The more I play with AI and build things, the more I realize that experience is the only thing that really matters.

If you're automating just to use cool tools, no one really cares. What people care about is how it makes them feel.

This week I built Lead Lasso, a Pokemon Go style game for catching sales leads. I took it outside, played it with my family and my toddler, and ran around a park catching leads. It's not the first game I've built, but it's the first time I've built something this fun that still ties back to business.

It gave me a level of joy and excitement about building internal tools that I've never felt before.

No one cares what tools you used. They care how it made them feel.

BCG surveyed 398 sales leaders: 91% expect higher quotas or productivity targets in 2027, but only 8% say AI has improved quota attainment, and just 6% have AI running full workflows on its own (BCG, Oct 19). Two days later, IBM's CEO said some enterprise AI spending is "completely out of control" (24/7 Wall St.).

Bigger quotas are always going to happen. I've rarely seen a quota go down. There's always pressure for the big number to get bigger, and one of the ways teams are justifying it is AI.

Even though few leaders say AI has actually made their teams more productive, I think it's twofold. One, it can be harder to see. Two, people need to adjust how they're using it.

But if you don't have product-market fit and a strong product, no matter how much you automate, it's still bad.

A bad process is still bad, even if it's automated.

Pricing advisor Ulrik Lehrskov-Schmidt says about 1% of users often drive 50% to 75% of AI costs, and one company's worst customer spent 380x more in tokens than it paid. Figma added credits and fair-use limits: margins fell from 92% to 82%, but growth sped up past $1B ARR (SaaStr, Oct 23).

We love pointing out that 1% of users burn all the credits. But I've worked in insurance and health-adjacent industries for years, and that's always the case. Your super users are what cost all the money, and charging them separately has always been hard. If you could solve this, you could make $1 trillion.

Insurance has been working on this for years with premiums. We're seeing a version of it now with Claude and other tools: weekly limits, tiers, and paying for overage. But there still has to be perceived value.

Super users will always be super users, so controlling them is a top priority. But with a new tool, you often get a huge burst of excitement, and then it drops off. I think we'll see more effective triage, whether that's tools offering consultative help to improve your workflow or education to minimize the big burn. A lot of market forces are heading toward unique implementation to adjust costs, the same way the insurance industry does it.

Super users will always be super users.

eWebinar's Melissa Kwan realized prospects couldn't tell what her product did. She built an AI "brain" from 650+ customer quotes and rewrote her positioning. Time on site rose 50%, and she closed a $40K enterprise deal against a $2,650 average customer value (Growth Unhinged, Oct 21).

The biggest issue in building a product is thinking you know what your customers want. It takes a very confident founder to realize the customer knows best.

Startups, sales teams and business teams all face this. They think their product is cooler than it is, or that it has unique value customers just don't get, or that customers are stupid. That's never the case. And even if it were, you're not going to convince someone you're selling to that they're wrong. You never want to insult the customer.

That's why it matters so much to build in a way of fact-checking yourself: understanding what customers are actually saying so you know where your positioning really stands. AI is very effective at this. Even though AI is biased, it can remove a lot of your own personal bias when reading those messages. It's almost a blind experiment.

Listen to what customers actually say, not what you want them to think.

Every month I redo my website or some part of its design. This is the first month where the 3D actually felt pretty cool.

The models clearly know 3D has been a weak spot. Creating 3D assets and animations for websites has gotten dramatically better over the last month.

3D on the web finally feels worth building.

Two problems: lead assignment can be boring and feel unfair, and scavenger hunts are dated. In Lead Lasso, a manager draws a fence on Google Maps, drops leads inside it and shares a link. Reps walk the area with a hot and cold meter and catch leads through their phone camera. Built in a weekend or two with Claude and free 3D asset packs.

It's extremely fun, and probably one of the best games I've made, especially for sales.

Watch the Lead Lasso demo

Recruiters use AI to screen, so you have to use AI on your resume, which makes everyone look the same. Proof tailors your resume, makes you verify every AI change in a "proof bank," researches the company, and builds a fit brief: a project made for that company and role that you can send before or after the interview.

Your line from the video: "You're already spending hours preparing for the interview. Why not spend 30 minutes trying to get ahead of the curve?"

One thing I'm focusing on more and more in my work: AI is a tool. No one cares that you used AI. They care about the outcomes.

Proof is built around that idea. What can you show? How can you prove you can do the work for the job you're applying to? No matter where the world goes, that will always be valuable.

Watch the Proof demo

Juniply Challenge #001 is a few weeks away, and it's free to sign up. The brief drops November 13, with swag and cash prizes.

It's very simple. If you've never used these tools, or you want to learn AI tools like Claude and Lovable, I'll be holding office hours to teach you. You'll learn with a goal in mind.

This one is more innovation focused than GTM focused, so it's a good intro if you're new to building. Join with your team, or send it to someone who's been wanting to learn.

Join the waitlist

I'm testing tools, building strange ideas, and learning what works in public.

Explore Juniply

Follow the 52 Weeks of Building challenge on YouTube

If you know someone who is also trying to understand where modern go-to-market is headed, forward them this edition.

If you're interested in building at all, even if you're brand new, join the hackathon.

At the end of the day, all AI is doing is allowing humans to be more human.

So make sure you're a good human.

Until next time,

From Brendan