I teach young UI/UX designers and try to explain to them that “design” is about creating a product, not about designing an interface. Steve Jobs has a similar quote: “Design is not just what it looks like and feels like. Design is how it works.” The focus of what designers do is shifting. AI reduces the time it takes to create UI options, but it does not solve the problem of finding product-market fit. AI cannot resolve conflicts between different stakeholders when creating complex products. Facilitation skills are now much more important than before. The principles of collaborative design, which have been developed since the late 1960s as part of Scandinavian cooperative design, are relevant again: designers help create products or environments, but they are not the sole authors. I think this is the direction in which agency work and the work of designers will shift: less “doing” design, more working as experts, facilitators, and professional researchers.
Small agencies often dont solve product market fit either, and still charge $25k a month just to keep the "product" team available... perhaps this "manifesto" speaks about those cases
Sorry but if this is your take then you might be a little detached from the actual design world. I encourage you to try the tools out there and then come back to this manifesto with a sense of what design really is and what it means to create PRODUCTION level design and applications.
Design adapts - If you’re not evolving your habits and processes then you’re not truly designing for the world around you.
I agree, Mike. But again, that's not the point - look at Figma's recent alpha release, the point isn't what it produces - its the speed that increases it.
Thanks Felix, interesting take. I'm an industry outsider (ex VC) but super curious on your points. I understand the role of agency falls somewhere along the lines of talent aggregators/editors, quality control and project management. I'm interested to see what tools are out there - that allow for designers to collaborate effectively ? e.g. Split proceeds, manage contracts, combine pitch (thinking part of metalabel's admin offer to creatives)
The toolset is changing for sure. AI is transforming the designers toolbox like power tools transformed the tradesman's toolbox. In recent history, the web transformed the print industry. Those who couldn't learn the new toolset didn't make it.
A new way to 'build' is quickly emerging. Embrace it or fade into irrelevance.
Agency has always been about gatekeeping, the taste-makers set trends. Everything has been done before, so everyone else just sample/copy that. The old guard are trying to hold on to what they can with legacy name brand. Networking will forever be valuable, it is hard to tap in, in an over saturated talent with the connectivity of the internet. I'm positive there is someone just as talented as you are, some where across the planet, playing with shapes on a crusty old laptop. Just hasn't got the opportunity. Besides going viral, can AI give these unseen people opportunity and visibility?
Love the thinking here Felix! Totally agree that the "Agency Model" is out the window! And has been for a little while now. It's truly antiquated and built to take a LONG time.
However, the take on how Notion and/or Figma can create a workable product in a week, three days, a day, or less with the use of AI. I'm just not seeing it, even 5 years from now.
A company, hell a person, would need to have a lot of things lined up in order for that type of result to even come close. I think that we need to stop talking about AI (or any of these "instant tools" Cap-Cut cough cough) as if it's a magical monster tool that will wisk away all the thinking and upfront work that needs to happen in order to create something that actually works.
The true devil out there, in my opinion, is that people are too focused on making a profit and monetizing things, and not focused enough on making something that actually works well.
Thanks for sharing Coop... on your last point, in my opinion the mindset that moves many in the product creation industry is: "this actually works, because its making a profit and I can monetize it more... forget about the design effort it took or my users, the faster i can impact the bottom line the faster I can move the needle in the business metrics that are important (ie ARR etc)....". Sign of the times...
I agree that AI can create many of the things you mentioned in the article. But there are two things often overlooked. First, designers are problem solvers. Second, human (animal, plant and environmental etc.) needs are constantly changing.
I accept that design and development tools are evolving. Allowing fewer people to work faster and more efficiently. But the target audience’s needs also evolve, and they continue to face new problems. I wonder that the companies determined their audeince new needs or problems after the AI was public. Or they add "AI chat feature" in their product only.
As a result, problems will never disappear; they will always exist as long as the world does.
That’s why companies need designers who bring communication, empathy, and the ability to identify and understand new needs and challenges. In fact, this means hiring more designers or design agencies, not fewer, to keep up with these changes.
