Pushing Claude Design to the Limit: Building a Startup Concept 🛠️

Let's try Claud Design together.

So hot off the press, Claude has just released its new design functionality and today we're going to be really pushing it to its limits, getting our hands dirty and seeing what we can break.

By the end of of Claude Design.

Now we're going to be doing this live, so you're going to see me struggle, you're going to see me potentially run out of tokens, you're going to see us really tear this apart.

but that's part of the fun, right?

So let's jump in.

So to get to Claude Design, how do we get the feature open?

So we come up here and we go Claude AI Design, press Enter and you're presented with this screen right now.

there's an option to build kind of a prototype and a wireframe.

You can do slide decks in here, and you can even So if you've got a template or a brand template that you've already used, you can put that into here and kind of even if there's a company name or a company organization that you really like their design system of, you can put that in here for inspiration.

So that's really epic.

But let's go back.

We won't do that.

we need to come up with an idea that we want to tear apart.

And one idea I've been thinking a lot about at the moment, is how can we encourage more use about energy saving behaviors in the household.

So we see a lot of these apps that are coming along about trying to encourage families to save energy and reduce their bills.

But often they're just targeted at the bill payer themselves, whereas it's the whole family actually using the using the energy system.

So is there a way that we can kind of gamify some of that a bit?

Maybe like a Duolingo or you know to have some kind of leaderboard with it within the family but also within and against other households, maybe in the area of a similar type.

So let's tear apart that idea now, we'll call it Energy Gamification App.

And I'm going to start with a wireframe because I want to throw a few ideas around with the tool and I don't want to go and burn a load of tokens straight away.

So we're going to start there.

Let's click Create and you'll be presented with this screen.

You can upload some Context, which ideally we would do.

But let's actually start by just typing a bit about that idea.

So I want to create a gamified app, to encourage all members of a family to save energy.

And reduce the household bills.

Currently this is mainly a concern of the main bill payer but we want to gamify it so all members of the household including children care.

We also want to create some competition with other households.

Ask me any questions to create about the idea UX UI to make a wireframe as effective as possible.

Now the reason I've done that last bit is I actually want to trade questions and I want it to pick apart the idea with me.

so we create a bit of a plan and we effectively don't waste a load of tokens going in the wrong direction.

So yeah, a pound in a pounding planning is worth a huge amount of value down the line.

That's possible.

So we'll click enter.

Let's press send.

So Claude design is now wearing away and you'll see a really nice interactive interface comes up asking me some of those questions.

Right, so primary platform, where will this live?

Yeah, on a, on a mobile app most likely.

who's the target family?

That's a great question.

I'm gonna say we want it to be kind of two parents and young kids.

that would be the typical kind of household and then it'll really hopefully get that element of fun with the young kids.

Should roles differ by family member?

I think ideally for this first everyone should just see the same thing.

but you know, maybe we could explore a few options down the line.

So we'll leave that there.

How does the app know energy use?

This is a great question.

My mind it has, this has to be a bolt onto an app from an energy company already that's kind of connected to a smart home.

So yeah, I think it will have to be really like a smart meter, live feed or smart plugs.

For example, what game mechanics should it feature?

You know, do we want quests, do we want points?

I'm really, you know you can click multiple things but these are all really cool ideas.

Right.

Like it's already starting to stretch some of our thinking which is, which is really great.

Definitely something about streaks.

maybe I'm going to leave this open with kind of explore a few options but I really like these kind of streaks collectible characters.

I feel like they kind of all tap into a lot of the apps that are popular with kids at the moment.

Right.

So if we crack on, let's continue.

Should it be the versus.

Should.

How should the other versus other households feel?

Match made Similar households?

I think so, yeah.

Because otherwise it's just going to be whoever's got the the best.

How you need the, the most energy efficient house that's going to win.

So I think you kind of want to have similar household matching.

But actually this idea of kind of private leagues is also really neat.

So that sounds great.

What would actually motivate the kids?

I mean I'm really open to a number of these ideas so I'm going to leave as explore a few options.

What would motivate the bill payer?

I think obviously seeing pounds saved each week will probably be the big thing.

Right.

what screens must we have in the wireframes?

Again, explore a few options but definitely we want kind of a leaderboard, maybe some kind of dashboard and a live energy view.

Visual tones.

I really want to play into that kind of duolingo feel.

So kind of I think and you know this is going to be won or lost by having the kids on board.

Right.

So I think we do need to have that.

And also if we're honest, we're all bought into Duolingo and the Green Owl anyway.

So I think this is great.

How many wireframes do we want?

Let's start with four.

how many screens per direction?

Seems like five.

Any tricky UX moments worth wireframing?

