Future of UX · 2026-08-13 · 19 min
Key moments - from our scoring
Substance score
41 / 100
Five dimensions, 20 points each
Patricia, host of Future of UX and founder of the AI for Designers Bootcamp, offers a rare behind-the-scenes look at her business operating on a three-day-a-week schedule (roughly 24 hours). This episode breaks down her summer projects, including custom private company cohorts for design teams and a personal side project called Melty (a branded popsicle recipe app built with Adobe Firefly) that informed three major observations about the evolving AI and design landscape. The core insight: Cloud Code has transcended its developer tool positioning to become a research synthesis interface; documentation has shifted from nice-to-have to essential infrastructure for AI output quality; and the industry is moving from copilot-style suggestions to multi-step autonomous agents that require entirely different design thinking around reversibility and risk. She also identifies a critical gap in most design teams - not tool knowledge but alignment and shared agreement on how to use AI. Patricia's honest failure: delegating too late despite three days allowing ample capacity. The next cohort of AI for Designers (launching end of September) will feature multiple guest experts rather than Patricia teaching everything solo, reflecting her realization that no single person can stay current across all specialties in this rapidly moving field.
Cloud Code has become a research interface for synthesis across dozens of transcripts and building internal tools that wouldn't be prioritized for engineers - not shipping code but organizing and processing research at scale.
AI output quality depends almost entirely on what it can read about your context, design principles, decisions, and constraints; teams with good documentation get dramatically better results from the same tools and prompts.
Copilots suggest and the human decides; agents act across multiple steps sometimes without asking first. This shifts the design problem from 'is the output good' to 'what happens when it's wrong and who carries the risk.'
Not tool knowledge but lack of shared language and agreement - no shared decision on what good looks like, who can use AI for what, and different team members operating at completely different quality standards with the same tool.
No; the constraint actually made the business sharper by forcing decisions about what truly matters, and most activities that got cut were things nobody missed anyway.
Our reviewer’s read on each dimension, with quotes from the episode.
The episode offers moderate insight density with some genuinely useful observations (documentation as infrastructure, context as skill, copilot-to-agent shift) mixed with substantial filler and tangential storytelling. The discussion of private cohorts, the popsicle app project, and personal anecdotes about work-life balance consume significant runtime without advancing substantive business or design knowledge. While the three main learnings are valuable, they represent relatively well-trodden territory in design and AI discourse.
context is now the skill
the moment the AI acts instead of suggest, the question stops being is the output good and becomes what happens when it's wrong
The episode presents mostly incremental observations rather than counterintuitive or first-principles thinking. The three learnings - documentation importance, tool skill gaps, and AI agent design challenges - are extensions of existing design discourse rather than novel frameworks. The speaker acknowledges lack of expertise and pivots to curating others' knowledge, which is pragmatic but not particularly original. Few claims here would surprise experienced B2B operators familiar with AI implementation challenges.
Cloud code stop being a developer tool
the terminal became a research interface
This is a solo episode with no guest present. The speaker is Patricia, a UX designer and AI consultant running a bootcamp, which establishes baseline credibility but lacks the caliber of bringing in practitioners at significant scale or proven results. The promise of future guest experts for the next cohort is mentioned but not delivered in this episode. The speaker explicitly acknowledges not being expert in everything, reducing her positioning as a definitive voice.
I am not an expert in everything and that's the point
I am a juror at the AIX Awards 2026
Evidence is sparse and largely anecdotal rather than data-driven. The speaker references 'three different contexts' for Cloud Code observations but provides no named companies, metrics, or quantified results. The 24-hour workweek claim lacks supporting data on business outcomes. The popsicle app project (Melty) is personal and illustrative but not substantive business evidence. References to other episodes and workshops exist but without specific outcomes or metrics.
