Podcast: Play in new window | Download
Subscribe: Apple Podcasts | Spotify | Android | TuneIn | RSS
Like many content designers in the fall of 2022, Andrew Stein was concerned about the possible negative impact of generative AI on content and design practice. And his concern was heightened by the large number of content designers on his team.
Since then, Andrew has discovered many ways to apply AI in his content design work, both in conventional digital-product design and in content work on AI products.
He has also discovered a happy additional benefit of taking the lead on AI. His expertise has led to exciting new collaborations and leadership opportunities.
We talked about:
- his work as a content design and AI leader
- his take on the best ways to use AI in content-design practice
- how to maintain focus on the fundamentals of content as you work with AI to create new content or manage and validate existing content, and a tool he is developing to automate this
- new content-employment opportunities that he sees emerging
- the clean slate on which content people can create their new AI roles and responsibilities
- some of his techniques for demonstrating how your content skills can help your AI collaborators:
- find opportunities to serve
- adopt a learner’s mindset
- “just do” – experiment with tools on your own
- some of the people he follows and resources he has consulted as he has developed his AI expertise:
- Noz Urbina
- Leah Krauss
- the conversation design community, in particular Maaike Groenewege
- his encouragement for all content designers to find a balanced approach to incorporating AI into their career
Andrew’s bio
Andrew is a Director and Principal Content Designer at a financial services company. He’s led content in smart home, social media, AI robotics, and FinTech. Andrew’s experience includes both consulting, and companies like Lowe’s, Wells Fargo, Truist Bank, and Meta. Andrew is currently focused on the way AI tools serve the content design process, and bringing a content-first approach to the development of new AI products and services.
Connect with Andrew online
Video
Here’s the video version of our conversation:
Podcast intro transcript
This is the Content and AI podcast, episode number 34. When generative AI burst onto the scene there were plenty of reasons for content designers to be anxious. Andrew Stein channeled his concern into a deep exploration of AI tech and how it might be applied in content work. As a design leader, he has discovered a number of ways that content designers can use AI tools, and build AI products. As an advocate for content practice, he has found that his AI expertise opens many new doors for influencing his business collaborators.
Interview transcript
Larry:
Hi everyone, welcome to episode number 34 of the Content and AI Podcast. I am really happy today to welcome to the show, Andrew Stein. Andrew is an independent content design leader. He works currently for a big financial services firm. He also has his own consultancy on the side, does various content things including AI stuff for folks. So welcome, Andrew. Tell the folks a little bit more about what you’re up to these days.
Andrew:
Yeah, very cool. Well Larry, super-happy to be here and as I mentioned earlier, I’ve seen all the episodes and get so much out of them every time, so really happy to be here. Yeah, right now, like you mentioned, I’m doing quite a bit of work both for the company I work for and on the side, working in both AI projects and traditional content design projects and really where those two merge together, both helping to build teams and build structure around how we approach AI from a content perspective, which I think is really key with all of this. And also how to bring AI into the work that we do as well as content designers working on traditional products and services as well.
Larry:
Yeah, I think that latter is probably the more familiar scenario for most of my listeners, I guess. I do know a number of people who are working on AI products, but I think the more common use case for many people is using AI in their day-to-day, just good old-fashioned content design work. Especially as a leader, how are you implementing that and encouraging your folks and just tell me a little bit about that.
Andrew:
Yeah, well I think at first, all of us were wondering does this do the writing for us? Does this replace us? There was quite a bit of fear and trepidation or looking at it very cynically like, “Oh, this thing can’t do anything for me. It’s not a writer, I’m the writer in the room.” I think there’s been a spectrum of views on it, but all looking at it as the writer. Is it going to be the writer? Can it replace the writer? No, it can’t. And what I’ve really landed on, or at least at this point in time is that, no, it’s not the writer, but it’s a really great assistant to the writer. And so that’s really the perspective that I’m coming at it from with the teams that I work on, with my own personal work is really seeing it as a really powerful tool.
