✍️ Quick Note

Before we dig in, I have a quick favour to ask you.

Each week I spend several hours crafting this email entirely for free. It’d mean the world to me if you could share the Voyager with just 1 person who you think would also enjoy it.

It’s super easy, and will only take 30 seconds. You can just forward this email or share this link 🙏 🙏 .

Voyager’s still young and almost 200 early subscribers, all from word-of-mouth, and all designers (+ my mom). Your contribution matters.

Cheers :)

Turns out the last newsletter I sent had too many things, and some didn’t even discover the last section. Those who did, found it somewhat confusing. Rest didn’t reply with any feedback (I’m eyeing you…👁️👁️)

So today I’m going to keep it concise, & firstly, going over that last section again. Here’s the bread, butter, drama —

• 🥪 Interface layer: The next big AI-UX trend

• 🔮 AI-UX First Principles

• 😡 WTF is “don’t just build a wrapper”

• 🤣 Meme of the day

Read time: 1338 words, approx. 5 minutes

With AI, we’re living in world of promises, that’s super clear. What’s the future promise this time?

Interface layer: The next big AI-UX trend

Up until now, the software apps and websites commoditized the operations layer — apps like Uber or Deliveroo made it easy for anyone to use the services without worrying about the on-the-ground logistics. This was the service layer.

1/ Operations layer - the on-ground logistics.

2/ Service layer - commoditized operations layers by providing sit-at-home apps like Uber & Deliveroo for easy access to services (logistics).

We’ve been surrounded by service-layer-apps. Times are changing though - what’s the AI forecast?

3/ Interface Layer. Predicted to commoditize the service layer (coined by Scott Belsky). It integrates multiple service apps together to provide one seamless experience. Instead of navigating apps to get a task done, the users demand what they need, and actions are taken on their behalf.

It’s about creating an experience around user’s intent.

Enough theory, let’s look at an example

An enterprise level example I came across recently was, let’s say you have a table with 50 columns and 10000 rows. You need to filter out column 3, 5, 6, and 24. Instead of designing an experience to make it intuitive for user to quickly do that, you design an experience where the user doesn’t have to do that at all.

Instead, the questions you ask are —

• What was the user trying to get from the columns?

• What was the user trying to interpret from the table’s data?

• When did the user come to see the table?

• What problem was the user trying to resolve?

• Can I just show the interpretation of the table’s data in the previous step, when the user faced the problem?

It’s still vague, but do you see how the questioning has changed? Designing experiences for intent means taking a step back and looking at the complete journey. Outside of drawing rectangles and aligning pixels, focusing on getting the solution faster to the user, not how the user can get the solution faster.

Read that again, there’s a difference.

Another example.

Let’s look at it from a day-to-day consumer angle —

We used to go to the library to research (unless you were born after the Instagram era, then you probably browse Tiktok and won’t know what a library is - a physical place with books and a ‘Shhhhhhh’ sign)

Now we Google and search the world wide web.

Arc is a “level 1” Interface layer app — you don’t have to search yourself, you just tell what you’re looking for and the app gets you the answer compiling the best + relevant results.

Another? Another.

Back in the 90s, people flexed their rizz (gen-z for charisma) to pick up lovers at the bar or house parties. The OG dating scene.

Fast-forward to day, we now have Tinder/Bumble. The love-of-your-life is just one swipe away.

The future? A personal GPT wingman.

Imagine a world where there’s a cloned version of you as a GPT that knows your interests, attachment style, what you’re going through right now, your hobbies etc. and you get a pre-filtered list of potential people that you can swipe on (because we all know you’d still want to “check them out”).

Swipe right, Match! A date is set, synced with both your calendars. You can now meet and vibe the old-fashioned style (or suffer through an awkward date when you shake her hand…as she was going for the bread)

Don’t listen to me, Bumble CEO said the same thing (FYI: I called it first 😎)

Oh baby we’re on fire! Another one.

Enter into the scene “AI wearables”.

We looked at the top 8 AI wearables last time, each with a different mission and a different execution.

They are taking it a step further - merging all your daily tasks into one seamless conversation (text or voice).

Order an Uber without specifying your location. Auto-text mom, because you always do before sitting in a stranger’s car. Remembered your fridge is empty? Your go-to kebab's on its way home before you are..

Everything done in 2 mins, all in one sentence.

puff, hard day at work huh.

I like to call this particular example “Screenless UX”, a user experience where screens are non-existent.

It’s all part of the next big AI-UX trend: interface layer. I believe in it, seems like the obvious route.

Do you think otherwise? I’d love to know — hit (reply) me up :)

Instead of advertising about the release of new ai-ux interactions, I thought I’d make it more meaningful.

🔮 AI-UX First Principles

We just added a new [WIP] guide to aiverse! It’s a set of principles & questions for when you're designing for AI.

#1 Principle — Don’t use AI at first.

#2 Principle — Break the problem into smaller problem solving loops.

#3 Principle — Determine the altitude of AI infusion in your UX.

More principles on the website linked to their relevant case study.

We even added a couple of questions to ask yourself when thinking about what AI features are possible.

Btw, if you’d like to contribute to the website in any way, hit me up :)

😡 WTF is “don’t just build a wrapper”

To whom is may concern (whom: those calling AI startups a GPT wrapper).

I don’t get the “don’t build a wrapper” movement. All the people I’ve seen say this are the ones sitting on the sidelines playing with a chocolate wrapper.

You can still win even if it’s just a “wrapper”.

AI is like water — ubiquitous, necessary, same in every bottle.

Water is freely available. You can drink it straight out of the tap. It’s everywhere.

Yet, you still buy branded water bottles selling “filtered” tap water. At the end of the day, it’s just positioning & distribution. Liquid Death is an example, it’s water but for non-alcohol drinkers to look cool at parties🤘 

How I see it is that there are 3 levels when building with AI —

• Level 1: Using a model to solve a niche problem (what people call “wrappers”)

• Level 2: Code-editors, multi-modal agents

• Level 3: Humanoid robots

Maybe you are building at level 1. Maybe that’s what gets you going, and that’s okay. You can still win. You’re just playing in a different league. There are players in each league, there’s intense competition in each and there’s money to be made in each as well. It’s all fun & games before AGI anyways.

Your wrapper WON’T work though —

• If it doesn’t pass the AI litmus test. Take out AI out of your pitch deck, is it still a good business?

• If the base model got better/cheaper/faster and added functionalities, will your product get better too?

• If you don’t have a strong distribution. Brand. Messaging. Marketing. GTM. Pricing. Customer service.

(Data + Model) x UX x (Distribution + Perceived Value to Customers) = Your new AI MVP

UX and Distribution are key.

OpenAI’s got the model, I’ve got the UX covered :)

🤣 Meme of the day