This is my perspective as someone who uses AI in both professional and personal life.
Before Ai, a person’s value in the workplace was in whether they could demonstrate proficiency in hard skills. That’s how people were hired and the soft skills took a lesser role. But anyone who’s worked with a jerk star surgeon or seen the unedited side of a social media influencer or actor can tell you, there is another side that Ai will not so easily replicate; the ability to connect deeply with people. The big four soft skills (communication, collaboration, creativity, and critical thinking) have always been around but they are now, in a post Ai world, even more critical to master. We are entering a phase of intelligence where the Ai tool can build the website, write the copy so that it converts, and seemingly communicate as good as whoever it was trained on. It seems as though it can communicate, think critically, and creatively as well. But what is the fear that remains about Ai? Our greatest power, lies in our ability to connect and collaborate with others. This is inextricably tied to our character. Only those who master soft skills, areas most difficult for Ai to master, will be able to function in a world where the Ai is able to do 90% of all tasks. Humans crave and live for connection and that means we should retain a slight edge for the foreseeable future so long as we are willing to invest the required effort.
Interesting read, but I’m skeptical. The manifesto leans heavily into AI hype and oversimplifies the role of design. Speed and affordability are great, but design is more than quick outputs, it’s about strategy, empathy, and solving real problems. This feels more like a pitch than a balanced critique. AI is a tool, not a replacement for thoughtful design.
Are design agencies not adopting AI? It’s funny that serval days ago, I went to Frog Design, and they showcased how they deliver creative results with the help of AI, and even guide their clients on AI adoption.
I’ve been teaching UX & Product designers how to use AI in their daily work for almost 2 years, and my experience so far is that designers or not, everybody is using AI in some form. I’m sure there are a some agencies act slower than the others, but when it comes to survival, I’m pretty sure people evolve with new tools and new business models.
There are lots of great design talents and wise leaders in the creative agencies I know of, and I’m not too worried about them. I’m more worried about a small portion of some in-house designers who refuse to adjust and redefine their job description.
That’s the whole point. I accidentally stumbled on this post and I genuinely appreciate how the piece is written. It's rhetorically sharp and deliberately straddles the line between hype and realism. But beneath that balanced tone lies a quiet misdirection.
By overstating the idea that agencies are not adopting AI, it fabricates a convenient tension that flatters its own manifesto. The writing is clever precisely because it disguises simplification as nuance by anticipating critique and folds it into the narrative to appear objective.
In reality, most agencies aren’t rejecting AI; they’re negotiating its thresholds around precision, governance, and ethical velocity. The real frontier isn’t adoption, it’s calibration. How we tune accuracy, accountability, and discernment will decide whether AI scales mastery or just amplifies noise.
I disagree with the notion that the design agency model is dead. Agencies are adopting AI more quickly and effectively than in-house designers are. We don’t need to look far, or hard, to find top-level branding work (that uses AI) being shipped by renowned agencies (Collins, Pentagram, Smith & Diction, etc.)
I honestly don’t believe that individual designers can consistently outperform agencies just because of AI. Good work comes from the tension between different ideas and experiences. Agencies excel on that, solo players don’t.
Well, currently it is really hard for me to see the strong value (and real speed increase) in AI generated UIs. They are pretty basic, and you need to still to come in, and explore as you used to - not seeing huge time improvement there. For the proof of concept? Sure. Maybe some super quick and dirty MVP, to raise funds.
Sure - MidJourney can generate amazing artworks, that can be used as part of the brand visuals, or serve as an inspiration to take things further. ChatGPT / Gemini can generate great placeholders or real copy. But for now, what I see, is clients (reading articles like these, or posts on X where everything can be done in seconds) are changing their expectations. But it's not fully grounded in what AI can offer. At least not yet.
The "Anti-Design Agency Manifesto" confuses speed with value. It’s a dangerous oversimplification.
