Greg Nudelman
The Role of UX in the Age of AI
In this episode of Brave UX, Greg Nudelman offers insights into how AI is transforming UX 🌊, what UX professionals need to do to stay relevant 🏄, and whether the demand for designers will decrease in the future 🤷♂️.
Highlights include:
- Does AI pose a threat to the role of UX practitioners?
- What’s the quickest way to derail the design of an AI project?
- Which UX skills will still be valuable in a world shaped by AI?
- How can we ensure that AI models create objectively ‘good’ design?
- Will using AI to support design best practices lead to greater respect for design?
Who is Greg Nudelman?
Greg is an internationally recognized UX author, speaker, leader, and veteran of over 30 AI projects 🤖. And he has recently launched the website UXforAI.com, which features a newsletter dedicated to leading with UX in the age of singularity.
He has written five popular UX design books 📚, including "Designing Search," "Android Design Patterns," and "The $1 Prototype." His sixth book, "UX for AI," is coming in late 2024.
Currently, as a distinguished designer at Sumo Logic, Greg is responsible for special projects, including innovative AI/ML solutions for the company’s security 🔐, network, and cloud monitoring services.
Greg has also notably been a Senior Director and Head of Design at Logic Monitor, Head of UX at Aible, Senior Director and Head of UX Research at GE Digital, and Head of UX at Cisco Cloud Analytics 📊.
Before joining the corporate design world, Greg was the Principal and Director of UX at DessignCaffeine ☕️, a consultancy where he provided product design and UX strategy services for Fortune 100 companies like Cisco, IBM, and Intuit, as well as major non-profits.
Transcript
- Greg Nudelman:
- The good news, you already have all the tools. As a UXer, you already have all the tools to solve that, so all you need to do is just be brave. Like the brave podcast. You need to be brave, a brave UXer, and ask the right question. Ask these uncomfortable questions and men act on that.
- Brendan Jarvis:
- Hello and welcome to another episode of Brave UX. I'm Brendan Jarvis, managing founder of The Space InBetween, the behavior-based UX research partner for enterprise leaders who need an independent perspective to align hearts and minds and also the home of New Zealand's first and only world-class, human-centered research and innovation lab. You can find out more about me and what we do at thespaceinbetween.co.nz.
- Here on Brave UX, it's my job to help you to keep on top of the latest thinking and important issues affecting the field of design. I do that by unpacking the stories, learnings, and expert advice of a diverse range of world-class leaders.
- My guest today is Greg Nudelman. Greg is an internationally recognised UX author, speaker leader, and veteran of over 30 AI projects, and he has recently launched a website called UXforAI.com, which features a newsletter dedicated to leading with UX in the age of singularity.
- He has also written five popular UX design books, including Designing Search, Android Design Patterns, and the $1 prototype. His sixth book UX for AI is coming in late 2024.
- Currently as a distinguished designer at Sumo Logic, Greg is responsible for special projects including innovative AI/ML solutions for the company's security network and cloud monitoring services.
- Greg has also notably been a senior director and head of design at LogicMonitor, head of UX at Able senior director and head of UX research at GE Digital and head of UX at Cisco Cloud Analytics.
- Before joining the corporate design world, Greg was the principal and director of UX at Design Caffeine, a consultancy where he provided product design and UX strategy services for Fortune 100 companies like Cisco, IBM and Intuit, as well as major nonprofits.
- He has designed software used by millions worldwide, helped author 24 patents and delivered over 100 keynotes and workshops, and today he's here with me for this conversation on Brave UX. Greg, a very warm welcome to the show.
- Greg Nudelman:
- Thank you so much for having me, this very kind introduction. I really appreciate it.
- Brendan Jarvis:
- You made it really easy for me, Greg, you did all these things and if my sources are correct, I understand that you first developed your appreciation for research when working as a chemist. What's the story there?
- Greg Nudelman:
- Well, actually it's interesting. A lot of UXers, at least in the age when I joined the industry, a lot of them came from different biosciences and I think it's sort of this appreciation for letting whatever happen happen and just sit back and seeing what will happen if we tried this and with the chemistry, it's sort of almost like the hand of God or the hand of the universe, if you will. If a certain reaction fails, it may be you or screwed up, or it may actually be just never going to work. No matter what you try to do, that particular reaction just is never going to happen. There's no explanation, there's no sort of mental model that needs to be put in place. It just doesn't work. When you approach design from that perspective, it teaches you a lot of humility and I think that's something that a lot of folks probably could use more of.
- I think all of us could probably use more of it, myself included, but if you approach it sort of from the standpoint of humility and just see what works and keep an open mind and keep a happy mind even if things don't work and don't get so attached to it being a certain way because 10 papers tell you this reaction is going to work when you try it on this different slightly different compound, it just doesn't work and that's okay because 10 other ways to do it, and that's up to you to figure out. Now just having that kind of open mind, I think that taught me a lot when I joined the industry.
- Brendan Jarvis:
- Yeah, it's making me recall that. I'm going to mess it up, but there's that famous Churchill quote about success consists of going from failure to failure without a loss of enthusiasm or something to that effect, just keeping on keeping a positive attitude and keeping on going.
- Greg Nudelman:
- Exactly, and it's very much like the Thomas Edison I believe had a very, I love a Churchill, anything he wrote, it's fantastic and one of my favourite writers, hugely underrated. I think so yeah, very much like what Edison I think tried when they said, Hey, your light bulb failed 300 times. He's like, well, I just found 300 ways it doesn't work. And I think a lot of what we do as UXers has been reduced to guessing correctly or guessing in the most mediocre, safe fashion, and that is not going to work in the age of ai. I can tell you right now, we got to really put ourselves out there. We got to go back to the roots. We got to go back to sketching and creating and thinking about things from the ground up, from the customer perspective, from the business perspective, bringing those together and creating something completely new that hasn't been tried before because the stuff that used to work isn't going to work now the stuff that gathers here is not going to get us to that next stage.
- Brendan Jarvis:
- You've written plenty recently on UX and its relationship with ai and as I mentioned in your introduction, you've also founded a website called UX for ai.com, where that writing's available. I just want to quote you in your own words that are on the website here. You've said that the website is dedicated to leading with UX in the age of singularity and perfecting the art of communication with our robot overlords. Now that last part about the robot overlords is somewhat tongue in cheek I gather, or is it or are we humans, US UXs? Are we on the way out?
