Design is Dead Final, Final

Welcome to The Daily Sprint.

Today we're talking about the future of design.

It's not dead, but there's a lot of confusion about where it fits in with AI.

I'll show you the one constant you need to know no matter what the latest tool is.

I'm Darrell Estabrook, 30 years into UX Product Design.

Yes, I've seen design feigned death a number of times since then, but we're still just getting started.

I'm also the founder of designing.

I coach designers into design leaders through real-time interactive product specific guidance, skills like turning vague stakeholder requirements into clarity, and designing value into your product at every level.

That doesn't just happen by accident.

It's made on purpose.

Find out more and get on board with a free newsletter at designy.com. That's “design” with a ”y” dot com.

So design is dead.

It used to be that UX was dead, but there's a lot of chatter going around about design, or at least it's implied that design is dead.

It's like there's the new thing in town and the old thing is dead.

What do they say in England?

The king is dead.

Long live the king.

It's definitely sensational, but get you a headline.

But we're talking about AI, specifically generative AI in design.

And there's a whole spectrum of AI that we should map out and and discuss at some point, but really, there's kind of this bucket of AI that generates an AI that analyzes as kind of a in between.

Although it's always a mix since even the analysis is generated, right?

There's some generation at point.

But the biggest challenge is in all these cases, the output is good enough.

Right?

It passes for human output.

And for casual uses, this is fine.

Because it doesn't matter.

You're asking questions like you would in search engines.

You're getting generated results based on a body of work.

But for serious uses, there's always a confidence score.

How accurate is the output that the model thinks the output is or, you know, based on what it's looking at so that you know as a user, okay, even though this has authority, it's 75% accurate.

Maybe think of it a different way for predictable output, it's fine.

Like code that has to be a certain way in order to function.

Right?

That's it's predictable.

It's structured.

There are definite rules to follow and you can repeat it.

But for multifaceted output, stuff that requires a lot of different maybes, it's very opinionated and can be myopic, just singular.

Like visualizations, image generation.

We're going to talk screen generation, complex analysis, or think about it whenever it depends is part of the mix.

It just gets very complicated.

And that's because there's a risk that it could be wrong.

That's because there's a risk, it can be wrong.

And it says so in the little disclaimer below the chat box.

AI can make mistakes.

And it's funny for all the years of UX. Discussions around the cardinal rule of you have to tell the user to do something, then the system isn't doing it well.

It was poorly designed.

It's kind of gone out the window, somehow saying that it might make mistakes, doesn't instill a lot of trust or shouldn't instill a lot of trust in the output.

And yet, we're using it anyway.

It was very interesting.

So kick this all up a level.

And we're asking AI to generate screens and flows of complex apps and calling it a design teammate.

We've elevated AI to be a coworker.

And it's amazing because it's instant in its output.

And what it produces is very authoritative in, in its presentation.

And it fills in gaps that it has in context or anything in the prompt, it fills it with assumptions.

And so all of that combined together puts the designer on the defensive instead of offensive, right?

They're being channeled to follow the path. Instead of using the tool to push itself to a different solution.

You don't really have to work at it to go the other direction.

So designers need more engagement, not less.

So take a look at this.

Post on LinkedIn this past week on the design process and a proposal.

It was a graphic and it had the old process and the new process, like 4 bullet points, 2 columns.

The old process was shown as these 4 parts, cascading one to the other.

Think, specify, design and build.

Like that's the old process.

We think, then we specify.

Then we design it, then we build it.

The new process was shown as build, evaluate, refine, and document.

So it's like build it 1st, then evaluate it, make some tweaks, and then document it.

And the old process in this graphic had a big X through it.

It's dead.

Old process.

And the point was, instead of building at the end, now building is at the beginning.

And so what do you think about that?

How does that strike you?

I get it, right?

AI allows one to start experimenting with working software immediately.

And there's an advantage to that, there's definitely a lot of things you can't do in figma or static designs.

And it's a path in a tool, an ability worth having.

But we're not really talking about that and the ability.

It's really a question of where to start.

And starting with building has always been the issue for years of starting with code.

There was a big movement, and even tools like web flow, where it's a visual designer, but it's coding behind it, but you're still stuck in the rules of cascading style sheets and you can't move things around where you want to.

And there's a lot that you have to think through technically when you're designing.

This is more of that, except now we've gone to working code with data and all of that.

So the biggest question you should be asking when looking at this new process, the new process of build, evaluate, refine, and document.

It's how do you know what to build?

The proposed new process is kind of saying just start.

Which is good advice when you like to procrastinate and, you know, you're pushing something off, just start.

Just do something.

That's that is a good thing.

But we're talking about intentional business related software design problem solving.

Even with just starting, like, I'm going to sit down with a prompt box.

You have to have an idea.

At least you have to you have to have something that sparks it.

And even that, you have to have more than just an idea.

It has to have a purpose because you're aiming for something.

It's not just to create a thing with AI.

And think about the old process that was crossed off in that diagram.

It was think, specify, design, and build.

Now that's out.

We don't need to do it in that order, or maybe at all.

Just build.

But these are critical concepts.

And I say them as concepts because you should never have a step, one of these 4 steps where you're not thinking, specifying or designing.

And in fact, you should never have a time when you're not building, evaluating or refining either.

The documenting thing, nobody likes that anyway.

So that's a separate process, and it has a separate purpose.

If you don't need to document, you shouldn't, but that's a whole other topic.

The big D design is this whole process of translating concepts into outcomes, right?