While I’ve used GPT models since the playground days, I see a glaring gap in the current AI hype: accuracy. Speed is seductive, but for mission-critical design work, it's a vanity metric.
"Vibe coding" is like using slot machines—great for rapid prototyping, but a costly way to burn through funds when you need strategic precision.
Baymard Institute’s research confirms this, showing concerningly low accuracy rates for AI in user-facing applications. When you need 100% accuracy, "mostly right" is 100% wrong.
AI is an accelerant. It speeds up brilliance, but it also speeds up garbage. Without proper governance, you get situations like the Deloitte AI scandal—where systems generate confident, plausible, but entirely incorrect outputs.
This is where designers become indispensable. Design agencies who understands this will still be in business.
Maybe. But if ‘good enough’ wins, we’ve automated mediocrity and mistaken speed for intelligence. Without accuracy or discernment, innovation will just scale human error faster.
The future of design is less about being replaced by AI and more about working with it. Eventually when AI crosses that border to meet accuracy most jobs will get redundant even influencers and hype marketeers but it still needs to cross that border first.
The work is shifting from execution to oversight:
1. Strategy & Context-Setting: Defining the 'why' before the 'what'.
2. AI Governance & Guardrails: Building the systems that prevent costly errors.
3. Model Training & Evaluation: Curing the AI of its biases and blind spots.
4. Quality Assurance: Filling the critical gaps AI leaves behind.
Investors who can't see this, risk funding beautifully fast garbage. The true value isn't in the tool, but in the master craftsman wielding it.
As someone who’s been in design for many years, I relate to these points. The way clients expect things now has changed a lot. For me, working faster and sharing the real process openly has become a normal part of projects. I’ve also noticed that clients value honesty and a fair price over creating a sense of mystery. Change can feel awkward, but it’s exciting to see new ways of working together. Thanks for starting this honest conversation.
I teach young UI/UX designers and try to explain to them that “design” is about creating a product, not about designing an interface. Steve Jobs has a similar quote: “Design is not just what it looks like and feels like. Design is how it works.” The focus of what designers do is shifting. AI reduces the time it takes to create UI options, but it does not solve the problem of finding product-market fit. AI cannot resolve conflicts between different stakeholders when creating complex products. Facilitation skills are now much more important than before. The principles of collaborative design, which have been developed since the late 1960s as part of Scandinavian cooperative design, are relevant again: designers help create products or environments, but they are not the sole authors. I think this is the direction in which agency work and the work of designers will shift: less “doing” design, more working as experts, facilitators, and professional researchers.
Small agencies often dont solve product market fit either, and still charge $25k a month just to keep the "product" team available... perhaps this "manifesto" speaks about those cases
Figma plugins generate polished UI in seconds.
Sorry but if this is your take then you might be a little detached from the actual design world. I encourage you to try the tools out there and then come back to this manifesto with a sense of what design really is and what it means to create PRODUCTION level design and applications.
Design adapts - If you’re not evolving your habits and processes then you’re not truly designing for the world around you.
I agree, Mike. But again, that's not the point - look at Figma's recent alpha release, the point isn't what it produces - its the speed that increases it.
Thanks Felix, interesting take. I'm an industry outsider (ex VC) but super curious on your points. I understand the role of agency falls somewhere along the lines of talent aggregators/editors, quality control and project management. I'm interested to see what tools are out there - that allow for designers to collaborate effectively ? e.g. Split proceeds, manage contracts, combine pitch (thinking part of metalabel's admin offer to creatives)
Very insightful.
The toolset is changing for sure. AI is transforming the designers toolbox like power tools transformed the tradesman's toolbox. In recent history, the web transformed the print industry. Those who couldn't learn the new toolset didn't make it.
A new way to 'build' is quickly emerging. Embrace it or fade into irrelevance.