- Greg Nudelman:
- I'm a huge fan of science fiction and that's definitely tongue in cheek at least at this point in time, but I think a lot of folks, even on sort of the investor side, as much as I'm not a huge Elon fan after all the stuff that he's been doing, he's been doing, there's one thing he does talk about is sort of the dangers of AI and maybe what he does about it is stuff that helps Elon and helps his businesses. In other words, let's not have little companies experiment with this stuff because it's going to interfere with my agenda, that kind of thing. But he has been, I think correct about saying, look at this technology not in its current incarnation, but look at it in its future form. And the impact of AI I think cannot be understated. Of course a fair bit of hype that's included and we've seen this, but somebody who's been at it for 15 years and we've seen go up and down but consider the impact for example of internet and then the mobile right after that, mobile happened a lot faster and the changes were much more sweeping and to the point where we cannot imagine our lives without our mobile overlord.
- Brendan Jarvis:
- So
- Greg Nudelman:
- True. As we spend hours praying to our devices, literally we spend more time praying to our devices than we spend praying to our gods. By and large,
- Brendan Jarvis:
- I sat beside someone today in traffic and I looked over and they were driving, actually they were stationary, but they continued to drive with TikTok on in the car watching it as they were driving. So yes, it is a problem
- Greg Nudelman:
- Of course, and just think of AI as going 10 x faster and having 10 x times the impact. Yes, we are in the words creating DiUS x Machina, we're creating our new guards from the machine and from all the data that we're generating, but I think the ultimate power of this thing is not necessarily to transform our lives in sort of a negative way. I don't think it meant as a negatives. I think of it as augmented intelligence instead of augmented intelligence. I'm thinking of it as instead of artificial intelligence, I think of it as augmented intelligence. In other words, we're taking what is an artificial intelligence creation and then we work with it to then empower ourselves to create something that we would not be able to do on our own. Now, humanity's superpower has always been this ability to come together in large bands and create something that a bento, chimpanzees or bonobos just cannot accomplish, or even a band of Neanderthals, right?
- Just this ability to tell a common myth that ties everything together and enables people to independently pursue a similar agenda and work toward the same goal is I think we reaching an unprecedented turning point in that and AI is a critical aspect of it because it's able to take all that information from all these folks and then put it at your disposal and just put the essentials at your fingertips so you're not overwhelmed by all of this, but you just got the essentials enough for you to then take the next leap into the unknown and create something new. And so in this way it is enabling us to be the best next version of ourselves, but it's also there's a lot of potential for misuse and the folks that are going to be able to use it better, of course they're going to prosper in the next 10, 15 years and the folks that really struggled with it are going to be unfortunately left behind and there's not a lot that we can do about it, but from all of these standpoints, it is kind of our next overlord if you will, even more so than mobile has ever been.
- Just think tenex, that is
- Brendan Jarvis:
- That existing digital divide in our society already, which it sounds like you are suggesting that's even going to be more greatly amplified by the arrival of AI and the adoption of AI by people that are already on the more privileged side, if you like of that divide. I'm curious about your perspectives on the emergence of generative AI as a father. As I understand you have two children, one in junior and one in senior high. How rosy is your perspective when it comes to the world that they might find themselves living in as adults say in 10, 20 years?
- Greg Nudelman:
- I think I'm actually extremely hopeful because they grew up as sort of digital natives. I know this term has been overused and abused, but what it really means is they're comfortable with part of their mind existing out there or part of their identity existing out there in sort of the interwebs. And they're also comfortable with the ambiguity and not being always right and not knowing everything because they understand how much there is out there to know and it's impossible for a single human being to really do that. And I think that is a necessary, it's required and necessary condition for us to move past where we are stuck today with a lot of the unfortunate right-wing tendencies. This is brave. We can talk about that I suppose, where people are really having a lot of grievances about stuff that they feel so sure about, feel so convicted and convinced that they're right and the other side is wrong and it becomes sort of concentrated down to this very minute, almost prosta slice of information that they're super stuck upon.
- And I think moving on to our next evolution as species and as humans, I think it's going to take us moving past that and say, I'm okay with not knowing, and that's kind of what I see in my kids is they may not know everything about ai, but they're comfortable maybe trying things out and going, well, I don't know, but maybe I can find out. As a parent, I feel like that's the critical thing you can give your kids is maybe to say, I don't know it yet but I can go figure it out. Just having that empowerment, having that sort of posture for lack of a better word with which to approach the world is I think a very powerful and empowering posture and I feel like I've done the duty to do that and either way either have done it or I haven't.
- They already passed the age where I can do anything about it, so they're off to the races, but I think that's kind of the key and I think they're very comfortable with it where I think we are failing them, if you don't mind me expanding on this a little bit more, where we are failing them is our education system. Education system is really is stuck on, well, they have to do their essays by chiselling on the stone pyramid and they have to tie it to the pigeon leg and it said the pigeon out of the castle window that's so archaic, it's unbelievably archaic and that's not the real world at all. Now, that's not to say that there's not a lot of value in writing pure essay for instance, but a lot of the essays that they are asked to work on, a lot of thoughts that they're asked to explore are extremely formulaic to be honest.
- I mean growing up in Ukraine we've had a lot more interesting conversations and maybe they were more about math and chemistry and things that were more safe because we couldn't talk about the Soviet regime and we couldn't necessarily discuss the role of Stalin and his relationship with Hitler, for example in the commun of World War II and the massacre that befall the Soviet Union in 1941. Those were not safe topics, but talking about molecules and really exploring that relationship and having in-class discussions about this stuff and then bringing in bits and pieces from works and the research and just kind of making it sort of a heady brainstorming mix to try to really understand and poke the idea. And I think what the US and by extension I think a lot of many other developing nations are falling into the strap of saying, well, no, we're not going to use AI in the schools and we're not going to allow this.
