Quick question before we get into this:

Today’s edition is exactly related to the poll. I’m going to show you how to use AI to start a services business.

And if you like today’s edition, do reply so I’ll do more editions like this. 

Let’s get into it. 

Think about the last two decades, notice how every new capability follows the same curve.

A small group figures out the tool before the market understands what it's worth, charges old prices against new costs, then watches margins compress as everyone else catches up.

For example: In 2013, a solo developer who understood Shopify better than an average business owner could charge $5,000 to build them a store.

By 2018, Fiverr was full of people doing it for $200, so that window lasted roughly four years for that small group.

Now my student can build one for herself during a 60-minute class without much effort at all.

Right now, AI design tools are at the 2013 Shopify moment.

The cost of producing a professional website has collapsed, and the market hasn't adapted yet.

That's the gap I find worth exploring today.

What happened

On March 18, Google shipped Stitch 2.0, a full rebuild of their AI design tool.

Figma's stock dropped ~10% the same day (coincidence?) 

You can take a screenshot of any existing website as input, give Stitch a brief about how the site should feel, and it returns a complete redesign in under a minute. 

An infinite feed of fully-formed design directions generates automatically as you go, different layouts, aesthetics, and visual languages, without writing another word.

What these two tools have done, together, is collapse the cost of producing a professional website redesign.

And three months in, barely anyone outside of design Twitter has heard of either.

The market they cracked open

There are yet a ton of small businesses that are using legacy designs (most in the area you’re staying right now).

The average expected spend on redesigning their website would be $1,000 - $4,000.

Agencies are quoting exactly that, sometimes three times more. And their timelines run three to six months.

All of this is still important, You can't really surpass it but you can accelerate it.

Earlier, these were the starting points before anything was mocked up. Now the mock up is easier and hence these conversations can be sped up.

But still, the average small business owner can’t be bothered.

He continues to live with the dated website for another year, and as a result, loses leads to competitors whose sites look like 2026.

This is the gap.

Millions of businesses with a known problem but no affordable solution, until now.

What I did

I needed phone repair shops in Bombay during my recent visit as I dropped my phone and cracked the screen.

I visited the one with the highest ratings despite their website being this:

The glass was shattered in the camera bump and body (connected).

The hero section had nothing to do with the fact that they were in Bombay. And they had a form to get leads…

I uploaded it on Stitch with the brief: redesign this to look clean, modern, and trustworthy for a phone repair business.

Output:

Then I ran the same site through Variant with a slightly different brief, asking it to communicate 24-hour turnaround and urgency alongside the trust signal.

The total time from a screenshot to something I could show a potential client was 20 minutes.

And I am not even a designer. (I know my way around Canva, though)

This becomes your foot in the door.

The math

Do a few projects a month and you’ve got yourself a well oiled machine. (Getting that first paying customer is a whole other topic I may cover in a future edition)

The best niches right now: home services (plumbers, HVAC, landscapers), dental and physio clinics, solo lawyers and accountants.

Pick any city. Google "[dentist / plumber / accountant] [city]."

Open the first ten results.

At minimum, six will have websites that you would cringe looking at.

I was applying for my daughter’s admission and found some of the best schools in my city. But they had such terrible websites, it initially made me second-guess those options.

That's how fast trust erodes.

Why this window closes

One technical caveat worth knowing: don't publish directly from raw HTML output. 

Run everything through Framer as the delivery layer, as it would handle server-side rendering properly.

This week

Find one business with a bad website in a niche you understand. 

Run their homepage through Stitch / Variant. 

Send the owner the redesign with no pitch: "I redesigned your homepage, here’s how it looks. Would you be interested in me redoing the rest of it?"

Reply and tell me what comes back, and if you convert them :)

Nandini



Keep Reading