Let's think about this later.

but generally we want to continue to politely nudge to good behaviors.

Anything else I should know?

No.

No constraints.

Okay.

So we can, we can then click continue.

So as it whirrs away, so already you can see some kind of directions that it's planning.

It's come out with kind of household pets.

Each family member has a virtual pet that thrives on saved energy.

Some kind of garden, maybe like a power squad.

I'm interested to see how these ones turn out.

What I am noticing is it definitely feels like it takes a longer time to get to a first version than perhaps some of the other Claude functionalities.

Which is, which is understandable.

But there's maybe other stuff we could start doing in the background as well.

I'm also really intrigued after this to see how it compares to some of the other tools for creating presentations and decks.

And I've seen some videos where they're actually also experimenting with kind of creating videos from this as well.

So we'll have a play around with that after.

So Whilst that's wearing away in the background, maybe.

See what starting a new chat does.

I'm going to take that initial prompt and whilst we're waiting I'm going to say help me create a venture, a deck to raise 2 million from Sequoia Capital for this idea.

Make it using best practices, make it very outcome focused.

Send.

So we'll see whilst that's wearing away, there's some, some questions to ask here about the presentation deck.

what's the company product name?

We haven't got there yet but placeholder name like Wassail.

I kind of like that.

I found a background.

And why now?

just fill in a template here for the time being.

I'm going to launch across the US first.

Please fill these in revenue model decide for me.

I'm just going to leave this fairly blank and see how it manages.

Right.

mix of the above.

So it's mostly going to be smart meter API.

Let's make it bold and punchy visual direction.

let's kind of really condense it down to like 11 slides max.

no deco only wireframe separately.

No, we don't need we can do some, maybe do some short bullets of speaker notes.

Let's see how it does.

Now that's really kind of like a one shot.

Right.

I haven't really prompted it at all.

So we'll see how that goes.

But in the meantime we'll go back to our game and see how that's doing.

It's still wearing away.

Right.

You can see it's already kind of working on number one household pets.

So he's really investing a lot of energy into this.

Right.

It's still working on number one which is kind of interesting in itself.

We're really throwing everything at it now.

So the wattle deck is going.

Let me know if you think there's better names for this idea.

Wattle maybe seems like a good version one.

Okay, cool cool cool think these tips are always quite funny as well.

Right.

They're like very designer focused.

So it's already kind of outlining some of the palette as well.

Telling us.

here we go.

Something's working.

Which, which one first?

I think the wireframes.

Right.

This is interesting.

Still firing.

Something's happening.

Okay.

so something's broken.

Let's see if it works itself out.

What's it saying?

It's saying it's had some Okay, cool.

So this is the deck that's come through, wow.

It's like very big and bold.

I like it.

Turn the family energy bill into a game.

I mean, it's kind of, it's a pretty good hook.

It's not, it's not anything amazing, one person pays, everyone burns, bills arrive monthly behaviors happen hourly.

The payer lectures, the household tunes out.

Energy waste is a coordination problem, not a tech problem.

Now I really like that.

okay, it's got the kind of AI tells of, it's kind of got the AI tells of.

Energy waste is a coordination problem, not a.

Like it's X, not Y.

But it's still pretty, it's still pretty good, right?

Like, I think what is really nice is it's really hit into that kind of personal element.

So that's pretty strong.

we'll come back to the wireframes in a second, but.

Okay.

so I think this is pretty cool.

It's like really getting that personal element out.

We can probably all associate with this a bit.

only one in five family members can name the household's monthly energy costs within $50.

I mean, that's also pretty cool.

I think these are good stats, right?

Like, there's a big problem there.

Now what it hasn't got right is I think like the color palette's pretty, this kind of cream on white's not super strong.

But besides that, I kind of do like the big bold kind of font and format.

and what's this saying?

Why now?

Okay, so we've reached a point where the infrastructure's there.

number of.

this is like about the bill paying increase, right?

So there's a pay in there.

And, and also we've kind of got this captive market of kids.

So I think this is kind of neat.

You know, it's.

Has it got to the essence of the problem and like put some stats behind it?

Yeah, I think that's pretty cool.

behavior change dies on monthly bills.

It lives on real time feedback loops.

So bill cycle, 30 days.

Thermostat app, one day duolingo streak, 24 hours.

I kind of see what it's trying to do here, but I think it would probably need a bit more backwards and forwards because I'm not like super clear on, you know, why it's introduced these numbers, but I think basically it's starting to get to the sense that there's reoccurring behaviors.

We haven't quite got to the solution yet.

Like every.

What is a score?

Every kid is a player.

I mean, that's really?

That's epic.