- We're not going to have them do this because it's interfering with that sort of pure educational content and I think that's just pure bullshit if I can just come on and say it. It's so not true. I think there is a time and place for everything. We're looking at the letter of the law rather than looking at the outcome and it's not just they're going to employ AI when they come to the workplace in a large way, but it is now is the time to really figure out what can be next and how can we apply, for instance, AI to music novel interaction paradigms that come with creating digital art and what does it mean to tell a story that other humans understand and why shouldn't we learn from a collective knowledge of billions of other humans that is put into this tool that can kind of guide you toward perhaps a better level of expression?
- Why wouldn't you want to do that? It just seems so valuable. Yes, there's definitely a lot of questions. There's still a lot. We've got to figure out how do we grade these papers and so on, but we can't just put our blinders on, put our head in the sand and say, well, we can't figure it out. So no AI in the classroom, that's just not going to work because people, they're going to do what they're going to do. And with the TT four oh, I think what we see is it's actually an excellent tutor. It comes down to the level that you're at and it explains it in multiple different ways until you really understand. It gives you a lot of examples. It draws pictures, it does diagrams and video and to the point where this just wasn't available even a year ago, and to reject that wholeheartedly and completely is absurd.
- It was absurd. It's like saying, well now you have to make this desk with one tools from 200,000 years ago. It's absurd. It's just not going to scale. So we need to find a better way as a matter some urgency. So I'm glad they're graduating out of a public school and it's a great education system where we live fortunately, but it's one thing that's lacking is sort of vision for what is next now that AI has come into its own and is here to stay and it's affecting how people are learning, working and communicating. Let's use it. Let's figure out how are we going to do this Next thing,
- Brendan Jarvis:
- Let's talk about that in the context of designers and I'm going to include, call it design people, so anyone that was working in the broader field of design, whether you're a researcher or a designer, you were recently interviewed on the future of UX podcast and you said, and I'll quote you again now, you said, I think the good news is we're actually going to need more designers than we have today. I think it's going to be a very useful skill. The challenge is what parts of the skill are going to be useful? So just picking up on what you were saying there about how are we going to use it to best effect possibly this unsurprising question off the back of that quote, what parts of the skill do you see will be useful?
- Greg Nudelman:
- I'm so glad you asked that. I think it's fantastic and a lot of, I guess it's not just me saying this, but if you don't believe me, you can listen probably to Jacob Nielsen and his podcasts. I think he's got a lot of the same kind of insights that are coming in from watching the industry and seeing what emerges. So I'm in very August company indeed, I would say so it is definitely an honour there, but the skills that I think are going to be very useful are the skills that I've already been interviewing folks for my teams for many years is really, can you draw me a picture? Can you tell me the scenario? Can you use, if I tell you the scenario, can you draw me a storyboard? And then using that storyboard, can you give me a set of screens, a potential workflow that this person can go through the system and accomplish what it is that they want to accomplish and understanding and asking good questions.
- Sometimes uncomfortable questions like How do we make money? Is this really what they want at the end of the story? Let's say they get a bunch of IP addresses and is that really what they want or do they want, can we take it a step further and really give 'em an answer? I think with AI there's a lot more possibilities and this ability to imagine something out of nothing is literally what design was defined as. It was a definition of design. Can you imagine what is possible to do with this technology based on the person's need? How do we address this person's need using technology that we have or can we come up with the tech that is going to help them actually do that? The very first thing I think that really was a significant advance, at least according to literature that I've read, has been the spreadsheet.
- And so when you think about the spreadsheet is just a very infinitely useful paradigm, the student, the test of time, can you think as a creator of a spreadsheet about your particular user and your particular problem, can you apply the same sort of overuse the term, again, design thinking, but literally just means problem solving, understanding enough technology to say, if we took this piece A and piece B and we did something with us together, can it not produce the outcome for this person that really needs this outcome Y and then it will create that benefit Z for them. Ability to do that I think is critical and I think it's lost because so many people are just laser focused on how can I use Figma to draw my responsive layouts or the layouts? People just fixated on that stuff. Well, let me tell you, that is gone.
- It's dead men walking, that kind of stuff because that's exactly the kind of stuff that AI is really, really good at and it's going to figure that out and forget the Figma. It's going to go straight to react, so you basically going to get a react component that already has a built-in CSS box and whatever else that it needs to be fully accessible and certified and pluggable into all kinds of different layouts and so on. What we really need to figure out is what are we really solving for? What is the benefit to the customer? What is the benefit to the business? How can we bring them closer together in an ethical way that is actually doable with the resources and the components and the backend in information and the models that we have in stock? And if you answer that, you're going to be, I think a very happy customer in the next five, 10 years.
- If all you focus on is all the layouts for Figma and how to do Figma stuff, you may find employment in a way that is going to be more design system focused and ability to take components and creating components and react, but you're going to have to learn a lot of react and or whatever else comes after the react, but you're going to need to bridge that part or you're going to need to tell more of a story because what I think 80% of the folks are doing is high def Figma pictures essentially is not going to be very useful. Even I think a year less than a year from now. I anticipate it's a matter of months if not weeks to be honest. It's going that quickly and it's going to be that disruptive.
- Brendan Jarvis:
- Yeah. You've previously said before, and I'll quote you again, it should be obvious by now that we will need far fewer designers. Today's designers like barefoot peasants, cutting omega harvest of wheat stalks with a sickle and threshing it by hand, sweating over producing and maintaining every single pixel for table after table. Gosh, that sounds dreary, doesn't it? On one hand though, that was back in early 2023, I think it was an article you wrote called UX is Dead, so that's kind of a year, 18 months ago where you've sort of said, look, less designers or fewer designers will be required. And then more recently in that podcast interview I mentioned on future of UX, you suggested, oh, we need more designers. So I know with the passage of time people can change their opinions, new information comes to light. I'm not going to hold you to what you said in 2023 if it's changed, but I am curious to understand what has changed for you that's taken you from a position of framing it as fewer designers required to a position of framing it as more designers are required.
- Greg Nudelman:
- Yeah, I think both are correct in a way. We definitely need a lot fewer designers that are doing what is currently done as understood by design, which is in many cases it's, pardon my French, it's purifying something that PM already created in their mind or in the sketch or in balza or on a piece of paper requirements or in Jira ticket, whatever. It's if you are a designer that takes in basically ready-made requirements from somebody else and is in charge of converting them into high depth Figma designs, we can need far fewer of that type of skill. What we are going to need more of is figuring out how do we incorporate AI into the products in a way that benefits the customer, actually solves the customer needs. It's a very different skillset. We kind of lump 'em together as designed. It's calling everything designed, but I think that's why it creates this kind of confusion like we're definitely going to need less figma jockeys that convert PM requirements into high do wireframes.