I've now seen it in three different contexts
Six months, three days a week
This is a solo monologue format rather than a dialogue, eliminating opportunities for sharp questioning, follow-ups, or productive disagreement. The delivery is conversational and warm but lacks the intellectual rigor of challenging one's own claims. The speaker presents learnings as conclusions rather than exploring tensions or counterarguments. There are self-promotional pitches (survey, waitlist, Adobe workshops) interwoven with content that blunt substantive engagement.
if you are self employed and you would be interested in working together
I would really love your help with because I am looking to work with more freelancers
Computed from the transcript - who did the talking, and the words that came up most.
In this first quarterly update I talk about the private company cohorts I'm building, the free Adobe Firefly workshop series, and three shifts I saw in the field over the last three months. Plus the honest version of how I actually work: three days a week, what that constraint did to my business, and the one thing I got wrong and waited far too long to fix. IN THIS EPISODE Why Claude Code stopped being a developer tool and turned into a research interface How documentation moved from hygiene to infrastructure, and why designers are well positioned for that shift What changes in your design work the moment an AI feature acts instead of suggests Why being an expert in everything is no longer the goal, and what replaces it How working three days a week made my business sharper instead of smaller Why design teams rarely struggle with tool knowledge and almost always struggle with agreement The thing I waited six months too long to fix RESOURCES Work with me: if you're self-employed and open to collaborating at some point, fill out this short survey. Completely non-binding, takes two minutes.
Transcribed and scored by The B2B Podcast Index.
Speaker A: I work 24 hours a week. That's it. That's my full working week. And I'm telling you this at the start of this episode because for the last six months I've been quietly building a business that number. And it has changed almost everything. How I work, what I say. Yes. What I hand off and what I stop doing. And to be completely honest, it made the business even better and not worse. But it also created a problem that I didn't see coming. And I want to be really transparent with you. Hello and welcome back to the future of ux. This is the podcast where we look at how AI is reshaping design products and the way we actually work. Not just the tools, but the structures underneath. If you're new here, my name is Patricia, uh, UX designer, AI consultant and I work with product teams on designing AI powered experience. I run the AI for Designers Bootcamp and, and I spend a lot of time tools, making things and figuring out what actually holds up in real projects. And today's episode is a little bit different. I am starting a new format. Every quarter I want to do an update like this one. I have done a couple in the past. A proper behind the scenes and here's why. What I share with you is completely finished. Like a framework, a workflow, structured way of thinking about something. And that's useful, I think, but it also hides everything that happened before the framework existed. Like the messy part, basically. So the projects that taught me the thing, the stuff, um, yeah, the stuff I got wrong, maybe. So this is that episode. I will walk you through what I've been building this quarter. What actually changed in the field. Three learnings that shifted how I work and yeah, also one very honest fail. So let's go. Part number one, what I've been building. So the biggest thing I've been working on this quarter is something I haven't really talked about publicly. I've been building private company cohorts. So basically the AI for Designers program, but fully customized for one company and their design team. It runs through four to five weeks as live elements, pre recorded, uh, video assignments, everything. And I have to say that I find this extremely cool. It's honestly, yeah, a super fun that I'm working on right now because it's a completely different problem than the open cohort that I'm working on. When you teach an open cohort like the AI for Designers bootcamp, you're teaching individual who each go back to their own context. And when you work with one design team inside one company or not just teaching the tools. You're really changing how the team works together. So there's a lot of alignment. You look at their actual design system, their actual research process, the actual handoff, and then you build the program. Many of those private company cohorts start in autumn and I'm deep in the planning right now, so I'm very, very excited. By the way, if that's interesting for your team, you know where to find me. Going to pitch it here. I just wanted to share. Yeah. What I'm really, really excited about at the moment. Another thing from this quarter was the workshop series with Adobe Firefly. This was super, super fun. As many of you know, I'm working on a lot of personal projects on building my own tools, my own products just for fun, but