How to win the upcoming AI SaaS revolution

👋🏼 hello again! Welcome back to the ProductFoundry newsletter.

I didn't intend to write on AI again, but ChatGPT and other generative AIs still capture everyone's attention.

This week I'm sharing the four ways you're seeing AI come to the market:

Two of which you should be thinking about
One of which is a good starting place
And one to flat-out avoid

Four Types of AI Adoption (and which to chose to win)

Those who win will be the ones who are innovating with AI.

Simply rehashing other people's work or shoving ChatGPT into their apps with a new shell isn't a winning strategy. Those who integrate it deeply into their product or invert their product experience entirely will stand out.

👎 Shallow Application

Much of what we're seeing today are shallow applications of AI—often regurgitating what's available via the base API. We're also seeing a lot of gimmicks, surface-level fun, but with no lasting substance. (aka Candy)

Avoid spending time here. This is flash-in-the-pan stuff and won't lead to lasting outcomes.

Examples:

All the ChatGPT apps that made it into the App Store
Anyone just adding ChatGPT to their apps as a help system (there is an interesting path for help systems, but this isn't it).

It should be obvious. Don't waste your time with shallow applications. You're not learning anything, and nothing you're doing will last.

👌 Fun Experiments

Lots of innovation comes from experimenting. And having fun captures the public imagination on what's possible.

But, the fun can turn into shallow applications fast.

So watch out. Experimenting with AI is excellent; doing it in your existing product can make users feel it's shallow. Most early experiments are best left as side projects.

Examples:

All the fun avatar generators
Twitter bots (it's more fun to call people out on their use)
Most wins here are fleeting at best.

Most wins here are fleeting at best, but they can be an excellent way to inspire you for something meatier.

🤩 Augmenting People

Augmenting your customers is where people start winning with AI and half the reason I'm so bullish on the tech we now have. Because people buy software for three reasons:

To make money
To save money
To reduce pain and frustration

And Augmenting people can do all three!

Think about this.

You're a busy Customer Success rep. You've got a portfolio of customers, and you need to be nurturing your relationship with them. This can be very time-consuming.

But now, enter your AI assistant built into your CRM.

This AI assistant can help you write letters (using the backlog of emails it has with them as context) to help prompt their next step. It can use user analytics to spot anomalies so you can reach out before a customer churns. And so much more.

You've now helped people make money (the next step could be an expansion), save money (stopping the churn), and be more productive (reducing pain).

Cha-ching.

Examples

HubSpot is now doing augmented emails. They're helpful as a productivity hack, but they can go far deeper, and I'm sure they will.
Apps that do sentiment analysis, sure, a human can do it. AI can scream through it.
The same is true for transcription services and text-to-speech.
Levelsio's work with interior design is another example.
- But pay attention to the audience. If an interior designer is using it, then it's augmenting them.
- If it replaces the designer, that's the following approach.
Apps that help authors write books, make suggestions on plots, etc.
Grammarly is another excellent example. It's your "editor in a box."

🤯 Inverting Systems Entirely

If you want to win, find a way to be here where AI does the work for you. This is where everyone is afraid it'll take away people's jobs.

Yes. It will.

But with every new tech, we see a window of time where the world has to adapt, but after, new jobs arise. We're about to be there again.

When AI inverts the entire solution, and at best, a person has to curate and tune it, you're helping someone make money, save money, and reduce pain on steroids.

Now we're talking 10x cha-ching.

Here's an example I want to see.

Bubble's a very popular no-code solution. But its database system is confusing at af. I was somewhat shocked at how it works, given how popular Bubble is with people. Then I discovered everyone finds it confusing.

But what if you've yet to deal with data outside of a spreadsheet and barely at that? Couldn't AI do it for you?

Yes, it can! Just say what you want. "I need to track people. I want to keep the name and email of each person. I want to keep track of the emails I send to them. I want…" And that's it. You never have to see your database. It just gives you all the hooks into your widgets and more.

ChatGPT's context system already solves "the what" someone is saying. It can write code, so we're 90% there.

Other Examples

Analysis systems that do the work for you.
1. Many "boring" enterprise business systems require human analysts to do work that is more suited for AI. This is where a significant revolution is about to happen.
Again an interior design system (but now give it a budget and have it provide the list of furnishings and create the order for you).
An AI website agency. DALL-E meets ChatGPT meets Carrd, anyone? Just give it your company name, logo, and a description of what you do. Next, toss in some product pictures and descriptions. And out comes an optimized, stylish eCommerce site, complete with SEO pages and more.

I could go on.

Nearly every digital agency is going to see AI augment or transform it. We're in for some fantastic disruption.