- Lemme put it that way. That part is very much dying. The other part, however, is not only alive, but there's just nowhere near enough people at the moment that I see that are capable of doing that, which is how do you tell a story? How do you convert a challenge into a solution? How do you actually understand what's workable from the technology perspective? How do you engage and create a vision for a product? How do you do competitive analysis in a way that helps you and the team understand what they're capable of and where your sweet spot is and how your customers may be different from their customers? That sort of thing. That's very in depth, being able to talk to customers, being able to converse and extract from that ideas and then recombine that tube and create something new. And I cannot do any of those things.
- It probably never will, but what it is really good at is taking a requirement and making a wireframe out of it using components that are already created and react. So if you say, I need an address form or I need a table, well, guess what? They can do that, right? You can say, I need a table that has a name and ID and a time, and then every time you sit down and you make a table like that, you should be thinking that's dinosaur work. That's robot monkey work that's gone. That is waste of time and it's always been a waste of time and it is definitely a waste of time today. What I should be asking, is this table going to answer the question that the customer really wants to know or is it better presentation something completely different? Should I write a paragraph?
- Should I have AI write a little paragraph explaining what it is? Or maybe I can suggest solutions. I can say, well actually it looks like I've analysed the data for you. It looks like you're having this and this and this and these are the three things you can do to solve this problem. Which would you like to do? I mean, that is a much more capable thing. Most products don't do that today and most people aren't thinking that far ahead. So if you're getting requirements, I need a table that has ID name and time of day, maybe some description or something, and all you're doing with your day is creating a table like that in high definition that is just not going to cut it a year from now, it's just not going to cut it. But the other one definitely there's not enough people because there just aren't that we've practised for better or worse, I think the economic pressures has reduced our industry to doing 80% of the high def wireframe of a table from a requirement, and that is something we need to collectively fight against.
- And the way that we do that is not by fighting but by providing value. And that is very much what Book six is about. How do we transition from being what economy has forced us to be and into something else? And I think a perfect example of this is content management, if I may, because I think that Brendan, you hit exactly I think the nail of the head. This is like the heart of the discussion. If you look at content management things like now 10 years ago, and I may be dating myself very, but I've been in this industry for a very long time, over 20 years. I'm not going to say exactly how long because that's really embarrassing, but over 20 years I've noticed this trend, right? But 10 years ago we've had an actual, I was at Wells Fargo doing consulting for this mobile project and there was a content person present at every kickoff that was a requirement for the kickoff.
- To be successful, we had to have a content person and that person had to review all the artefacts and they had to comment on it, and they designed every label, every button, everything had to be aligned, and then they wrote all the docs for it and only then it was considered acceptable and allowable for shipping. Now, today, or even fast forward five years, let's say a ge, we had two content people that were mostly working on the docs portion and then we ourselves did all the work on the buttons and the labels and the error messages and warning messages and so on, which is actually pretty important stuff. If you think about your BP disaster or something like that, that GE was not involved in, it was before that relationship. But if you think of the magnitude of stuff going wrong, potentially if you tell the wrong message or if you signal the wrong way, it's pretty critical.
- This is literally billions of dollars, huge environmental impact, potentially even people's lives at stake, et cetera. So it's a very, very important aspect of the work. So making sure those messages are correct is a very big deal for a company like Genie because again, you're talking about liability and so on, so not having enough content people to look at every single button, every single label was kind of a challenge, but it is something that the team has absorbed. And then between the PMs, the devs and the UX people, we were able to figure out what the buttons were and if we really had a question, we would then go to the content writers. Now, fast forward another five years, what we're finding is the content people, there's just not enough content people to even look at that kind of stuff. They don't look at it.
- All they're engaged in is creating guides. They create guides for, here's a voice, here's the kind of error message that we typically give. You should use the third person. We don't make jokes, we call this thing X and it's a capital X versus this call this Y, and that's a lowercase y, and that's how we refer to things, and that's driving that sort of consistency. Now, fast forward another year, so what you're seeing is a trend of literally fewer and fewer content managers need it. Now, fast forward one year, guess what? All the LLM stuff, all the copilots, all the assistance, that's all content, everything, you train them on all the suggestions, all the utterances that copilot does, the copilot voice, the copilot instructions, every single thing that about the copilot is content. And so content is suddenly thrust into the middle. It's the alpha, the omega of the entire experience.
- It's the content. If you think about there's almost nothing there, right? You look at Microsoft copilot, which is quite successful, Amazon Q applications, there's almost nothing from the UI perspective, there's not much there and what is there? It changes quite a bit. The whole thing centres around is content. So all of a sudden you have content. Again, it's the pillar against everything around experience evolves, and so you suddenly need a lot more content people and you need them to do more than right guides. So if you think about that sort of arc of evolution, that's exactly what is happening to the UX. So you're going to need fewer and fewer people drawing tables. I'm sorry, I'm picking on tables, but hey, if I never see another table that I have to make, that will be too bloody soon. But I mean we have to uplevel our conversations to talk about the money and the outcomes and the ethics and what is it that we're doing with this long-term with technology and what is this really delivering? Is this doing the right thing for our customers, for the environment, for the planet, for our company, for shareholders, et cetera, and that is where we need to be and there's not enough people doing that. So I hope that helps answer the question.
- Brendan Jarvis:
- Yes, I think Bill Gates said back in 1996, that content is king and listening to you talk there, it seems like perhaps that is finally being realised and truly understood, at least in the way that you see things panning out there regarding content. I want to touch on the importance of training these models for the purposes of design. We've been quite hard here on visual designers. The people that are focused on Figma are focused on creating layouts, rightly or wrongly. That's their current focus. That's what they're doing. Think about the models that their data that they're trained on, at least to my understanding, so correct me if I've got this wrong, is critical in ensuring that they create a positive design outcome. And what I've been wondering is technically for a designer, whether they're visually based or otherwise, their ability to create better design depends on the training in inverted commas that they have undertaken throughout their career and how I suppose reflexive, they've been paying attention to what works and what doesn't work. And so I'm curious about how feasible you feel it is for an AI that's being trained to effectively distinguish between something that is objectively good design, which perhaps that needs some definition and something that isn't. How will we ensure that the models that are being used to create UI in the future are objectively creating good design?
