A pattern I've noticed in AI Native Design

Hey, I’m Shane.

Welcome to Design for Builders, a newsletter for founders, operators, and builders who want to improve their design skills and create better products.

I know I have been typically using this newsletter list to talk about large-scale design trends, news, and updates, but I just want to talk a little bit about my experience here today.

What if the following was true

AI Didn't Add Features, It Deleted Navigation?

Your best users should barely use your product each day.

They should accomplish 10x more while spending 90% less time in your app. They should forget your interface exists.

This sounds insane. But after 6 months deep in the trenches with AI-native startups, I've seen the future: software that builds itself around user intent, eliminating navigation entirely.

Products where the interface doesn't exist until you need it.

Let me show you what I mean.

In Traditional SaaS Navigation

Every SaaS product today is a Swiss Army knife. You pay for 20 features, use 2, but carry all 20 everywhere. Salesforce has 300+ features. Your team uses 12. But everyone navigates through the same bloated interface, hunting for what they need.

We've accepted this as normal. We even have metrics for it: feature adoption rates, time to value, user engagement. The entire SaaS playbook is built on getting users to navigate MORE of your product.

Here's the dirty secret: SaaS companies are incentivized to keep you trapped in their interfaces. They dress it up as "powerful functionality," but it's just complexity that justifies the price tag.

An accountant using traditional software has their ritual: spreadsheet in one tab, government filing sites in another, client emails in a third, document storage in a fourth. When a client emails asking, "How much should I set aside for taxes?" the dance begins:

Open spreadsheet tab

Find last year's files (2 minutes of clicking through folders)

Navigate to tax calculator

Check current regulations (new browser tab)

Calculate scenarios (manual data entry)

Format visualizations (wrestling with chart settings)

Draft response (switch to email, copy-paste numbers)

45 minutes. For one question.

The accountant knows this dance by heart. They've internalized the software's limitations. They don't even question it anymore.

The Revolution: Interfaces That Don't Exist

Now the AI-native version. Same accountant, same client email. But instead of navigation, one action: answer the email in their AI workspace.

What happens next breaks every UX principle we know:

The AI doesn't open a pre-built tax calculator. It CREATES one. Specifically for this client, for this question, at this moment. It generates a complete custom interface:

Pulls 3 years of client financial history

Calculates scenarios based on current run rate

Factors in state-specific regulations that just changed

Shows its math like a proof every calculation transparent

Generates an interactive visualization the client can adjust

Drafts the email with the right tone for THIS client

Suggests setting up automated quarterly tax savings

Five minutes. No navigation. No hunting through menus.

The AI didn't automate the workflow. It invented it.

The Paradox That Breaks Every Metric

Success now means users spend LESS time in your product, not more.

Traditional metrics are suddenly toxic:

Time in app? Wrong you measure time saved

Feature adoption? Wrong features don't exist until needed

Engagement rate? Wrong invisibility is the goal

Imagine pitching investors: "Our best users barely open our product. They accomplish twice as much in 1/80th the time. Our engagement metrics are plummeting and that's exactly what we want."

Maybe this is an existential crisis for SaaS. I don’t know.

What Happens to Designers and PMs?

Let's address the elephant: if AI generates interfaces, what happens to the people who design them?

They don't become obsolete. They become LEGO designers.

Instead of crafting screens, they're building the building blocks AI uses to CREATE screens.

This is actually harder and more creative than traditional design.

Take a smart spreadsheet component. Old way: design rows, columns, formatting options. New way: design a component that:

Understands what data to pull without being told

Knows when to pivot itself into a chart

Connects to email components to share relevant excerpts

Spawns calculation modules on demand

Explains its formulas in plain English

You're creating intelligent building blocks that combine in infinite ways.

The Three Stages Every User Goes Through

Watch someone use truly AI-native software for the first time:

Stage 1: Confusion

"Where's the dashboard?"

"How do I find reports?"

"There's no menu? How do I navigate?"

They're looking for the Swiss Army knife. It's not there.

Stage 2: Revelation

They accomplish something complex in one prompt.

"Holy shit. I just did my weekly reporting in 2 minutes."

"Wait, it built a custom workflow just for this?"

The penny drops. They've been trained to think in software limitations.

Stage 3: Liberation

They stop thinking about the tool entirely.

No more "How do I do X in this software?"

Just "I need Y to happen."

The interface becomes invisible. Only outcomes matter.

The Transition Phase

Let's be real: most companies are still bolting chatbots onto existing products. That's not what we're talking about. Although it’’s not entirely incorrect.

Phase 1: Parallel Paths )

Keep your traditional navigation but add AI shortcuts. Users can navigate OR command. Let them discover that commanding is faster. Perplexity does this well you can click through categories or just ask.

Phase 2: AI-First, Navigation Backup

Flip the default. AI command bar becomes primary, navigation becomes secondary. Show users what's possible through AI, but keep the safety net.

Phase 3: Generative Only

Navigation disappears entirely. The AI generates everything. This is where the real magic happens, but only after users trust the system.

The New Competitive Moats: The 3Ds

When anyone can prompt ChatGPT to build features, traditional moats evaporate. But three new ones emerge:

1. Data Compounds

Every interaction teaches your AI about domain patterns. The accountant AI doesn't just know tax law. It knows:

How THIS type of business typically structures finances

Which deductions get flagged in THIS state

What questions THIS industry asks next

Your AI gets smarter about your specific domain with every user. That compounds. A new competitor can copy your features but not your accumulated pattern recognition.

2. Design Becomes Aspiration

Emotional, harmonic, beautiful and intuitive. Users don't want accounting software. They want to be the accountant who handles twice the clients with half the stress.

It's crafting the narrative of who users become.

3. Distribution Lock-in

Own channels others can't touch:

Government API access for compliance

Banking integrations for real-time data

Professional association partnerships

Embedded positions in platforms

The AI accounting tool that files directly with the IRS has a moat that better AI can't cross.

The Future: Software That Disappears

The companies that win this decade won't have the best AI. They'll be the ones brave enough to let AI redesign what a product is.

Not software with AI features. Software that doesn't exist until you need it, then builds itself perfectly for that moment.

The accountant who forwards an email and gets back a complete solution. The designer who describes an outcome and watches the right interface materialize. The founder who states a goal and sees the path appear.

Navigation is dead. Menus are dead. Features are dead.

What remains: intent, outcome, and software that shapes itself to serve both.

That's not just a better product. It's a fundamental redefinition of what software is.

The builders who understand this won't just win markets. They'll create entirely new categories while traditional SaaS companies desperately add chatbots to their Swiss Army knives, not realizing the entire concept of carrying tools is obsolete.

Your users don't want to use your product. They want to become the person who effortlessly accomplishes what once felt impossible.

Build that. Everything else is already obsolete.

What do you think? Are you building for navigation or outcomes? Reply and let me know what you're seeing in your products.

-Shane

P.S. - If you're a founder trying to make this transition, I'd love to see what you're building. The best examples I've seen are coming from unexpected places accounting, legal, healthcare. The "boring" industries are actually where this revolution hits hardest.