The two-minute value test

Later today I’ll have two minutes to explain an AI workflow that took months of experiments, failed runs and head-scratching decisions to build.

I’ll be one of seven speakers at an Exit Five live event, with more than 1,000 people registered. Each of us gets two minutes to walk the audience through an AI agent build. If I make it through to the next round, I’ll get an additional four minutes to go into more specifics and share a case study.

Have you ever tried to start a two-minute clock and explain something somewhat complex?

Try it now. Pretend you have to explain what your product or service does to someone who has no idea about it.

What you’ll find out is that two minutes is the perfect amount of time to force you to be effective and efficient with your words. One minute is barely a few breaths: you can give a pithy pitch, but you won’t even consider adding backstory or context to it. Five minutes is an actual demo. But two minutes—that’s the sweet spot that pushes you to condense your pitch into 1) what you’re talking about, 2) what’s in it for your audience and 3) the next step they can take.

That’s it. The features, architecture, systems and tech have no room to exist in your two minutes, which makes you focus and prioritize.

And this is just the first big realization. Once you start creating the slides and designing the two-minute speech flow, you then start questioning more and more stuff. For example, what does the audience need to understand before I get into any of the tech or operational details? That’s a whole other ball game.

Just to give you some context on what I’ll be presenting in those two minutes, it’s an agentic workflow that takes your knowledge base, uses it to test a piece of copy with synthetic or AI personas grounded in what you know about your ICP, and finally gives you a report of the findings and insights you can use to decide what deserves further testing and validation with a human audience.

Showing everything under the hood would have been an attractive proposition. I could have tried to prove how sophisticated I am, how much effort I’ve put into building this system, and how much of a freaking tech genius I am (not really).

Problem is, a junior marketer doesn’t give a shit about which models I use, whether my agent is open source or not, or what API I’m using. They only want to get their job done better and faster.

And this is also why it’s hard to decide what to cut from your product demos, pitches or really any copy. You feel like including the features, the specs, and signaling how much you’ve worked on it all, IS the product. But if it weren’t for the value your product delivers, there wouldn’t be anything really to sell.

In my case, for example, when I had to crystallize the main big idea behind my workflow, I started going into how fast this copy-testing method can be and how scalable it is. But in reality, I had to focus on a way better key idea:

And that became the core of my whole two minute speech. It also kept me focused around that narrative, so that even when I got to talking about how we build our “project brain,” for example, I could always go back to why we built it, what that has to do with understanding your customers better and how it helps the audience.

That’s why I think you should try a little experiment any time you’re prepping for a new pitch or launch: the two-minute value test.

Here’s a quick framework to use:

Situation → Change → Reason to believe → Receipt and boundary

• Situation: Who is trying to do what, under which constraint?
• Change: What do you simplify, improve or make possible?
• Reason to believe: What minimum viable mechanism makes that change credible?
• Receipt and boundary: What can the audience see for themselves, and what should they not infer?

Want the team version?

• Ask product, marketing, sales and a founder or subject-matter expert to prepare their two-minute version separately.
• Do not let them compare drafts first.
• Transcribe each version and highlight the buyer, problem, value, proof, mechanism and next step.
• Compare what each person led with and what each person omitted.
• Do not average the versions into committee sludge. Use the disagreement to identify the positioning decision your team needs to make.

Now, if I get to the final four-minute round (my inner nerd hopes so), I’ll still show the stack, albeit a simplified and streamlined version. You need to show rather than just tell, even briefly, if you want to earn credibility and offer something practical to implement. But the whole architecture of the presentation will now follow a compelling value story, mostly because the first two minutes forced me to create it.

That two-minute version is the actual product; the extra four minutes are just the proof that I actually did stuff and can help you implement it.

Talk about the importance of your first two minutes in any pitch!

P.S. I’d absolutely love it if you wanted to join me—and, yes, vote for me—later today. Hope to see you there.

Discovery

​How I Use AI Agents to Run My Messaging Studio​

For this video, I deliberately went in the opposite direction of concision. YouTube is my messy workbench for showing the thinking process, so this 53-minute walkthrough goes into the nitty-gritty: the context layers, plans, agent handoffs, failures, custom review tools and live examples behind how I run my messaging and copywriting studio.

​

​Speed matters: Why working quickly is more important than it seems​

James Somers argues that speed does more than help you finish more work. It lowers the perceived cost of trying, which means you try more often and get more chances to improve.

Don’t rush to your pitch, but find a tighter constraint to reduce the distance between thinking, trying, getting feedback and trying again.

​Compound Writing: The Ultimate Guide​

The team at Every just published an excellent guide to writing with AI. It covers the mental models and best practices any writer, marketer or knowledge worker should internalize to get better results.

I especially like that it treats the back-and-forth with AI as part of writing rather than an inconvenience you should automate away. You can use the system for everything from interviewing and outlining to feedback, while staying close enough to decide what the piece actually needs.

Resonance