- Greg Nudelman:
- That's a great question. If I could correct you. I guess what I'm saying is I think the visual desires are going to be just fine. Don't get me wrong. I think where a lot of people that are going to have a problem coming in is they're using a design system for example, and then they're taking what is already in the design system and then they're creating a local example of that for a particular purpose to fit a requirement that's been pre-written. For example, a PM says we need an address form. Here's a crazy situation. I've actually had this happen at a very large company that I'm not going to name to protect the guilty, but we had to make a login form a kid, you knob three months of effort to create a login form. It is absurd the amount of money that was spent on that.
- Do we call this login or sign on? Is it login or login? Is it, should we have a finally button on the cancel? Where does the cancel go? If this is a model? Oh, it's a model. Okay, well how responsive is it? Let's talk about all that kind of crap. So it's just on and on. The conversations with literally the ROI is zero. Like yes, you need an accessible login form, but there is like a billion login forms out on the web. This is where AI can come in, just knock one out, you're done, don't need to talk about it, we're done. Let's see if it works and let's not worry about it. I mean you and I can look at it and go, looks pretty good. Or better yet, what AI can do for almost for free is give you 10 versions of it.
- It can go like, here's 10 versions literally for free in high definition using the design system components that have already been created and it's been trained upon so it knows the primary button and it knows the secondary button and it knows every time there's a cancel, it's a capital C and it knows it's supposed to cancel that back and so on. It has all these rules already built into it based on thousands of pages that it has been trained upon and maybe doesn't know all of the best practises, but it knows enough from just regular web content ju exposed onto your design system so that it can create one. And not only that, it's just going to create one straight and react. So there's no hanky panky, there's none of that. Is it eight pixels? Then you've got some renegade UX design that goes, well, I hate eight pixel grids, I hate mods.
- Well, I'm going to do this. I don't like the blue that they're using. I think they're too close together. They just going to go and create all kinds of nonsense for you, and this is where 90% of the time, 90% of the full designers are spending their time today and it is absolute waste of time. We take the stuff down the roads where we just don't need to do, and yes, we need to watch what we trained us on, but majority, I think our input into the process will be a really robust design system and by meaning it is robust. AI itself is probably going to tell us, Hey, you're missing a checkbox in this particular thing. It looks like you have an indeterminate state of the checkbox that is missing. Well, that's actually very useful information. Can you create one for me? Great. Here's an indeterminate state of the checkbox.
- It probably has a minus on it. It probably has the highlight colour that everybody else is using throughout the system. It's probably smart enough to put that together and goes, do you like this? Or, I have six other versions for you. If you want this minus sign to be like two pixels from the edgings or one half pixels, whatever it is, you can then make a judgement call and say, yeah, that's the right component. Go ahead and make that and react and go ahead and use that. The next time somebody says, I have a select all type of control that I need, or I have a table with a select, well, guess what? This probably needs select all. I cannot tell you how many times in my career bull as an ic, the visual contributor, and as a team leader, I've had designers come to me and just present these out of left field designs around search and tables and things like that around.
- I do a lot of SaaS work, so tables is kind of the bread and butter of a lot of these displays. Even if it's a data visualisation at the bottom of it is a table, so people would forget that it has to be sorted a certain way that a sorted mechanism needs to be present, that certain things need to be sorted together, that it's not an alpha sword if it's a time sword for example, things like that. Just no thinking involved at all, but this is exactly the kind of stuff that I think AI is going to do so much better at because it's going to just uplevel it and not only that is going to give you the exact table already made and react and it's going to give you six or seven or 10 versions of it. And not only that, while you're talking even further while you're talking to a customer and the customer says, yeah, this is really great, but what I really need is a column that has the total in it or I need to know when that total was computed in addition to this, well, AI hearing this can actually go in and insert another column and it's going to be a number and it's going to be right justified because all numbers are right justified around a decimal point, and it's just going to put that in the right order and then it's going to prompt you as a moderator say, why don't you ask them how is it supposed to be sorted?
- Do they want it to sort by this life or fife, luton first out or a different kind of sort? And so that type of library, then as you kind go through and create it starting with your current design system is going to be the fuel that is going to make this AI smarter and smarter as the time goes on to where it's going to get more and more things right straight out of the box. So all you need to do is just tell it what form you need with basic information around it or what kind of table, what kind of basic fields. I want an id, I want timestamp, I want, who made the last change? I want a description. It's just going to build one out for you and it's not going to forget that it's supposed to be sorted by time or it's going to prompt you, Hey, do you want to sort it by time? Because a lot of times that's been happening. So I think is kind of the essence is that's the part that's going to train it and you interacting with it, it's going to create this idea of augmented intelligence. So rather than it doing the work for you, it's going to just empower you to go that much faster and delete all of the robot monkey work that it's already really good at. And humans are actually pretty bad at to be honest, and designers are terrible at it
- Brendan Jarvis:
- Considering what sounds like some significant trauma you've experienced when it comes to the design of tables and other UI elements over the years,
- Greg Nudelman:
- It
- Brendan Jarvis:
- I can see why you would be very happy with the utopian AI assistance to create these design systems and the patterns within them a hundred percent. And I can buy that technical argument, right, like the attention to detail that it can pay to the finer points that humans just can't pay. I can buy the argument that it will help us to create overall more effective designs depending on how you judge the outcome of design. I can understand all of that. What I wonder is given the scenario you gave of that company that took three months to design the login screen, and I think I've heard you talk about this before where literally there was over a hundred mind changes as to where and how different elements on that simple page should be and should look and should work. Do you feel that introducing AI into the mix here in terms of its support that it can give UI in establishing best practises will get any more respect than designers have been in similar situations like you've described?