also to get a little bit more productive. And one project that I was working on this summer was, um, actually inspired by my daughter, one and a half. And she really loves ice, even like ice cubes, like melted water. But she also loves popsicles. And all our neighbor kids are a bit older, so they're already eating like a lot of these, like, uh, you know, sugary popsicles. And I thought, okay, I need to experiment a bit and try maybe a healthier version because she's still so little. So I experimented a bit. I found some really good recipes online and now I'm making her basically her own version. Of course, can also get a healthy one if they want. And I wanted to structure it so I can find it that I can also share it with my partner and with friends. So I created my own mini app just for fun, where I gather all the healthy popsicle recipes. Make it fun. And also to challenge myself a little bit, I created a really nice branding around it. The whole brand for this concept is called Melty. I created really beautiful visuals and mock ups for this brand. So this was just like a really cool summer project. I shared the whole process of how I created the visuals and the mock ups and the mood boards and the videos, everything in three workshops. And I collaborated with Adobe. I was doing everything in Adobe Firefly boards and I basically shared the whole process. It's completely free and it was super fun. You can still sign up and um, have a look at all the recordings. So this was like a really cool thing this summer. Yeah, really excited. This really big project and really big collaboration together with Adobe. Super, super excited. So usually summer is a little bit slower for me. I usually do a little bit of like time off. I, you know, spend time with, with, with the family. Um, outside, usually we go camping. We do all these things in Switzerland. And I did the same actually for this summer. So I was working on a couple of projects. So a lot of time for me to reflect what I want to do and down a little bit. So I did a couple of projects. Um, also focus a little bit on relaxing. All the relaxation and the, uh, time to experiment and to try things out also led me to a few realizations that I want to share with you, because there are a couple of things that really stood out to me over the last couple of months. The first really big shift for me and how I work. Cloud code stop being a developer tool. And this is the one I'm most confident about because I've now seen it in three different contexts. Cloud code has maybe a little bit of a branding problem because the name says code. So designers and researchers look at it and then think, not for me, but some of the most interesting use cases I have seen this quarter. Nothing to do with shipping code. It's research synthesis, it's working across dozens of transcripts, it's building internal tools that no one would ever have prioritized for, you know, an engineer. I actually had a full podcast episode about that with Caitlin Sullivan a while back. I think it was like, maybe like a month ago or something. If you haven't listened to this one, go back and do it. It also changed how a lot of people in my audience work. So this was a very inspiring episode. Also the very first video, uh, podcast episode. So you can see the video. You can also see Kaitlyn's screens, what she's sharing, how she structured everything, and I really love that. So the pattern I summarize is the terminal became a research interface. Nobody predicted that. This was like my first realization, stop being a developer tool. The shift number two for me is that context is now the skill. The second really big shift is about documentation. For a really long time, documentation was the thing everyone knew they should do and nobody wanted to do. So I would say it sit at the bottom of the priority list. And that flipped a little bit this quarter because the quality of what AI produces you depends almost entirely on what it can read about your context, your design principles, your decisions, your constraints, your past work. Teams with good documentations are getting dramatically better output than teams without it. Same tools, same prompts, completely different results. So documentation stop being nice to have and more something that you need, like an infrastructure. And I think designers who understand that early, ah, will have a real advantage because we are often the ones who already hold that context in our hands, what is documentation? It's basically a document where you write things out. So it's not that complicated. It's a document where you document what are your design rules. And then a really big shift. Shift number three is the whole jump from copilot to agent. This is happening in real products. And I also noticed that in mostly in my client work a year ago almost every AI ah, feature I looked at is maybe a chat, chatbox, like a chatbot. A copilot suggests something and then the human decides this quarter or even like a little bit before that changed, I'm seeing more features where the AI actually does something multi step across systems, sometimes without asking first. And that's completely, that's a completely different design problem because the