Right?

That's like a really pithy line.

connect the smart meter, assign avatars and budget run missions, earn points.

I mean you know the wireframe could maybe do a bit of work.

But like I think as a kind of initial first first go considering how little we did this is pretty strong.

and imagine if you actually already had your kind of MVP products or dex to feed it.

I think it would be even better.

Right.

So this is just like okay, our stuff works again really neat showing the average saving like that there's actually retention.

I think they'd be some of the big concerns with this.

Right.

So again like maybe it needs a bit of work but and you know the design here is a bit funky but what is the -18% but it's not, it's not bad.

10 figure retention market.

So here's our market sizing US households and smart meters.

Average electricity spend target attach rate consumer utilities.

I mean it's had a go kind of global residential energy software.

What will target Samsung?

you know if you had done your own calculations on this anyway and so I think actually the visual format's not too bad.

two rails, one product kind of.

Okay, so there's a family subscription and maybe a paid for by the utility.

I thought this could you know is there a more exciting kind of play here around incentivizing good behaviors with kind of rewards and stuff?

but it's decided to go down this route.

Utility B2C Again you'd have an ingoing view but it's kind of.

It looks good.

Okay, so this is just a team slide.

Fine.

here's the spend slide.

bill payer has fought alone 50 years.

We're giving them a team.

Nice.

I mean that is a really nice close.

Right.

It's kind of like a big hairy audacious clothes.

I really like it.

let's quickly whilst in the background let's just go to the deck and say looks good as a first version.

Some of the white font is hard to see.

But look I think if we zoom out right.

If you've prompted slide creation before on any of the other tools I this is actually I think best in class.

Right.

It's got some really pithy structures.

The visuals for the most part look good.

It's.

It doesn't look like a deck that's been made by, by AI so you know I'm really impressed by this.

I definitely am.

So at the same time let's it'll whittle away at that.

But let's go back to our wireframes.

Awesome.

So this is basically it's split out some different wireframes by four different ideas.

Okay, so first one, household pets.

Each family member has a creature.

Okay, so this has come back already.

hi.

This is actually an issue throughout.

So it's come back and said that, it's only an issue on two slides.

Please look at the whole presentation.

Okay, so let's go back to the wireframes.

four distinct directions, one gets chosen, then we prototype it for real.

So what I like about this, right, is there's kind of all like sticky notes and working notes over the wireframes.

Almost like it's had its own whiteboarding session, which is really cool.

and you know, this is like, it's kind of.

What's this first one?

It's kind of given each individual like their own character, to kind of train and how much they're saving.

it's also got the smart meter data and like, I guess like energy savings of feeding the pets, which is kind of a neat, a neat Duolingo esque thing, right?

Giving some quests and like a family leaderboard and also kind of a street league out off the bat, right?

This is a really neat idea and you can kind of imagine yourself sitting around with your team and using this to kind of go, okay, which bits do we like, which bits do we don't like?

just like a real design session.

So, okay, number two, pet sequels.

Emotional hook.

Careful.

pet getting sad, can stress kids out.

So it's giving some notes as to like comparing the options as well.

Which is clever, right?

So maybe instead it's a garden plot.

you know, weather in the garden.

kind of think this is a bit clunkier, right?

Like it kind of doesn't necessarily feel, quite clear that it's linked to energy and stuff and maybe not as motivating for kids, but yeah, it's good.

It's still interesting.

Power Squad family, the football sports team.

I mean it depends, right?

And almost like you could almost choose options here.

Maybe the user could actually choose the UI and the game that they want to play based on what they know their kids like.

So You've got a kind of football thing here with this.

You know, you do well, your team's going to do better in the, in the scorer board and you've got a league table against, friends.

This is kind of Neat.

Neat.

It almost reminds me like a fancy football league.

Right.

So this can be pretty neat.

And who's the best players?

Yeah.

Awesome.

I think that's like, for a certain type of family.

That.

That and the pet ones are kind of my favorite options at the moment.

What watches?

it's kind of like weight watches.

Right.

I kind of think that name would, would stand out in a bad way.

but look, this is kind of a bit more focused on the actual bills and probably closer to kind of normal apps, so probably further away from our kind, of gamification concept, but, you know, could still be interesting.

So now we've done these wireframes for me.

I'm going to look at like 1 and 3, this kind of Power squad idea and the household pets idea.

And what it's asked us to do is four low wireframes, five screens each.

use tweaks in the toggle bar.

Yeah, let's do this as well.

Like, it's cool that you can kind of change some of these and maybe they change live.

It doesn't look like they have.

Right.

but you can turn things on and off.

You can turn annotations on and off.