- Greg Nudelman:
- Wow, that's a very different question than a terrible table trauma. I think we've actually gotten worse. I think when we started out we've had to fight for getting seat at the table proverbial seat, but I think too many of us have traded to paraphrase Pink Floyd, a walk on Pardon in the war for a lead role in a cage, and we've kind of traded this uncomfortable questions in this role of disruptor and kind of a pirate renegade thing. If you've ever seen Reddit two A, that's a really good discussion there about who the chefs are and she says, Hey, we're all just artists and pirates and we're disruptors in our own way, even though they're all cooking based on these established recipes, they're all quite individual in their approach and they all have their own quirks that are reflecting themselves in their work. And I think that's something we need to go back and treasure and treasure, and that is how we gain respect by creating value, and it starts with respecting our own internal customers.
- Two, often I see designers staring down this lofty ire tower and saying, well, I am the only or arbitrary of what is beautiful or what is functional. Let me tell you this and I can tell you a hundred different examples. I just want to remind, I'll write this down. Alright, so now that I wrote that, I won't forget, but what I'm saying is we need to be of service and we need to bring value and we need to come down from the lofty hates heights that we've put ourselves on and be part of the team. Really too often, if you look at agile methodology, we've been the chickens, so I dunno if you know the story, but maybe if some of your listeners they'll know it. But the story goes that you have the chicken and the pig starting a restaurant and the chicken says, oh, whatever.
- And the pig goes, what do you mean it's my bacon for you? It's just the eggs. So the moral of the story is you need to be the pig and not the chicken. And too often we're the chickens and we just create these eggs and we say, oh, that's a beautiful egg, and we just kind of throw it over the wall and say, yeah, good luck, go omelette. And that is no longer where we can afford to be with the coming of ai. A lot of these functions that we've been tasked with doing these purifying things or working against a set of requirements that have been provided from on high, it's not going to cut it. It's just not going to work. So more and more you've seen what has happened to the content designers. In the old days, the content went into the rest of the team.
- It didn't disappear, but it went into UX and product management and developers. So the same thing is happening to the design. By using these tools that can very quickly developer can say, well, give me a nicely laid out address form. Well, is it going to be the most beautiful address form anyone can make? Probably not, but it's going to be pretty damn good and it's going to be accessible and it's going to use the components I already have in stock and it's going to take me no time at all to make, and instead of somebody spending months and months creating the slogan form, I can just knock it out in an hour and we're done. We're just moving on and if we ever want to change it, we can go back and change it. This is a proverbial two-way door that Amazon calls it. There's one-way door and two-way door.
- This is a two-way door. You can always change the login page. So one designer comes to mind that I worked with who, I'm not going to name any names of course, but he said, I bring beauty to the world. And I said, oh, really? Okay, interesting. So he spent six months with his team of six people working in this design that was supposed to be used in the desert in bright sunlight and on the little tough book pro little Panasonic, but yay Big bright. And these people had gloves on because their oil workers and the design that he came up with is tiny little letters, old black, very low contrast stuff. And of course a lot of the oil workers are middle age and above, they're males and for one reason or another, some of them are white. And so you end up with a very high incidence of colorblindness and unfortunately what he chose as the marker of trouble versus not was this very subtle colour change between a bright fluorescent green and kind of a subdued red and there's no change in shape or anything like that.
- So that according to him was beautiful. I would say it's an absolute garbage of a design and has proven to be that that's the case. So you've got this absolute incredible arrogance for somebody to come in and say, I find this beautiful. So you should implement that. It is just beyond arrogance to not ask any of the questions that you need to be asking as a designer and not consider for a second where this thing is going to be used by whom, in what context, what device, what sort of PPE or whatever else they're going to be using, and just assume that it's beautiful and therefore it's good. And in these little half pixel bright fluorescent lines on black background, that's going to be in direct sunlight. I mean it's just absurd, that kind of stuff. But that happens over and over and over again.
- So for whatever case we've created for ourselves a reputation being assholes, I mean, I'm just going to come out and say it. We've created, we're not team players. We don't take criticism well. We're pri Madonna and we're super touchy and we dislike other people doing any kind of research or talking to customers. God forbid a product manager or developer, oh my God, developer talking to customers like, wow, that's never should happen. God forbid, that's just baloney. That's ridiculous. Everyone should be talking to customers and yes, we are better trained to ask better questions and we probably can interpret it a little bit, but that just means we got to bring other people along or we got to join them on their discovery journey. It's going to take a village for everyone to come together and do this. And only that AI is sort of driving these larger conglomerations conglomerations of larger sets of humans working together on stuff that not one person can come up with that whole myth of a single hero that's going to pull through and come up with everything That's always been baloney, but now it's just absurd. It's to the point of being silly, which is where utterances by world leaders that only I can solve this problem. Only I can be your saviour. It's absurd. In the same way, designers cannot say, only I know what is beautiful. Well, no, it's just not true.
- Patently false. So let's just start with the posture of humility. Let's start with what can I do for the project? How can I contribute? Here's some of the things I can do for the team, which of these would be most helpful? Let's start by doing what we do really well. Listen to the discussion and take notes. So that's a very unpopular thing. AI is taking notes. If AI is taking notes like the way it does on Zoom, maybe we can sketch design ideas as we go live on the zoom meeting and say, is that what we're talking about? Really what I heard is this and this, and this goes to this, to this, to this. Is that what we really talking about? Then everyone can be on the same page and you're providing clarity and you're adding value every step of the way, whatever methodology you're using, if it's inflexible and to the point where you cannot do that, you're going to have to change that methodology because you've got to be bringing value every single second that you are on this project. You got to be earning your keep. That's how you earn respect. So rather than asking how can we get them to give us respect, we got to start by respecting them, our partners first off on our team and not considering ourselves to be holier than thou, and really just looking to contribute and looking to add that. That's my 2 cents of course, but that's where I come from.
- Brendan Jarvis:
- Perhaps I'm being like a dog with a bone here about that example you gave of the login page, but I want to come back to it one more time and see where else you might lead the conversation off this. So thinking about that failure that you experienced, not your own failure personally, but thinking about the failure that was chewing up three months of time, designing a simple screen, how much of that do you attribute to the stuff that is outside of the thing that people are looking at? How much of that, when you think about that situation, was less about the design that's in front of people and more about the relationships that exist between them?