moment the AI acts instead of suggest, the question stops being uh, is the output good? And becomes what happens when it's wrong? And who carries that risk. I talked about this at the beyond the Copilot episode. So I won't repeat the whole framework here, but the short version is reversibility. Confirmation gates and the legibility are no longer nice to have. They are the design work. I also did a, ah, proper workshop about that in June at the UX on the beach conference in Venice. So it was really, really nice. So now the part that actually changed how I want to get you a little bit behind the scenes and talk about some of the learnings that I had. Completely honest and I mean this shouldn't be a surprise, but I am not an expert in everything and that's the point. This is also a really big mindset shift for me this quarter for a really long time. Uh, you know, I have a teaching program, the AI Bootcamp. I should be the one who knows everything in it. And this is also what expert does, right? I don't think that anymore. The field moves so fast. There is genuinely no single person who is on top of every part of this. Anyone who claims otherwise either not paying attention or I think honest. So my role changed. I stopped seeing myself as the person who teaches everything about AI for designers and I started seeing myself as a person who also curate the people who can teach it best. I don't think that this is a weaker position, I think it's a stronger one. Because what you need right now is not one person's opinion. You need access to people who are genuinely deep in each specific topic. And that changed how I am building the next round for my program, AI for Designers, which I will be back to at the end. But I can spoil It a little bit. We have several guest experts for the next cohort. We have a lot of support from different experts, um, who are also supporting the cohort. I am collaborating with different companies. Yeah. What can I say? I'm very, very excited about this huge upgrade. Okay, what was my tool? My second constraints made my business better. Maybe you have already seen this in the podcast episode topic or, um, in the beginning, second learning. And this connects back to the 24 hours. Also, here's the very, very honest version of my setup after my baby break. So around a year ago, I started with one day a week, about four months. Can you believe it? I only had one day per week. And then I moved to three days a week, which was this January. Pretty much that's still where I am today. Three days, that's my week. Monday, Tuesday, Wednesday. Sometimes I'm moving things around. But this is my schedule. And I want to be really transparent because I think a lot of pressure around this topic in both directions. For me, work is, uh, generally my me time. I mean, I always love working. I still love it. Not doing three days because work is a burden. Trying to minimize. Absolutely not. I build a business that I love. I have work that I love. I have wonderful clients. I am so grateful that I am basically living my dream life. Dream job. I would say dream job. And um, if I look back five, six, maybe 10 years ago, I was dreaming about exactly this work situation and I'm still dreaming about this. I think it's, it's fun. I love it. So I am doing these three days because I want to stay flexible Sometimes my daughter is sick thing, you know, uh, is happening. So now here comes the learning. For me, I expected the constraints the business smaller. The truth is it didn't. It made it much sharper. Because when you have three days everything you have to actually decide. And it turns out that most of what I nobody missed, including me. The constraint didn't limit the business. It forced me to figure out what actually mattered in my learning. Number three is teams don't struggle with tools. They struggle with agreement. My third learning was and this comes straight out of the company code. When I go into a design team, I expect the blocker to be the tool knowledge it almost never is. The blocker is there is no shared language and no shared agreement. One designer uses AI for everything and doesn't tell anyone. Another one touch it because, you know she's not sure if she's allowed. Nobody has decided what good looks like. Nobody has decided what's off limit. So you get five people using the same tool in five completely different ways and five different quality standards and no way to any of it. So there is a huge gap between um, people who are already pretty advanced using AI, know what they're doing, they're experimenting, they're trying out, and people who are a little bit, I don't want to say it in a mean way but who are a little bit behind, they're just starting out and they're uh, like on the other edge of this corner. This is a little bit of an alignment and this is what I realized working with lots of different teams right now, yet there's a lot of alignment needed with different. So maybe you are in a similar situation where you feel you are a little bit lost in your design team. You don't know what to do, you don't need me for that, right? Like you can start your own alignment, really gather, exchange ideas, find a little bit of work of how to work, found