To AI or not to AI: A Founder’s Guide

In partnership with

Hire the highest quality overseas talent with Oceans.

Our offshore talent has more impressive resumes than your employees. Seriously.

Most outsourced firms can’t source the type of EA who can work with a $100M executive, but Oceans can. Our EAs take over all of your backend admin and operational work, so you can focus on what actually matters: scaling your business.

Work with the firm that sources the right-hand to Morning Brew, Beehiiv, and Magic Spoon.

Get full-time, highly experienced talent for only $3,000/month.

Learn more about Oceans

Hey Shifters!

Greg Isenberg recently made an interesting point about AI startups: while many are just ChatGPT wrapped in a good-looking UI, they're winning by building user habits and workflows that stick.

This got me thinking about how founders should approach AI in 2025.

I have a few friends who are currently getting their PhDs. All of them talk about how trash the job market is. They can either choose to work in academia and compete for a professor job, making next to nothing, or they can go ‘industry’, meaning that they take a job at a tech company or in consulting.

To help them and other students, I’ve been playing with an idea for an AI-powered resume generator for PhD students entering industry. Initially, I dismissed it – "They could just use ChatGPT!"

But that's exactly Greg's point: the opportunity isn't in having the best AI; it's in solving specific user problems in ways that fit their daily workflow.

AI is moving incredibly fast, and nearly every founder I talk to is wondering how or if they should use it. After spending the last year thinking about how we introduce AI at Chezie ) with AI at Chezie (note: we haven’t), I've developed some thoughts about when to use AI, when to skip it, and how to think about it strategically.

Should You Add AI to Your Product?

You may have to have an AI feature in your product right now, and you may not be wrong. But before you jump in, here are some things to consider.

Enterprise Pushback

In recent months, several enterprise customers have asked us if we use AI in our product because they don’t want their data being used without their knowledge. I won’t call it resistance, and it hasn’t slowed our sales cycles down (mainly because we don’t have any AI features yet), but it’s worth noting. If you sell to enterprises, be sure you’re using AI in a way that won’t bring up any red flags with their procurement/infosec teams.

Cost Considerations

AI ain’t free, and while costs seem marginal right now, founders need to consider the cost of scaling API calls to AI models.

In the same way you budget for hosting costs from AWS or Google Cloud, you have to do the same for the cost to tap into AI models, plus the cost of engineering resources that maintain the AI.

User Adoption

Think about how much people push back on companies using their data for ads that they didn’t ask for. That same pushback is coming for AI. Especially for founders operating in regulated spaces like healthcare or finance, you need to be exceedingly clear on how you use your user data.

To see how not to handle this, here’s a Reddit thread of one of my recent favorite tools Wispr Flow. It’s a voice-to-text app for Macbooks. It works exceedingly well, but people are dubious the ethics of the company.

The founder eventually replied in this thread to explain the company’s practices in detail, but it would’ve been better if he’d been upfront about all of this.

Note: this hasn’t stopped me from using Wispr, but I was hesitant after reading all of this.

The Key to Building with AI: Augment, Don't Replace

The best AI makes humans better at their jobs than trying to replace them entirely.

Think about Jasper versus ChatGPT. While ChatGPT can write content, Jasper succeeds because it's explicitly built to help content creators write better and faster - not to replace them. They understand that the best AI tools enhance human capabilities rather than eliminate the need for human input.

When building AI features, ask yourself: "How can this make my users better at what they do?" rather than "How can this do the job for them?"

This approach not only leads to better product adoption but also addresses many of the concerns above:

• Enterprises are more comfortable when AI augments rather than replaces

• Users are more likely to adopt tools that enhance their work than they are to adopt something that replaces them

• The value proposition is more straightforward when you're making people better at their jobs

What AI Means for Founders

Companies aren't investing billions in AI to make employees more efficient. They want to replace expensive knowledge workers with AI agents that don't need sleep, benefits, or time off.

The average knowledge worker – engineers, designers, marketers – will be replaced by AI in the next 5-10 years. But for builders and founders (that's you), this is a golden opportunity.

Why Builders Win

While AI can execute tasks, it can't (yet) build audiences, create compelling visions, or understand deep customer needs. These uniquely human abilities are only becoming more valuable.

Using AI in Your Daily Workflow

It should be part of your toolkit even if you're not building an AI product. Here's how I use it:

Rapid Prototyping

I use Vercel to turn product specs into working prototypes in minutes. I write out the use case, market alternatives, and desired changes in ChatGPT, then upload that to Vercel. Two minutes later, I had a prototype to share with our designer. This saves hours of back-and-forth and reduces design costs since our contractor (who’s paid hourly) can build exactly what we want.

Content Creation

AI won't write the perfect LinkedIn post, but it gets you 80-90%. Where it used to take me 30 minutes to write a LinkedIn post, I can now produce one in 10-15 minutes. Get AI to draft the post, add your personal touches to the opening and closing, and you're done.

The key is using AI to amplify your capabilities, not replace your judgment. Let it handle the repetitive work while you focus on strategy and creativity.

The Bottom Line

AI is moving fast, but don't let FOMO drive your product decisions. Adding AI features just because everyone else is doing it is a quick way to burn cash and confuse customers.

Focus on:

1. Building AI features only when they solve real customer problems

2. Using AI daily to amplify your capabilities and run lean

3. Identifying what AI can't do (yet): building genuine connections, understanding deep customer needs, and creating compelling visions

For us Equity Shifters, AI is a chance to level the playing field. When the barriers to the building are lower, what matters most is creativity, hustle, and deep customer understanding – areas where underrepresented founders often excel.

The future belongs to builders. And with AI, we can build faster, smarter, and more efficiently than ever before.

See you next week,

Toby