Something seems to be wrong with the accent color, but interesting.

tell me which direction resonates.

I like the options 1 and 3 for further iteration and to make more high five, medium phi.

Let's go with those.

So we'll let that wear away.

But look, we come back to our deck and things are looking a lot better.

Right?

And.

no, guys, look, we've hit our limits.

so I'm going to add some credits just in the background.

But, you know, it goes to show.

Right.

Like everyone says, this stuff burns through credits.

And, that's kind of case in point.

So I'm just going to buy some extra usage in the background.

Kind of every video I've seen about this is how much it burns through credits, so it's really interesting to see that play out in real life.

Okay, so I've topped, up.

Okay.

Retry.

I've topped up credits.

Let's see what this does.

Okay, so we're back going.

but let's look at the deck now.

Like with that black font, I think, you know, that actually looks really good.

It's.

It's far better than kind of any of the other screen tools that slide tools that I've actually used before.

There's still some kind of formatting issues here.

But you know what I like about it is it's got some really kind of really good copy within here.

so yeah, I think it's really strong.

let's go back to this.

Can we continue?

So something's happened.

We may have broken it guys.

So let's go out and in again.

So we'll go back here.

We'll go back here.

it.

Interesting.

Doesn't seem to be liking it.

So we're back guys.

We've topped up the credits and it's now wearing away and creating those two more detailed versions for the two options that we liked.

It does go to show that you know this is very intense and hungry right.

it will eat up your allowances and I've seen that from everything kind of every video that's out on Claude design at the moment.

So you see it's also kind of compared to Claude code probably bumping into errors more frequently I'm noticing and then going through and kind of fixing all of that.

So you know that probably feeds into to why it's so hungry but we can quickly flick back through the deck in the meantime.

I think this looks really sharp now.

Right.

it's also kind of done some speaker notes for us which is pretty neat 10 figure attention market Nice.

Still wearing away.

Okay, cool.

It looks like it's loading now.

So as a reminder it's now like taking it up a level from those wireframes and.

Wow, this is pretty neat.

Right?

it's already giving me kind of the Duolingo type feel.

We've got these little avatars now.

and each, each family member's got a little pet.

That's pretty cool.

and let's scroll.

So you've got a bit of a streak going.

You've got some kind of live energy monitoring and I really like this quest concept as well.

Right.

It's kind of it's just really clever.

Biscuit is happy you fed them two stars.

Cool.

Like I don't know about you but whenever I'm looking at Duolingo it's, it's, it's this kind of super simple UX kind of really bright if it feels like it would be a good fit.

And this is also, this is the the football example.

So like we've started to got a bit of a kind of a win streak.

It looks really like a fancy football app.

let's have a look Patela.

See and you've got kind of league table again, pretty neat.

Cool.

Well I'm super impressed with that.

let's say what it said verify claimed already.

When you are pick a direction or a mashup and I'll push it to hi fi with real interactions.

Let's, let's do.

I'm personally going to go with the pet side there.

Hi Fi of the pet side air.

What I'm also going to do is whilst that's going okay, what phone do we prototype most deeply?

let's do the kind of both equal depth which end to end flow should actually work.

Daily check in maybe decide for me.

Decide for me.

So I heard for me which tricky moments matter most to solve.

Yeah, I kind of want to see how we can encourage engagement throughout all of this.

So that's definitely a view.

I think we should still keep that deep view for parents interaction and animal fields.

Pre bouncers, Let's go.

Okay, let's continue with that.

now another thing I want to try a bit more of is some people have said okay there's a slide deck there.

There's also you can do some kind of animation.

So energy app.

I don't want to create from a template but let's see, let's see if we need a template here.

animated video.

Okay so let's go back to go back to our energy app and maybe we can just do it in here.

energy gamification app.

And if we click plus and we do a new chat, maybe we can turn on.

Okay, maybe not.

But what I'm going to do is I'm going to go back to this and I'm going to see how good it is at creating, an animation separately.

Right.

so we'll go over to the energy app.

I think there's an animated video there.

actually, do you know what I'm going to do?

I'm going to go to the gamification app and I'm going to download the slides and use that as an input to creating a product video.

so if we go to present to share, we go to download, export as PDF.

Let's do that.

Okay.

Print PDF, let's do that.

And then I'm going to go back to our app, and let's see if we can make a product user, video for which I will use to launch my app.

See the slides attached for contact.

Okay, So let's see how it does with the animations.

Let's just make it a short teaser.

YouTube, playful, and friendly.

Decide for me.

Decide for me, Decide for me.

Continue.

I'm really intrigued to see how it does with video here.

Right.

So let's just go back whilst we're waiting.

think that's going to churn away.