workflows or something that you can repeatedly do. Build skills for your design, build documentation that you can use. Build this foundation yourself. This uh, is my biggest learning, not only for my work but also I think genuinely for the design industry right now. Okay, and now comes the honest part because my fail this quarter is not a project that went wrong. It's something much slower and more, much more annoying than that. I waited way too long to delegate. Six months, three days a week. And that entire time I really kept doing almost everything myself. Like the content, the program, the client work, the operation, the ad admin, all of it through me. And it was pretty capable because of AI to be completely honest. It's not that I was stressed all day. I automated most parts that I could automate and that worked really well in very specific tasks. Which brings me to the one thing that I would actually love your help with because I am looking to work with more freelancers and self employed people on different projects. My projects on um, content on program work on all kind of things. So if you are self employed and you would be interested in working together at some point or uh, maybe working for one of my clients, you know, I'm working with lots of different clients all around the world pretty much and I can take on all. So I really would love to recommend someone. There is a super short survey like a Google form linked in the show notes, completely non binding. It just takes a couple of minutes. Fill it out and if something interesting comes up that fits do, I will come back to you and send you an email. I would really, really Love to see out there. So now what's coming? A few things on the horizon for me. I am a juror. I am a juror at the AIX Awards 2026 in September at the Future Product Days in Copenhagen. And I'm really looking forward to that one because, um, yeah, of the submissions and because it's a really good excuse to room with people building this stuff and I'm very excited. I went to the Future Product Taste last year, I did a workshop there and it was really wonderful, uh, conference with lots of great speakers. So I can highly recommend going there actually. Um, and you're also going to meet me then. The private company cohorts are going to start in autumn, as I mentioned. And behind the scenes I'm planning a lot for AR4 designers, which goes into the next round at the end of September, which, which is the last thing that I really want to tell you about AI for designers. The next round starts end of September and I'm currently rebuilding a good part of it. So here's the actual news and it goes straight back to learning number one. I am bringing in more guest speakers and bringing in specialists for specific topics because I don't want to be in a position where I'm teaching something that's slightly out of date or where I'm covering a topic at 80% when someone else could cover it 100%. And right now, in this very specific moment, I want to be sure that what participants see is generally the newest and the best version of it. And I'm not an expert in everything, nobody is. So the people listen around are going to get some really good updates and some excellent guest speakers. I'm uh, still finalizing the lineup and I will share more soon. But if you want to hear about it first on the waitlist, it's non binding, so just sign up for the waitlist link is in the show notes. You will get updates about dates, about guest speakers, about what's going on and you will also get the early bird bonus in case you decide to. So it's definitely worth it. I would really like to know if this format of the quarterly, uh, quarterly updates is useful for you. If you like this kind of episode, please tell me. Send me a message on LinkedIn, leave a review. You know that always helps me more than you think, to be completely honest. If you're self employed, seriously, go fill out that survey. Thank you so much for listening and I will see you or hear you in one of the next episodes of the Future of ux.
Other episodes covering the same guests and topics, from across The B2B Podcast Index.
How to Design AI for Scale From Day One with Tom Greenlees, Intelligent Core [090]The Business of AI · on Adobe Firefly80 / 100
IBM's AI Marketing Journey from Experimentation to Intentional Execution with Craig MillsAPAC's B2B Growth Podcast · on Adobe Firefly79 / 100
214. Daniel Thulfaut, Head of Product at saas.group - Why AI Is Killing Traditional Product ManagementThe SaaSiest Podcast · on design systems74 / 100
The Big A.I. Opportunity Hotels Are MissingCheck-In with Bryan · on design systems74 / 100
Lucas Josefiak, Widgetbook - Role of Design Systems in Software DevelopmentBuild To Succeed · on design systems74 / 100
How a Solopreneur Built a Business on One Recurring InvoiceSolopreneur Sessions with Fexingo · on design systems72 / 100
#160 AI Customer Research: Why Designers Should Learn Claude Code with Caitlin Sullivan63 / 100
#162 The Case of the Vanished Judgment#161 700 Components in One Day: How Design Work Is Quietly Moving Into Code#159 Beyond the Copilot: The 4 AI Patterns Every Designer Should Know#158 Why Junior Design Jobs Are Vanishing (And What To Do About It)