We have an intended outcome and we want an actual outcome.

And how do we do it?

And there is this, this entire battery of skills.

The little D design. Is the specific skill and talent of visualizing, skillfully arranging, and otherwise communicating a plan to be executed.

Right, we're going to fulfill this plan.

Are those discrete steps?

Well, they are linear because we do live in time.

We call this the daily sprint because we can only do so much in a day and we should be focusing on what we can do in this day.

But in order to get that outcome, a week from now, we need to set the dominoes up today.

So there is a linear flow to things.

And even if you speed it up, it's still sequential, it's just shorter.

But there is some value in the time and there are other people involved than just the designer.

So communication becomes a very big glue that isn't obvious when you talk about these processes and steps.

So that's why I propose the design-y framework.

It's so universal in that it's not tied to any specific tool or moment in the process.

And it's very scalable because you could use it and talk about the entire strategy of the app in these terms, or you can talk about an individual element that you're designing and run through it really quickly, like in an hour.

But they are always present.

And if they're not present and we're not thinking about them on purpose, then we're going to miss something.

We're going to shortcut short change or not know why we're doing what we're doing.

And that's huge, especially when there's other people involved in the process.

So it's 5 points.

And we've covered them before.

Well cover them again.

We'll go deep into these.

There's so much to really dig into these and how to apply them, but initially, the 1st one is define purpose.

It has to be defined.

The question is for you as a designer, is it clear what the intended outcome is, and how do you clarify it if the stakeholders can't?

Happens all the time?

Even with AI, what is the prompt?

Well, you got to clarify this outcome.

And usually it's not from you.

You are dependent as a designer on a strategist, on a business person that they're, they're the ones who are ordering this to be delivered.

The 2nd one is discover context.

Are you getting accurate information surrounding the purpose?

So much of that is tied up in research, but research doesn't have to get complicated, but it has to be done.

You have to get some kind of sandbox, barriers, borders.

You have to define what this is, discover the extent of it.

Number 3 create options.

Generating multiple screens is only as good as your purpose and context are clear.

Right?

Even then, are you skilled in knowing what to tweak and why?

AI generates it, but there's a lot going on in that screen.

Are you clear on what everything in that screen is doing, where it's going?

Number 4 is make choices.

Just because it's a screen doesn't mean it's where you need to go.

Like here, this is, this is, this is the plan.

It popped up.

Besides, how do you get designs approved when there are challenges communicating to non-designers like stakeholders and engineers?

Do you have explanations for the design decisions and then how do you guide them to make the strongest choice?

And then number 5 is measure results.

Are you getting accurate and real feedback on the outcomes?

There's only so much statistical analysis that would work, but what is your use case and your business for your users.

We are very much the same as people across the board, but we are also very different.

And so there are unique niches, business processes, things like that that really matter.

And not all the patterns are appropriate for every situation.

So yeah, these are specific techniques.

There's a lot that goes under each one of those, but they don't rely on any tool or technology to be effective.

It's definitely a framework.

You just have to know how and when to use them.

So where does AI fit into this?

Well, AI is a tool.

It's not a solution.

It's not a person.

It's not relatable in that sense.

But as a tool, it has uses.

Right?

A stapler isn't a hammer, but I've used it as such on a number of occasions, not with very reliable results, but it's a tool.

It absolutely is.

And AI is extremely sophisticated.

It's one of these sophisticated tools which has the potential to divert us, in our focus, unlike any tool in human history.

Someone compared it to the revolution of moving from an abacus to a calculator and just like, that's a, it was a step change in that, you know, concept.

Yeah, but a calculator is always going to produce the same answer regardless of the question.

But these are fuzzy questions we're asking AI.

We're asking it to think like a person, but there's so much that goes into it.

And then we're talking design and there's even more involved in that aspect of it.

Will it get there?

I don't know.

We'll find out.

People are still trying to find a place for AI now, and the tools are changing all the time.

So it's a moving target if there was ever a chance to use that word.

So all of that together, really as designers, we can't afford to forget these fundamentals of design, regardless on how sophisticated AI gets.

Because in the end, who will be the final arbiter of what is effective in these hyper-specific use cases of your product?

The answer is you are, if you're in the position of a designer.

If you're not in that position, then AI will make those choices too.

It'll design it and decide and deploy the code.

So yeah, design is very much alive.

And you need to ground yourself in it more than ever.

If you're interested in learning how to establish these principles, the designing framework and all that to the product you are working on, well, I can help.

I love working with people directly, interactively and specifically around this product.

I have a 3 month program.

You can find out more about it on academy.designing.com

So what makes us unique is how regular we focus on building.

Courses and conferences and even AI can't really answer your questions when you need them, to the depth you need them, with the context you need them.

But these are weekly one-on-one sessions to give you the skills that you can't get from just tools, like turning vague stakeholder requirements into clarity and making design decisions which add value into every level of your product.

There's unlimited private designy slack, channel for asking questions, reviewing designs, or even getting advice between these sessions.

And if you have a manager, send them the link too, especially if you need to approve a training budget, but they can see these challenges of designing for outcomes.

That's their goal.

They are vested in making sure that the designers on their team are really focused on that, are enabled to do it independently and regularly.

So that's academy.designing.com or whatever, that’s “design” with a “y” .com.

So keep getting involved, follow, subscribe to the daily sprint on whatever platform you're listening to right now, and just thanks for listening.

I appreciate it.

Glad you're here.

See you next time.

Remember, today is a great day to design with a why.

See you later.