Andrew:
Noz Urbina, I think he said, and maybe I’ve inflated the number, maybe it was 100 and now I’ve made it 1,000, but I believe he said, “Think of AI not as your superhero that’s going to do everything and the magic bullet. But think of it as like 1,000 interns that can do way more than you can, but they can also do way more than you can really poorly with poor instructions or really great as long as you give them great instructions.” And so that’s really the area where I think AI fits in as a tool for content designers. It’s definitely not replacing you, it’s not going out ahead of you and doing all the work and then you’re wondering where you fit into the picture.
Andrew:
It’s very much human in the loop before, during, and after, and it’s kind of like a companion or a sidekick that can help you do things, can serve as another person in the room or a lot of other people in the room to give you feedback on ideas. But very much from that perspective and not nearly as much as, “Oh, it can go out and do it for me. I don’t really have to think about it.” I think you have to think even more now to use AI well, but if you do that, it can be a really powerful tool and that’s kind of the spot where I think we’re coming-
Larry:
…I don’t know, I’ve managed as many as 15 people at a time, but that’s only 15. And I’ve also done a lot of volunteer wrangling at work conferences and things like that. And that notion of managing enthusiastic and pretty knowledgeable, but really just not as far along in their professional development as you are, managing all that requires a lot of guardrails. How do you manage that? Everybody would love to have 1,000 precocious, brilliant people ready to help them, but they just don’t know as much as you do about the job. How do you constrain that enthusiastic creative energy that LLMs bring to the game?
Andrew:
They’re definitely way too enthusiastic most of the time, and I think if you try and just generate some content without a lot of structure to what you’re trying to generate, you get that way over enthusiastic, too many words, too many repeated words, that young person that’s learning a lot of cool words and wants to use them all. I think that’s what we see a lot of times. It really is about, and I want this to be an encouragement to all content people out there, is that really the fundamentals of content creation are still there and if you bring those into the scenario, you’ll have a much better experience with them. But it really is about diving into those core principles.
Andrew:
So when thinking about creating content or checking your content, being able to connect whatever AI tool you’re using to really good sources and good existing content, so like a style guide or a content design system, having that built in and really fine-tuned so that anything you’re doing is within those guardrails is really important. And obviously you can expand out from there. I think that’s the base. So if you’re using something like ChatGPT to help you ideate or to check your work, it’s really about, okay, again, like you said, guardrails, what are the guardrails that keep that content generation or that content check or that brainstorming companion within the scope of whatever you’re working on? So if you’re in an organization, making sure that your LRC guidelines are built into that. That’s another way to have checks and balances on the content you’re creating.
Andrew:
Even tying into research and personas and having all these different pieces of data that can create this world that LLM or that ChatGPT tool can live within and work through is really important. But you can come at it from a few different angles. So you can think about it like that content creation and you’re generating new ideas and new concepts or new content ideas and it’s coming to you already within that framework, or you can take existing content and check it against those things as well. Does this match those guidelines? And that’s where I think if you’re using a tool, you want to build that tool in such a way that it’s giving you the reasons that it’s making decisions or the pieces of data that’s factoring into what it’s giving you.
Andrew:
I think that’s been a really key thing for us, for the projects I’ve worked on is having that validation that, “Oh yeah, this made that decision considering this piece of information.” Because not only does that give you a reassurance that yes, I’m within the guardrails. It also tells you where it’s getting it wrong and where you can go update those guardrails to get better outputs. And then the more people that use it, you know that they’re all working in the same frame … It is just like your style guide. You want everybody to be referencing it. Well, it’s the same thing now, but now you’ve got a tool that’s also referencing it. So really you’re just thinking of it as another team member that’s really ambitious, has lots of energy, but needs a lot of instruction and a lot of guidance.
Larry:
Yeah. I know that there’s some commercial tools, that everything from like Grammarly for basic grammar checking, all the way up to more commercial tools that do specific style and voice and tone and guidance and even, I don’t know if they do … But I guess if you had content guidance in a design system, that could be one of those things. Do you use or have you developed any tooling around that sort of guidance? Like both providing the guidance and then ensuring adherence to it?