It's got some ideas.

and we'll go back to our app in this other.

Okay, so let's see how everything's going on down here.

Continue.

So that's interesting.

If you exit the session, it doesn't run in the background.

That's good to know.

But we've got it going there.

We can go back and re review this as we're waiting.

Right.

So.

Trying to have a look.

Is there.

Anything else to call out?

Yeah, I definitely prefer this design.

It's it's a lot cleaner.

So it's over here creating our animation.

something's gone wrong.

It's really interesting flicking backwards and forwards as well.

Right from this kind of low fire to medium fire.

So you kind of see how we're going on that progression loop and it just really tidies things up.

from this kind of first one to.

The four creatures still, it's really strong.

Okay guys, so things have been wearing away.

We've got the, the demo trial video.

So let's have a look at this first.

Right.

Okay.

So ready?

It's a bit clunky, right?

Okay.

I mean it's a bit all over the place but like there's some interesting elements in here.

Right.

I think it's still loading.

So maybe we give it a bit more chance because currently it's feeling a bit of a nightmare.

but look, at the same time, let's have a look through this.

Right?

so the hi fi thing has come, come back.

it's like a clickable flow now.

So that's really neat.

You can actually now like unusual spike detected tap to investigate.

Cool.

So then you can maybe you could come through.

It's going to prompt you in a gamified way to figure out where the where your spikes coming from.

and look, this is really cool.

You can kind of automatically click through and get a real sense of kind of the ui.

I mean this is epic.

So Your usual is 22 minutes.

The oven's been on 47 minutes.

Someone's obviously cooking a big roast.

But.

wow.

Okay.

It's updated it in real time.

So this is crazy.

As you see, like you can be looking at something and it'll decide that it's hold my beer.

It's not done yet.

but look, this, this looks like he's nearly done now.

But okay, so there's some, even some animation in here now.

you get a nudge if one of the kids maybe hasn't checked in for three days.

That's really interesting.

And you can actually click through this whole kind of you like work workflow.

So there's still some stuff to go.

What's happened there?

I think we've broken up.

Okay, what's happened there then?

Okay, so we hit our usage limit again.

But that's, that's fine.

So look, I think we've got a good start.

So turn off the landing light.

I did it.

Wow.

Okay, cool.

This is epic.

You get a really, you know, that really gamified view.

how cool would this be?

Right?

Like quests, you kind of, you're getting like a real full, and you can change the colors now to we go to like a parent pack and also the kids view.

Right.

So maybe you can scroll, you can you can, you know, interact with other people.

You can go to your leagues, how you're doing versus Other families.

This is epic, right?

This is so cool.

Pocket money jar.

That's a really cool concept.

Right.

So what if you were like encouraging these behaviors in the kids and you maybe gave the, the saving on energy is their pocket money.

I don't know, maybe I'm being stingy there, but could be a way to incentivize some good behaviors.

but look, this is epic.

You can very quickly like okay, we're.

It's taken us about maybe 40 minutes to get to this point.

So you could easily do this sat around a table with a product team or with a client.

I think this is really powerful.

and that you've even got like the celebration ui.

This is just so good.

so let's go back to the the animation that was coming through which I think was here, right?

no, it's here.

okay, here we go.

This is a.

Maybe a.

Okay, I can see what it's trying to do.

Right.

But it's.

well that's cool.

Right.

Okay.

So like this is not perfect by any means.

and but you know compared to what this would have previously cost, like getting an animation team to do this.

Could you probably go backwards and forwards with this a bit and get to something that's pretty good and also like I really like the logo.

Give the bill payer a team.

What a tagline look.

So that's, that's all we're going to go through.

My, my summary is the slides better than anything kind of in a one shot that I've seen from any other tools.

So that's really exciting.

The design files again incredibly powerful.

and going on that journey from kind of lo fi up to what is kind of a fully working app I think is really incredible.

I think the animation probably some work to do and I don't think you could ever get to kind of a video style ad.

So I think there's better tools out, out there for, for that video piece.

but you know, all in all I'm really impressed.

I think you know what's clear is this is very credit hungry.

and you could seriously like if you were going backwards and forwards and you haven't really properly done a planning file, then I think this could become a real nightmare.

But you know how like what would have been possible with these kind of hi fi designs six months ago?

I'm kind of blown away by this.

so look, I hope, I hope you find that useful.

Definitely go and play around with the tool with yourself as we've said do your real good planning up front, and I think the credits will go a lot further.

and, yeah, let me know.

Let me know if you find these videos useful.

we're planning on doing some more kind of Live in Action, as new tools and features come out.

So, yeah, let me know what you think.