Agency has always been about gatekeeping, the taste-makers set trends. Everything has been done before, so everyone else just sample/copy that. The old guard are trying to hold on to what they can with legacy name brand. Networking will forever be valuable, it is hard to tap in, in an over saturated talent with the connectivity of the internet. I'm positive there is someone just as talented as you are, some where across the planet, playing with shapes on a crusty old laptop. Just hasn't got the opportunity. Besides going viral, can AI give these unseen people opportunity and visibility?
Maybe I’ll come back and process a better comment, but yes. Yes. Yes. This is what I’ve trying to formulate into better words. But. Yes.
My field is classic car creation and restoration. Most of the points you raise are applicable to my field. Massive changes will need happen.
Love the thinking here Felix! Totally agree that the "Agency Model" is out the window! And has been for a little while now. It's truly antiquated and built to take a LONG time.
However, the take on how Notion and/or Figma can create a workable product in a week, three days, a day, or less with the use of AI. I'm just not seeing it, even 5 years from now.
A company, hell a person, would need to have a lot of things lined up in order for that type of result to even come close. I think that we need to stop talking about AI (or any of these "instant tools" Cap-Cut cough cough) as if it's a magical monster tool that will wisk away all the thinking and upfront work that needs to happen in order to create something that actually works.
The true devil out there, in my opinion, is that people are too focused on making a profit and monetizing things, and not focused enough on making something that actually works well.
Thanks for sharing Coop... on your last point, in my opinion the mindset that moves many in the product creation industry is: "this actually works, because its making a profit and I can monetize it more... forget about the design effort it took or my users, the faster i can impact the bottom line the faster I can move the needle in the business metrics that are important (ie ARR etc)....". Sign of the times...
I agree that AI can create many of the things you mentioned in the article. But there are two things often overlooked. First, designers are problem solvers. Second, human (animal, plant and environmental etc.) needs are constantly changing.
I accept that design and development tools are evolving. Allowing fewer people to work faster and more efficiently. But the target audience’s needs also evolve, and they continue to face new problems. I wonder that the companies determined their audeince new needs or problems after the AI was public. Or they add "AI chat feature" in their product only.
As a result, problems will never disappear; they will always exist as long as the world does.
That’s why companies need designers who bring communication, empathy, and the ability to identify and understand new needs and challenges. In fact, this means hiring more designers or design agencies, not fewer, to keep up with these changes.
This is my perspective as someone who uses AI in both professional and personal life.
Before Ai, a person’s value in the workplace was in whether they could demonstrate proficiency in hard skills. That’s how people were hired and the soft skills took a lesser role. But anyone who’s worked with a jerk star surgeon or seen the unedited side of a social media influencer or actor can tell you, there is another side that Ai will not so easily replicate; the ability to connect deeply with people. The big four soft skills (communication, collaboration, creativity, and critical thinking) have always been around but they are now, in a post Ai world, even more critical to master. We are entering a phase of intelligence where the Ai tool can build the website, write the copy so that it converts, and seemingly communicate as good as whoever it was trained on. It seems as though it can communicate, think critically, and creatively as well. But what is the fear that remains about Ai? Our greatest power, lies in our ability to connect and collaborate with others. This is inextricably tied to our character. Only those who master soft skills, areas most difficult for Ai to master, will be able to function in a world where the Ai is able to do 90% of all tasks. Humans crave and live for connection and that means we should retain a slight edge for the foreseeable future so long as we are willing to invest the required effort.
Interesting read, but I’m skeptical. The manifesto leans heavily into AI hype and oversimplifies the role of design. Speed and affordability are great, but design is more than quick outputs, it’s about strategy, empathy, and solving real problems. This feels more like a pitch than a balanced critique. AI is a tool, not a replacement for thoughtful design.
Are design agencies not adopting AI? It’s funny that serval days ago, I went to Frog Design, and they showcased how they deliver creative results with the help of AI, and even guide their clients on AI adoption.
I’ve been teaching UX & Product designers how to use AI in their daily work for almost 2 years, and my experience so far is that designers or not, everybody is using AI in some form. I’m sure there are a some agencies act slower than the others, but when it comes to survival, I’m pretty sure people evolve with new tools and new business models.