Andrew:
Yeah. So yes, definitely. So I’m currently working on a tool that does that. It’s a tool for various content designers to be able to use that they can go into it and they can use it for all those different reasons. Well, they can check their work and say, “Does this violate any of our guidelines? Does this violate any legal risk and compliance dos and don’ts? Are we using the terms, the definitions correctly? So when we have certain words in here are those terms used correctly?” Being able to reference those definitions very quickly, these are things that it’s like you don’t necessarily need AI for, but it makes it a whole lot more efficient being able to tap into a glossary of words and say, “Oh, I’ve got that word out of place. That’s actually on the do not use list. We’ve got this other word instead, so I need to replace that out.” So yeah, there’s lots of data pieces that we tie into and that you can put your content through and get feedback on that content to make sure it is within those guidelines.
Larry:
And I’m assuming there’s … I’m just trying to picture what the AI take capabilities bring to that kind of thing. Is building a tool, because I guess you could have done this five years ago, it was just like an algorithmically driven rules kind of thing, but I’m assuming that AI must figure into the tooling itself as well?
Andrew:
Yeah, so I mean, kind of the cool thing about that is that you can. You can connect these data sets to generated content. So you’re getting that LLM feature, you’re getting novel ideas and novel ways of approaching things that can also be checking itself based on some older existing technology like what we’re talking about.
Larry:
Yeah. And so you’re talking about, that’s like an internal project, but you’ve told me you do some user-facing AI products as well, which I think is … I’m trying to guess what percentage of my listeners might be doing that. I know I’ve had a number of guests on who are doing that, but tell me about, so the folks who are doing that work, are they new people? Have you repurposed your content design stuff? How are you building stuff for AI or building AI products?
Andrew:
Yeah, so what I’m seeing is that, and I think this is a common trend where it is existing content designers getting repurposed into AI projects. That’s definitely been my experience. And interestingly, it’s led to conversations about bringing in new people and building that out because now – I think this is an exciting time for content designers. Again, at first we were all wondering what’s going to happen? What’s going to happen to us? What’s going to happen to our practice? I haven’t seen jobs going away. I know they have. I know some companies with very little foresight jumped the gun and thought we can replace writers, but the companies that know what they’re doing and that have some foresight and didn’t rush into things, what I’m seeing there is it’s creating new opportunities.
Andrew:
All the existing product and service design teams are still there, but there’s also new projects happening and somebody’s got to do those. And if the organization is smart and they see how fundamental content is to gen AI products, well they know they need content people in there. And that’s absolutely been my experience. So now what I’m experiencing is seeing new projects open up, people hearing, “Oh, this content person, he knows about AI too. This generates content, we need that person in the room.” And so those teams are being built out with content involved. And one of the really cool things about that is there’s this contrast I’ve been noticing where the existing content designer dilemma, always wanting to be invited to the room but not getting there early enough. When they do get there, they’re kind of standing in the back of the room rather than at the table. It’s something we’ve all struggled with and some of us have figured out how to do it, others of us are still working on it. But you’re always fighting against this existing mental model of what a words person does on a project.
Andrew:
Contrast that with new teams that are being put together and there’s no mental model for what that team should look like. There’s no structure for, “Oh yes, if you’re working on an AI project, you absolutely have these people and these are their roles.” A lot of it, it’s so novel that we come to it with a really open mind. And so if you are a content person that has some understanding of how this works and how content fits into it, and you’re able to be recognized for that, my experience has been getting invited way earlier than I ever would on a traditional product or service project, and not only being in the room but having a real seat at the table. Why? Because you have ideas and you show knowledge and you’re able to kind of move into that space and say, “Yep, I can help with this. Oh, well you’re doing this, but did you know that there’s this too?” And it’s, “Oh man, this person’s bringing some value to the team. They’re a real part of this team now.”
Andrew:
And so I think that’s one area that I really want to encourage so many people to do it. It’s not throw content design, the traditional work out the window and just be this AI person, but have AI as a tool in your belt that you know about because these projects are happening and it creates a really cool opportunity to get in at the ground level. And I always say that the doers now are the leaders tomorrow, and we haven’t gotten to this place where the AI practice is so established and all the roles are so defined. So now those things are being defined by how well these projects go and who contributes to them.
