out for them?

Liz: 40:07

I think what I would like to see in that room is a wide range of experience and anyone willing to share. I think that we are at a point in AI that we are all still learning from each other. And while Zohra and I have a lot to share with you about what our experience is, we want to hear about your experience too. And I feel like every day, almost literally every day, I learned something that I didn't know the day before about how I can incorporate AI into either my workflow, a tip. Sometimes it's not about work, sometimes it's about how am I going to get a new meditation? How am I like maybe something to lift your spirits? You know, setting up your fantasy football team. There's all kinds of tips out there. You can you're gonna find them on all of your social medias, you're gonna find them on your YouTube, you're going to find them on forums. There is no lack of information out there. There is too much information to digest. So, what I would really love to see come out of this is conversations with real people about what are their needs, what are their successes, what are their failures, and just the ability for us to learn from each other. I mean, we're coming in, we're starting that conversation. I'm hoping that they're gonna continue the conversation with us.

Zohra: 41:30

Great points, Liz. And yes, learning from each other is critical as we lean into each other. So that's my goal as well. So thank you so much for getting on the show and helping me unpack. I've learned so much from you. This is gonna help me with my outline now. Have fun, Liz. Thank you. Alrighty, I'll see you all at LavaCon. Bye. Thank you for listening to Inside Tech Comm. If there's a guest you would love to hear on the show, please let me know. And don't forget to follow and catch every episode on your favorite podcast app. See you next time.