There are lots of great design talents and wise leaders in the creative agencies I know of, and I’m not too worried about them. I’m more worried about a small portion of some in-house designers who refuse to adjust and redefine their job description.
That’s the whole point. I accidentally stumbled on this post and I genuinely appreciate how the piece is written. It's rhetorically sharp and deliberately straddles the line between hype and realism. But beneath that balanced tone lies a quiet misdirection.
By overstating the idea that agencies are not adopting AI, it fabricates a convenient tension that flatters its own manifesto. The writing is clever precisely because it disguises simplification as nuance by anticipating critique and folds it into the narrative to appear objective.
In reality, most agencies aren’t rejecting AI; they’re negotiating its thresholds around precision, governance, and ethical velocity. The real frontier isn’t adoption, it’s calibration. How we tune accuracy, accountability, and discernment will decide whether AI scales mastery or just amplifies noise.
I disagree with the notion that the design agency model is dead. Agencies are adopting AI more quickly and effectively than in-house designers are. We don’t need to look far, or hard, to find top-level branding work (that uses AI) being shipped by renowned agencies (Collins, Pentagram, Smith & Diction, etc.)
I honestly don’t believe that individual designers can consistently outperform agencies just because of AI. Good work comes from the tension between different ideas and experiences. Agencies excel on that, solo players don’t.
Well, currently it is really hard for me to see the strong value (and real speed increase) in AI generated UIs. They are pretty basic, and you need to still to come in, and explore as you used to - not seeing huge time improvement there. For the proof of concept? Sure. Maybe some super quick and dirty MVP, to raise funds.
Sure - MidJourney can generate amazing artworks, that can be used as part of the brand visuals, or serve as an inspiration to take things further. ChatGPT / Gemini can generate great placeholders or real copy. But for now, what I see, is clients (reading articles like these, or posts on X where everything can be done in seconds) are changing their expectations. But it's not fully grounded in what AI can offer. At least not yet.
The "Anti-Design Agency Manifesto" confuses speed with value. It’s a dangerous oversimplification.
While I’ve used GPT models since the playground days, I see a glaring gap in the current AI hype: accuracy. Speed is seductive, but for mission-critical design work, it's a vanity metric.
"Vibe coding" is like using slot machines—great for rapid prototyping, but a costly way to burn through funds when you need strategic precision.
Baymard Institute’s research confirms this, showing concerningly low accuracy rates for AI in user-facing applications. When you need 100% accuracy, "mostly right" is 100% wrong.
AI is an accelerant. It speeds up brilliance, but it also speeds up garbage. Without proper governance, you get situations like the Deloitte AI scandal—where systems generate confident, plausible, but entirely incorrect outputs.
This is where designers become indispensable. Design agencies who understands this will still be in business.
Would AI eventually change this? When it gets good enough, I guess is yes.
Maybe. But if ‘good enough’ wins, we’ve automated mediocrity and mistaken speed for intelligence. Without accuracy or discernment, innovation will just scale human error faster.
The future of design is less about being replaced by AI and more about working with it. Eventually when AI crosses that border to meet accuracy most jobs will get redundant even influencers and hype marketeers but it still needs to cross that border first.
The work is shifting from execution to oversight:
1. Strategy & Context-Setting: Defining the 'why' before the 'what'.
2. AI Governance & Guardrails: Building the systems that prevent costly errors.
3. Model Training & Evaluation: Curing the AI of its biases and blind spots.
4. Quality Assurance: Filling the critical gaps AI leaves behind.
Investors who can't see this, risk funding beautifully fast garbage. The true value isn't in the tool, but in the master craftsman wielding it.
As someone who’s been in design for many years, I relate to these points. The way clients expect things now has changed a lot. For me, working faster and sharing the real process openly has become a normal part of projects. I’ve also noticed that clients value honesty and a fair price over creating a sense of mystery. Change can feel awkward, but it’s exciting to see new ways of working together. Thanks for starting this honest conversation.