So by getting in now and contributing in a really effective way, well then down the road, you’re part of AI leadership or part of leadership on teams where content is really recognized and valued. So I feel like this is a real opportunity for content design to really make a name for themselves, to really step in and say, “This is a content-based thing.” I mean, honestly, I think all UX is a content-based thing, and we just kind of missed that. If you go back to the foundations of the internet, there were no content people.
Larry:
Hey, wait a minute, I was there. I didn’t do a good job, apparently. Sorry.
Andrew:
Okay, there were. There were. But there was a lot of engineers and developers writing content because they knew the technology. They were in early, they knew the technology, somebody had to do it, so they just did it. Well, I think we’re at the same place, but there’s this whole content community that can step in and say, “We don’t need to do it that way. This time, I can be in here and I can really help build this.” So whereas we kind of built the entire internet without content people and then for the last 20 years, content has been fixing this broken internet. Now we’re at the point where we don’t have to do that again. Content can be in at the foundation and we can really build it the right way, not missing that really important piece of the content brain that we bring to projects. We can build it the right way and in the process really establish content as a key role in building things.
Larry:
Yeah, that’s probably a whole other episode just to go into the details of that, but can you talk a little bit more about specifically, because that seems like a really exciting opportunity for a lot of content people to … Especially I know a lot of people who are chomping at the bit to get into more generic design leadership roles and things like that. So what do you think would be like two or three things? I know for you, from our prior conversations I gather it was like your just genuine curiosity and just learning a lot about AI. But are there, especially in hindsight, because you really are one of the true pioneers in doing this, can you think of, I don’t know, just two or three things that people might … Skills they might develop or things they might learn to be positioned to do what you’ve done?
Andrew:
Yeah. Well, it’s really interesting. The way that I got on my first AI project, it was actually just I found out about something and I reached out to somebody and I said, “Hey, you’re putting a pitch to put this past some leadership people. I’m a content person. Do you want some help with the pitch?” Had nothing to do with the tool or the technology or anything else. It was literally just offering your value to somebody who could use it. And that’s something that I often encourage people to do is if you want to get involved in something, it’s probably not best to show up and say, “Hey, I’m an expert in this and you need me on your team.” That might not get you the results you want, but if you come in, and this is really I think a foundational piece of advice for your career. Period. If you want to be a leader, be a servant, be somebody that helps somebody else’s job go easier.
Andrew:
I was actually speaking with a senior, an executive, and I said, “I want to grow in leadership. How do I do that?” And they said, “Be somebody that people want to follow.” How do you do that? You serve people. So that’s point number one is find projects that are happening and find how you can serve those projects wherever they’re at. So if it’s just reviewing a deck and saying, “Oh, we could say it this way,” and “Oh, your audience are these people and they’re really concerned about budgets, so let’s really focus on the efficiencies this is going to give us. And oh, over here, it’s not about efficiencies. These people really want to see improved experiences, so let’s tailor the message here for this pitch.” Those kinds of opportunities are huge that will get you in the door and working with people so much faster than, “I’m a content expert and you need me on your team.” So that would be like, here’s how to get in the door 101, do that every time and I promise you, people will respond.
Andrew:
As far as being able to show up and know what you’re talking about, man, there is so much information and so many AI experts on the internet and on LinkedIn, and I definitely don’t want to be another AI expert, but I do want to be a continual learner. And one of the ways that I’ve done that and I think helps me when I’m in those rooms is be willing to be an idiot. Ask lots of questions. This is so new and for somebody like me and you, we’ve been in content for a long time and we’re supposed to know what we’re doing right now. And it’s really hard, I think, sometimes to go in a room and say, “I’m totally ignorant. I don’t even know what LLM means. How does this work? Is an LLM the same as ChatGPT and what the heck are tokens?” And you’re like, man, I’m an expert and now I feel like a baby. What do I do? You do that. You lean into that. You lean into looking like you don’t know what you’re doing because you don’t, and that’s fine.
Andrew:
So I think really having a learner’s mindset and that takes humility and that’s hard for all of us I think on some level. But have a mindset of humility and look to quality resources. Anything that says AI is everything, probably away from that. Anything that says, “Oh my gosh, AI is the worst and here’s why. It had this hallucination and caused this problem. It’s going nowhere.” There’s also valid concerns that I think we need to consider. There’s a lot of ethical concerns. There are limitations to AI, so look for the people that acknowledge those, the limitations, the pitfalls, the concerns, and also see the potential in it and lean into those people. That’s something that I’ve appreciated from what I’ve seen from this podcast. You haven’t been shy about bringing people in that are very critical of AI.
