Why Customer Interviews Miss the Biggest Opportunities

Hello Predictable Revenue community,

Book update: we’re at 240 pre-orders, which is a great number, but wouldn’t it be better if it was a nice round 250? I need your help, the book launches on the 28th (Tuesday!) and if you’ve been on the fence about pre-ordering, now’s your last chance to grab the pre-order bonuses that have taken me months to build. So here’s my final hard press. 

This book will help you strengthen your product market fit, turn customer dev interviews into customers, and figure out how/when to build yourself out of the sales role. My goal was to draw a line between The Mom Test and Predictable Revenue. Isn’t that worth $20?

We have a paper book made with 100% real paper (I’m pretty sure), an audiobook recorded by yours truly, and a kindle version made with 100% real electrons. 

On top of that, you’ll get access to the 10 pre-order benefits I’ve built over the last few months (screenshot at the end of the post). 

I'll be honest - I didn't expect the response to these bonuses. Founders have been telling me the VC Investor List alone is worth the price of pre-ordering. The Founder Sales Guide has become the most downloaded resource (it's essentially my entire methodology in one place). And the Monthly Fireside Chats? Those are always a highlight of my month. 

But here's the thing: these bonuses end when the book launches. After that, you get the book. Period.

I'm not doing this to be difficult - I'm doing it because the people who believe in the work before it's "proven" deserve something extra. You took a bet on me and this book when it was just an idea. These 10 resources are my way of saying thank you.

If you've been on the fence, this is decision time. Lucy loved it and you will too. 

Source: Library Journal (gated content)

Onto the email… 

I can't spell HTML.

That's not self-deprecation - it's just true. I'm not an engineer. But right now, I'm elbow-deep in Windsurf, n8n workflows, MCP servers, and AI coding tools, building automations like my startup depends on it.

Because it does.

Every founder knows the playbook: interview customers, validate ideas, build what the market asks for. I followed it perfectly for CARB.IO. Talked to SDRs, managers, industry experts. Asked all the right questions.

Nobody - not a single person - mentioned the problem that became our entire competitive advantage.

I only found it because I stopped asking and started doing the work myself.

And if you're a founder trying to avoid getting disrupted in the next 24 months, you probably need to do the same.

The Pattern I Keep Seeing

Here's what happened when I started CARB.IO. I didn't start as the CEO with a vision deck. I started as an IC doing manual outbound work, trying to figure out how the hell to get meetings booked.

I got into manual mode because there weren't any tools available that could serve the process. And I had just burned my engineers out building a CRM that nobody wanted, so I couldn't ask them to build anything else. I could feel something interesting but I still spent a good year using manual tools, figuring things out, trying to see where the gaps were.

Taking on the process myself helped me see that cold email would evolve into tools to help you send more emails. But here's what my non-DIY competitors (who have done much better than I have) missed: SDRs would need help managing the conversation after someone replies (not just sending a ton of emails).

This is one of the reasons CARB did so well. Managers and users loved that we were solving two problems: sending emails AND what happens after a prospect replies. Our bigger, better-funded competitors were just helping companies blast tons of email.

Our insight into post-reply being critical shaped the product and is one of the reasons we grew so fast. (We eventually failed for technical prioritization reasons that I've documented well were my fault, but that's another story.)

Here's the key thing: I'm confident I wouldn't have seen this without doing it manually.

The post-reply problem came up in literally none of the customer development interviews. I interviewed SDRs, managers, experts in the space. Nobody "saw" this problem.

But I did. Because I was on the tools, doing the work of today, and thinking about what the knock-on effects would be.

If you're sending a ton of email and getting a ton of replies, the knock-on effect is that people are going to forget to follow up on some of those emails. I saw it because, despite my beautiful Gmail labels-based sales dev process, our team (myself included) were still letting things slip through the cracks.

Every single time I've followed this pattern (going back to IC work, getting hands-on with the bleeding edge) I've found something interesting to start.

So when I started feeling that itch again recently, I didn't ignore it. I went back in.

The Shift Nobody's Talking About

Ten years ago, there was a massive efficiency unlock happening. People were moving from manual spreadsheets and paper processes to basic software. Being able to click around in a CRM instead of managing a filing cabinet? Revolutionary.

That benefit isn't novel anymore.

In fact, clicking around in software is starting to feel... annoying. I noticed it in my own workflow. After a sales call, I'd go to send my client a recap. I'd copy the transcript, paste it into ChatGPT (even though I have a GPT project set up for this), wait for it to format, then copy it back out.

And I thought: I don't want to do this annoying work.

That friction - that "I don't want to do this" feeling - is the signal.

Because here's what I'm seeing: People are going to expect AI agents to do the work for them. Not help them click faster. Not organize their clicks better. Actually do the work.

The question isn't whether this shift is coming. It's whether you're positioned to build for it or get disrupted by it.

Living in Tomorrow

Right now, I'm using AI coding tools to build agent workflows in n8n. I'm automating call recaps. I'm building sales agents that update next actions. I'm training these systems to handle the work I used to do manually.

But here's where it gets interesting.

Step one is using an AI prompt to generate the to-do list. But if I can get that prompt to extract the work clearly enough, then when computer use agents get good enough (and they're getting close), I'll be able to use my weekly call to program the work.

Think about that. Not "take notes about the work." Not "create tickets for the work." Actually describe what needs to happen and have agents build it.

This isn't science fiction. This is 12-18 months away.

Building for the Knock-On Effects

Here's the framework that's working for me:

Start with a change happening today. AI agents are getting exponentially better at understanding and executing complex workflows.

Think about the secondary and tertiary effects:

• How will consumer expectations shift? (They'll expect instant onboarding, zero manual setup)

• How will business expectations evolve? (They'll expect your system to import their existing setup automatically)

• What will become the new normal? ("Why can't your tool just clone my current workflow?")

This paints a version of the future for yourself.

Then ask: "If all these are true, what gaps will be created in my market that I'm well positioned or excited to solve?"

For me, that gap is building tools that leverage AI agents to eliminate manual work. More coming on this in a later email…

The Climber Mindset

I was listening to Chapter 4 of Dirtbag Billionaire recently, and something clicked. The author talked about not being a planner but being a climber - seeing the next wall and working out the right path as you go.

You can't plan for a future you haven't experienced. You have to climb into it and feel your way forward.

That's what going back to IC work gives you. You can't strategize your way into understanding AI agents. You have to use them. Break them. Train them. Watch them fail and figure out why.

Then you extrapolate forward: If agents can do this now, what becomes possible in six months? In a year? What problems will exist in that world?

And you start building for those conditions today.

The Work

So here's what I'm actually doing:

I'm spending time every week building with Windsurf, n8n, and AI coding tools. I'm automating my own workflows. I'm training agents to handle the annoying work I don't want to do anymore.

And I'm documenting the journey - not because it's sexy content, but because the act of building is giving me the insights I need to see what's coming next.

You have to live in tomorrow to build for tomorrow.

What version of the future are you building for? Hit reply and let me know - I read every response.

Collin

PS - if you thought I wasn’t going to ask you to buy the book one more time, you were wrong.

I am once again asking for… you to stop overthinking this and just pre-order the book.

Please.

Here’s the link.

PPS - as promised - here’s a screenshot of the whole pre-order benefits list: