It’s been a weird kind of existence over the last couple of years trying to reconcile 30 years of design and product experience. Product and design teams are under immense pressure to keep pace with engineering workflows that have seemingly turned coders overnight into decision-makers managing code output.
That creates all kinds of new challenges, especially from an agency perspective.
What I’m seeing though are amazing opportunities to support teams in new ways using experience, intuition, and a different way of embracing AI. We ripped the band-aid off this week and leaned into an AI-first way of building and learning. Not halfway. Fully leaning into tools that support thinking just as much as production.
We launched Glare this week not only as a way to share these ideas, but collaborate with you to make them better to make them better. More human. More useful. And ultimately something that helps create more clarity for product and design teams that want to embrace AI while keeping a user-focused approach.
There’s what I call the AI brain fog that comes from managing “two workflows” and marrying the outputs with traditional work (curious if others get to this point). I think to goal is just start letting go.
Had some great lunches and meetings this week talking through these challenges (thanks Greg). Living in this transitional world creates all kinds of tension, and I’m seeing teams divide on how they think this unfolds. We want to help teams work through this challenge.
I dove into it more in this LinkedIn post:
Ultimately, we’re leaning in. I think the key idea is that we need to make problems smaller and layer in feedback as we go.
Following up on this post, one of the biggest opportunities for teams is figuring out how to bring more decisions into the build process. That’s harder than it sounds because it requires strong collaboration across functions in real time.
I’m curious how you are thinking about this @Kike_Pena and @nikhil_mahen, as it’s not just a shift in methods, but collaboration.
Ok, this is great. Last week, I was invited to sit with a bunch of guys from different backgrounds to discuss how AI has reshaped things in the tech industry nowadays, and without question, circling back to my initial thesis of how adopting AI tools and a mindset change three fundamental concepts in any company: https://uxdesign.cc/think-twice-before-adopting-the-ai-label-25d11f1a3ff5
I can confirm that all the new work dynamics that AI encourages are, to me, the new way of building things for the future. But beyond this new dynamic, it raises questions about the consequences of having tools that accelerate product development and shipping.
The big question now is: what can you do now that you have sped up the product-creation process? This topic is also beautiful to start considering, because “crafting” is no longer a barrier but a step in the process. Today, news skills like effective communication, imagination, and systems thinking are the new ways of design.
Your graphic about the new product workflow is 100% accurate.
Somewhere we all know the why the problem exists but as the world of AI requires in some cases unwavering structure, we don’t follow through with the intangible signals, irregular workflows etc. The way AI actually solves the problems is when it solves the felt problem. Most people either have hacks to solve the felt problem or have drudgery (repetitive manual work with a gazillion exceptions) to solve it.
AI often doesn’t take those into account.
Here is an example - we built an internal connection tool for a client in the construction space who can have access to all his company’s past and present data.
The challenge? that data is not followed or structured as consistently as one thoguht. Which means if there is a cost sheet, there are a dozen exceptions to the way you read it. However one only comes across the exceptions, when we try to implement it and get incorrect information. These experiences create the illusion of inefficiency and increase the burden of expectation for the AI to know what the humans have developed as unwritten rules. When the rules are not obvious or are only work arounds - AI cannot catch up quickly.
The new realisation is to set client expectations to feel this from day one. We designed a set of 30 questions (for each property in this case) which the client can ask in any form and have the answer reverted perfectly. Its not ideal but in the short term thats how we solved it.
In the long term the way to solve it would be to build the context graph of the department or the process. But that’s a longer discussion.
Lovely thoughts in your article. I particularly like your thinking around the dynamics. Well done.
So far, the way we build products and interact with each other starts to feel different (work dynamics) , the way we are integrating new AI features and solutions to existing products opens new user behavior challenges (AI experiences or IUI) , and last but not least, it seems like having UI/UX or product knowledge is not enough to prosper in product development, so new technical upskilling sounds like a necessity (technical knowledge) .
Work dynamics are hard for many to accept, but I firmly believe that traditional workflows no longer make any sense. I’m curious if you find a group or mindset that really limits the work?
Excellent stuff @Bryan . I like this: “make problems smaller”..
I built the Weave on my own( https://weave.jokull.io/ ) to make research much smaller for companies (and much better), and with my electrical efficiency startup we are using Lovable to build 10+ workflow enhancements + growing the work to other affiliate companies.
I’m also asking Lovable to just build in the analytics directly into the Weave as then I can 1. get exactly what I want and 2. I’ve build it as a chat interface. No dashboard (dashboards are so constraining). I tried to use Mixpanel for 14 days and the barrier just to understand it was too high (Micpanel is built to accommodate specialists .. not normal people with other things to do ..)
Anywhoo .. It’s becoming soo easy and fast to build stuff now I wonder if we are going back to the beginning of the internet when there were 7 million people writing poetry and 7 people reading it … will we get to a point where we can exchange the ‘poetry’ with ‘apps’..? How do you win in such an environment?
Taking people out of their comfort zone is always a challenge; in this case, the principal barriers to embracing these new dynamics come from the company itself and people who are not open or willing to accept new communication and building ways; this issue is more prominent in big or traditional companies where, over the years, they have been working on static, almost monolithic processes, and with that, roles and people in power are not open to creating a collaborative environment. In my new role, that’s part of the issue I’m facing. Very challenging.
Finding T-shaped people is difficult. Even harder to keep them engaged in companies, as most companies just want to find talented people to fill a role. The pendulum has swung in the other direction.
AI is forcing more people into these collaborative roles, and in many cases, it’s simply not working, as many people just want to stay in their vertical expertise.
Yes, on my list to dive into! Great to hear from you.
I miss those days. Lol.
I’m bullish on the idea that we can build businesses that work like “mainstreet” to support small groups of people. Service as software is a real thing.
I honestly love the collaboration part of work, which has probably been why I’ve been so open to taking on work that might fit my particular expertise (being a software engineer).
Feels like most will have to switch to that mentality.
service as software (SaS) I might need to read up on this. But what I can see is that it’s not focused on selling seats (license), but outcomes … are we finally ready for the outcome economy? https://medium.com/design-bootcamp/the-outcome-economy-17431e5e2991 I’ve been waiting since 2014
With the Outcome Economy the company’s incentive and motivation becomes the same as their customer’s. They can work together to get to the same desired result. The customer can trust the seller more to have their best interest at heart as it’s the same as their own.
Seems at a smaller scale this is possible with all the tools. People with loads of experience can provide some of the same technical expertise that larger companies have productized. Only smaller, service-oriented companies can align the solutions better at a fraction of the cost.
Love this focus on behaviors. It’s what we are trying to provide in Glare in an open way… providing Skills to help people work through their decisions. Still a work in progress, but the user needs (which aligns with a broader situations has an overview, playbook, techniques and resources that shape the Skill.
I came across the youtube channel of O2ui which has this brilliant, but tiny thumbnail promoting one of their videos discussing the launch of Claude Design: “Designers can cook now”
I felt that really hit the spot. It’s not that we don’t need designers with brilliant ideas anymore, it’s that designers with brilliant ideas need fewer of everything else to prove their experience/idea, get users and start learning.
I assume speed to learning and quality of learning will be a critical measure moving forward.
Yes. Making decisions is predicated on good learning. Crap in, crap out. AI will accelerate what you put into your thinking. We need thinking systems to match the speed of AI outputs. In the short term, this will have to be considered as business stakeholders look for shortcuts.
has an overview, playbook, techniques and resources that shape the Skill.