Andrew:
And so you learn from people that have gone before you and those are the people that I want to learn from. So I’ve obviously gained tons of knowledge from this podcast and then going and following the people that you bring onto this podcast and digging deeper with them. You had on Leah Krauss recently, I immediately reached out to her and I was like, “You’re doing some amazing things and you’re really focused on these principles that you talked about. How can I learn more about those principles?” And so she pointed me in the right direction, and actually they’re public, so we can all learn from those principles. So I’d encourage anybody go find those principles she posted and learn from them. I’m incorporating those into the work I do now. I don’t think there’s a single source you go to and you learn everything you need to about AI. It is very new and there’s lots of people that focus on a certain angle or another angle. So take your time, learn from those people and be willing to reach out.
Andrew:
One of the things I love about the content community is the generosity that we have. I can’t name a time I reached out to somebody and said, “You’re doing great work. I want to learn from you,” and didn’t hear anything back. I mean, it’s just unbelievable how much generosity there is in our community. So reach out to those people, reach out to me, reach out to anybody and say, “How can I learn more?”
Andrew:
And then just do. That’s the last thing. I was in an organization where we didn’t have access to any GenAI tools, but everybody knew the need for them and wanted to use them and wanted to learn. So we got creative and went outside the walls and we found creative ways to start building our own GPTs and just experiment with them because we couldn’t do them within the confines of the organization. So we did our own, not related to the organization, but we started learning just by doing. And so I think it’s about 20 bucks a month at this point to be able to have the account with ChatGPT with OpenAI where you can build your own, but you can build your own GPT and you can build it and then realize how poorly it works. And then you refine your instructions and you make it better. And then you connect with other people and they’ll give you tips like, “Oh, when you’re writing your instructions, be more clear about this, but don’t be too specific here because it limits everything.”
Andrew:
Little things like rules with ambiguity do not work great in a GPT because they don’t know how to differentiate the context-specific rules. So being cognizant of those, but you don’t learn that until you really get hands on with it. So get hands on with it. I mean, I think that’s the key to growing in this. You have to get your hands dirty.
Larry:
Yeah, I hope you can do a blog post about this. Get in the door, put on that lifelong learning mindset and then just do it. But hey, one thing you mentioned, thank you for mentioning the podcast, but there must be other resources out there. Do you have any favorite go-to educational places or just people that you follow that have helped you? Because you’ve obviously learned a lot this last couple of years.
Andrew:
Yeah, it’s just being immersed in it. But yeah, so there’s one person that I am really attracted to and it’s because of how all in they are on AI and also how incredibly critical and cautious they are with it as well, which is Noz Urbina, which we both know. I’ve so enjoyed the content he’s been putting out recently, the talks he’s been doing. I think it’s really important to not just think about the practical or how does this serve my business or how does this serve my personal brand or my growth, which is very important. But also how does this fit in to society? And what is this doing? And how do we respond to that? How do we hold tech accountable? I think that’s really important. So Noz has been a great one to follow.
Andrew:
Again, going back to Leah’s … Her appearance on here, she pointed me to some resources that Microsoft has. I’m sorry that Google has. Google has some phenomenal resources out there, just tons and tons of resources with workbooks and principles and modules. So I would encourage people to go check that out. I think there’s a lot you can dive into there and it really helps you to look at, if you’re working on an AI project, if you’re working on a tool that’s going to incorporate GenAI, how can you approach that in an ethical way, in an inclusive way, in a way that’s fair for everybody and accessible to everybody, and that gives proper disclosures? All these things are really key to think about, and it really takes … You have to think about it through that UX lens of what’s the experience of this person and how can they interact with this in a way that’s not off-putting, but also tells them very much what’s happening with their data, how it’s being used?