- Greg Nudelman:
- You hit the nail on the head once more, Brendan. It's exactly that. There just wasn't enough trust. There wasn't enough relationship building up to this point, and there wasn't, the decision maker wasn't identified, and most importantly, we did not know how to judge the definition of success. So you told me a lot. I mean, this was one of the most valuable, it's a complete waste of money for the company, but for me it was very, there's valuable because I've learned on this and many other failures of a similar kind because that literally what makes you an expert is you've screwed it up in every possible way that you could. You've screwed the pooch and now you know what not to do. All the ways it doesn't work. I think that's literally what makes you an expert
- Brendan Jarvis:
- When it comes to AI projects of which you have become somewhat of an expert in, right? You've done over 30 of them, you've been involved in these for the last decade at least. What is the single biggest sin, if you like, that design leaders are committing when it comes to increasing the chances that they'll screw up those projects?
- Greg Nudelman:
- Oh wow. Well, that's my favourite topic ever of conversation. Well, there's actually eight or nine that I've identified, but I think the biggest one is where we just torpedo it straight out of the gate is we go after we got a hammer, let's go find some nails. Rather than asking what is the customer really trying to accomplish? One of my favourite examples is when I worked for a smart irrigation startup, the founders had this idea that they're going to sell irrigation insurance. So they're going to go to the farmers and they're going to say, well, we are going to help you to make sure your crops is efficiently watered. And so these farmers were, a lot of them were second, third, 10th generation in one case, which is like boggles the mind, right? Tim generation farmer who's been working this land forever since they were a little little boy.
- And so you try to tell this person that their crops is sufficiently watered, they would just laugh us out of town. They have a very simple thing. They would go out and do this kick test. They would literally sink the heel of their boot into the ground and if the heel goes in crops is officially wooded, if the heel not go in, then water a little more. So that part was not clean, it was not the issue. So I immediately pivoted to say, well, what us skipping you up at night? Well, it turned out there was plenty. There's new water regulations, there's water scarcity due to climate change. There is all sorts of new acfi pressures because multiple people are drawing from the same acfi, and so there's a limit of how long and when you're supposed to pump, and then there's the cost of pumping electricity has to do with it.
- I mean, there's on and on. They would just rattle it off to me over and over again. So I said, well, wait a minute. What I recommended as a consultant is we immediately pivot the whole company to asking, how can I save the water? How can I help you water less while maintaining your yield? Really what they care about was the bottom line. If I save you time and money, if I'm able to pump during the night when the electricity cost is low, if I'm able to water it very little or as little as possible using timed pulsed irrigation, which is the latest, greatest thing, and I can ensure that it's sufficiently watered down at the roots while I'm using as little water as possible. Now you're talking, right? Yes, I'm interested. Sign me up. Oh, I can save like 20% of my water. I can save all this cost.
- That goes straight to my bottom line. These people understand that the more money they output, the less they get to keep. They know it very well. It's a zero sum game for them because there's only so much yield that they can get out of the land, and so that's a much better question to ask. So as UXers, we're often too often we drop onto, again, it's the same problem the PM told us that's the use case, so we're going to go with this use case. Well, you need to retrain ourselves, I should say, to ask those uncomfortable questions. Is this the right use case? Is this really what keeping customers up at night and how many customers did you talk to? Did you just show them something and said, do you like that? Is that all you did? Is that the kind of research you did?
- And then, okay, great. Let me go ahead and honour you as my partner and go ahead and make that mock-up, but spend as little time as possible on that, and then literally tell the story you're trying to say, and then let's go to the customers immediately in as little as a day or two preferably. So don't spend three months, pardon my French, dicking around with stupid colours and doing fluorescent green and fluorescent red on black background, if that's what they're doing with a team of six people who cry out loud. That is just absurd, right? Don't do that. Spend a day doing that. Make a mockup that fits the basic story and then go immediately to your customers and say, does this make sense? Is this literally what you want? If we gave you this, how much would you pay us? That's literally the question you should be asking, is this bringing value to you?
- Is what you want? And if you were a magic wand king or queen for the day, what would you want it to be? Let's say you have unlimited budget. We have this very smart ai. What do you want AI to do? Most people will tell you that this is not it or it is it. Or if we tell you something cautiously, I'll say, well, it doesn't really work for me, but Dan down the street, oh yeah, he's really going to buy that. That's not going to work for you. We should all have enough training in our ear or to know that's actually means a no. That's a very polite way of saying that's not going to work for me. So that's where we fail most of the time. I think as designers, we're not asking uncomfortable questions with ai too often it is in the hands of developers and is the hands of data scientists and AI is just too important to the future of your company, to the future of your job and your career and to the future of humanity as a whole. For you not to take an active role in asking what is really the right use case to solve. That is absolutely the number one way that these things blow up. And the good news, you already have all the tools. As a UXer, you already have all the tools to solve that. So all you need to do is just be brave. Like the brave podcast. You need to be brave, a brave UXer, and ask the right question, ask these uncomfortable questions and then act on that.
- Brendan Jarvis:
- There's a lot in there, and I feel like you have started to touch on this idea that's popular in our community, which is that we are well positioned as UXs to provide somewhat of a defacto ethicist role when it comes to the evolution of product. And I've been thinking about this over the years on this podcast and I'm not sure I buy that argument. And the reason for that is for a start, most of their designers that I've come into contact with don't seem to be the ones that are calling the shots as to what gets built. They might have some influence over how it gets built, but they're often not the ones that are in charge of making those big strategic decisions. And the other thing that I come unstuck on with this argument that we should be providing that lens is that it makes us, it positions us as somewhat of a roadblock to people wanting to get certain things done. And maybe that's necessary. Maybe there do need to be roadblocks and checks and balances, but I just wonder how well that positioning is working for design as a broader field if we are perceived to be roadblocks rather than enablers. So I'm curious, that's a bit of a download from me, and I'm not suggesting that you believe anything really, but I am curious to understand, based on what I've said, how do you see that working either for or against us playing into that role that we seem to have cast ourselves in?
- Greg Nudelman:
- I think we were not naturally fall into ethicist role. I think you're absolutely right that it doesn't come naturally to us and it is a mistake to paint us as a, it's what happened to accessibility for better or worse, I think it's the same story. And I think it's what in the larger microcosm I think happens to, it's happening to the left in the us for better or worse, they're seen as this kind of the thought police and the police of a particular language use and so on. And the few equates a really large digress, a really large and egregious form of racism with a just innocent misuse of a particular gender profile. Oh my god, he called them a she. And that's the end of the universe. And I actually do know what I'm talking about because of my daughter, I'm not going to get into it, but the point is maybe just being a little bit more tolerant about that would've actually helped accessibility rather than hurt.
- And so we tend to come down, just bring the hammer, tht, hammer of God on anyone who dares to suggest we ship it and then fix it later. Or that God forbid, it's AA compliance to the AAA compliant and it only works on a certain platform. It only works in mobile or desktop or whatever. And I think we have to look at a larger benefit. And so I don't think it's served us well in the past to do that. And I don't think, at least I'm hoping, I'm not suggesting that although people tend to misinterpret and they'll take things out of context all the time and they'll say, oh, Greg is kind of six, usability is bad. That's not at all what I'm saying. I want to make it very clear. I know Jacob Nielsen got into the same situation with something he wrote and I wrote a lot in support of what he was trying to say, which I think he's basically saying that with ai, we actually have a shot at it because we can't actually customise it in a much more direct fashion for a specific person.
- And it doesn't need to just be all in one, all accessible or not at all accessible. We tend to have this absolutism about us. And I think with ai, it's very much the same. And I think it's giving a lot of folks a lot of anxiety in our industry because AI is moving so quickly and it's very difficult to tell. And then what is happening today is fine. And then a week from now we've changed the algorithm slightly and suddenly it's completely not fine. And so it's kind of an ongoing anxiety drama situation, which a lot of folks in our industry already prone to a lot of drama and anxiety, myself included. And so it doesn't play well for us to worry over much. So I would really caution against being the roadblock or being even seen or perceived as a roadblock. I think you're absolutely right.
- It's very delicate thing. Where I think we have to call it is where there's literally potential harm that can happen. And that's when you pull out all the stops and you cry wolf when the wolfs actually come and you don't be the boy who cried wolf because you've seen a kitten or you've seen an ant in the grass, but literally wait for that exact moment when you need it. For example, January 6th, perfect example. The whole Facebook stuff and the algorithm that they shipped that did not surface any of the actually cooler articles, maybe even from other Jones or stuff that was a little more balanced, they completely went into grievance mode and they said, oh, grievance is generating clicks. Fantastic. Let's just double down on all the grievance stuff, whether it be left or right, and let's just give them they want more grievance, they're clicking on the grievance stuff, let's give 'em some to be aggrieved about.
- Well, there's always something to be aggrieved about. And so that is exactly where we needed to be stepping in, calling it. And I think because we have not been doing that or we have been sort of crying wolf a little bit throughout and not like that's the one time when we needed to do it, and nobody has done it. Nobody in the Facebook, nobody in Twitter, nobody stepped forward and said, this is unethical. This is a problem. This is going to cause our democracy to fall apart. Potentially, yes, you're making a few extra bucks, but this is absolutely something I'm not going to stand behind. There's a couple of individuals that have left the company since then that have called it out, apparently that came to light. So it's not like nobody has done it, but it is I think, critical that we don't react to every little thing and we actually do have a lot of humility.
- But when the situation does call for us to act, that we act and we step into our roles and actually do it. But to be in that role all the time is a terrible, I completely agree with you, man. It's a recipe for disaster for us and being dismissed outright. But I think where we need to be is we need to be putting ourselves into this role of an expert consultant over visier perhaps, or gal to Aragon. And then everything the Gandalf said that people listened to, no, but by and large, and one comes to mind, they should stand and fight. Instead, they're running for the past and I fear a much bigger disaster. So we need to have that larger view of a global village that I think Gandalf showed so well in the return of the king, but at the same time acting locally something that is helping the leaders to create that vision.
- And I think that's where you become a trusted advisor. You become somebody who is respected in the leadership team as a voice, like you're just adding your voice. You have that seat at the table after all that we fought so hard for, let's use it now. What does it take to do that? It takes engaging, it takes, maybe you have to step up and do it from the ground up, but eventually, hopefully where you get to is this role where you help provide a vision for where this AI is going to go. And then maybe even add your voice to the questions that come from this vision and be maybe somebody who tests the vision with the potential customers and discusses the viability of the vision and what is the long-term concerns and issues. And that is how you get to be counted. Now, don't expect everything as even Gandalf never got everything that he was worried about resolved. But by and large, if we work hard, if we have the best purpose of humanity in our hearts and the best hope for the enterprise, I think we find that we're going to be in a better place than we would've been had. We just complained about every single little piddly thing.
- Brendan Jarvis:
- Greg, this has been a really insightful conversation about AI and its impact on design. Thank you for so generously sharing your stories and insights with me today, and also for your voracious contributions to the field over the last 30 years or so.
- Greg Nudelman:
- Thank you so much for having me. It's been a pleasure, Don provoking questions. Yeah, really, really good time. Thank you.
- Brendan Jarvis:
- Most welcome, Greg. If people want to connect with you, want to follow along your contributions to the field, your thinking on AI and UX, what's the best way for them to do that?
- Greg Nudelman:
- UXforAI.com is where the blog is and all the greatest happenings and the conference appearances, the workshops, et cetera. So that's really the best way, UXforAI.com. And I do have a book six coming up with a lot of these questions. Really rather than trying to be the oracle of truth and the only source of veracity, what I'm trying to do with the book is ask a lot of the questions and then have the readers have the practical tools to explore what it really means for them and how to drive value from their own unique perspective that they bring to the field. So look out for that book, UX four a.com, check it out,
- Brendan Jarvis:
- Check it out people. Thanks Greg, and to everyone that's tuned in, it's been great having you here with us. Everything we've covered will be in the show notes, including where you can find Greg in all of the things we've spoken about.
- If you've enjoyed the show and you want to hear more great conversations like this with world-class leaders in UX research, product management and design, don't forget to leave a review. Those are really helpful. Subscribe so it turns up every two weeks. And also just tell one other person that might get value from these conversations at depth.
- If you want to reach me, you can find me on LinkedIn, just search for Brendan Jarvis, or you can find a link to my profile at the bottom of the show notes or head on over to my website, which is thespaceinbetween.co.nz. That's thespaceinbetween.co.nz. And until next time, keep being brave.