Andrew:
Those types of things are really, really important. And if we’re not careful with them, it can get us all in a lot of trouble. And I think any organization you’re working for will incredibly value you for bringing those points up and raising flags and saying, “Hey, if we go down this path, that could lead to some misconceptions that could really come back to bite us.” So having a lot of those resources, I think Google’s resources are phenomenal to help with that. So that would be another area that I would definitely point somebody to.
The other one would be conversation design. Just digging into that. So I recently brought in Maaike for a talk. You’ll have to pronounce her last name for me because I’m not very good at it.
Larry:
I am just starting to learn Dutch, it’s close to this – Groenewege.
Andrew:
That was beautiful.
Andrew:
Yeah, I think she’d be happy with that and more happy that I didn’t try to pronounce it poorly. Maaike has been a phenomenal resource and also a genuinely amazing person to connect with and learn from. So I would definitely say to check her out. Again, I think she’s one of those people that is very excited about it. She’s leaning into it and she’s not married to her own ideas. As the technology changes, she’s willing to say, “You know what? This might kill my job.” Every career she’s been in, tech advances have killed her job and she’s happy to move on and say, “All right, now what does it look like?” And that’s really important. You have to be nimble. You have to have a malleable mind to be able to hold on to the core principles of what makes content design work and what makes UX work. And also be willing to say, but the rules of how we do that are changing drastically, and I think she’s a phenomenal person to learn from in that space as well.
Larry:
Yeah, I think there’s good reason that Maaike is everybody’s favorite conversation designer. Hey, I can’t believe it, Andrew. We’re already coming up close to time, but before we wrap up, is there anything last, anything you want to revisit from the conversation or just make sure that we share before we wrap up?
Andrew:
Yeah. When I started in AI, really what led me to it was seeing the content community faced with a really perplexing scenario that they might all be replaced, we might all be replaced, or our jobs might … This industry might go away entirely. What’s going to happen? There’s so many different responses to something novel like this, and I think what drew me to this space was I saw all of that going on, I saw a lot of chaos, just anxiety and fear and confusion, which I think is normal for anything, anytime you have the rules changing very quickly. It’s unsettling. And so what kind of drew me to this space was, okay, I want to be a voice of reason and I want to be a help to my community in this space. What does that look like?
Andrew:
And so my encouragement to all content designers are find that middle ground. I don’t think AI is going away, but I also don’t think it’s the end all, be all, and everything’s going to change, and everything will be automated and nobody’s able to work anymore. And whether you think that’s going to lead to sipping margaritas in Isla Mujeres in Mexico, or you think we’re all going to be suffering, find that middle ground and say, “Okay, if this is a part of our world moving forward, if this is a part of the UX landscape and how the internet works to some extent, where do those fundamental content principles meet this changing environment and how can I contribute to that?” I think if you just start there, you’re going to be in a really great spot to participate. Yes. What I would say to people is, yes, I think content design will change. I don’t think it’s going away and the name might change because the name changes all the time, and we’ve already got 15 names for it anyway. Don’t worry about that.
Andrew:
Writing has changed throughout history. The way we write, the way we create, it’s always changing. So that’s not a new concept. This particular iteration of change is right here, right now, and we’re here for it, but that’s not new. This happens all the time, and it’ll continue to happen. In five years, there might be something else that is changing the way we do our work, but don’t worry about that. Just find, okay, what is it that I know and I bring to the table, and how does that fit into a new way of working? And I think we’re going to be okay.
Larry:
I really appreciate your perspective. Thanks so much, Andrew. Hey, one very last thing. If folks want to connect or follow you online, what’s the best place to find you?
Andrew:
Yeah, so the only social media that I’m on is LinkedIn, so that’s the best place to do it. I’m there and I’m pretty active. We do content design office hours once a month on there, so you can find me in those office hours, do a talk or two every now and then, and I’m just always open to chat in the DMs or connect on … Well, actually ADPList as well. I’m on ADPList so you can-
Larry:
That’s right. You’re a mentor there.
Andrew:
Yeah.
Larry:
Yeah. I forget how we … But anyhow, one of those media is how we first connected. So anyhow, great to connect and great to reconnect. Thanks so much, Andrew.
Andrew:
Yeah, thanks. I’m super happy to be here. This was fun.
