I’ve demoed on stage in front of thousands of people. It is scary and fun in exactly the same breath, and the honest math on it looks something like this: for every minute of mainstage time, you’re logging 10 days of prep, and that’s the average. Some demos run hotter than that. Demo built and rehearsed, talk track memorized to the point you can jump in at any point and carry on, contingencies built for all the things that can and will go wrong, the staging environment resetting itself at the worst possible moment, the feature that worked flawlessly in the run-through deciding, live, in front of a few thousand strangers, that today is not its day.

I have seen every possible thing go wrong, from losing venue internet right before the keynote starts and ending up running via a tethered mobile connection (hello TechEd 2013) through to that perfectly reliable execution process failing on the live run.

That’s where console comfort saves you. Not the script, and not the slide with the perfect talk track. The muscle memory of actually knowing the product, well enough that when it breaks, you can route around it live, narrate through it, or pivot to a different flow without the audience ever feeling like you did that. You can’t fake that from a spec doc. You either have it or you’re praying to your deity of choice.

Let’s talk about why more PMMs need to be the best demo people in the company.

The credibility tax

A pattern I’ve watched play out in more companies than I care to count, is a PMM who can write a killer positioning doc, nails the MSD, builds a beautiful narrative arc for the launch, kickass presentation … and then hands the actual demo off to an SE. Hashtag fail.

What they have done is self-identify as the “slide person”, and once “the slide person” label sticks, it doesn’t stay contained to demos. It bleeds into everything else that PMM says. Their positioning input gets treated as theoretical. Their objection handling advice gets filtered through “well, do they know what happens when a customer tries that live?” The credibility tax isn’t fair, and it isn’t announced, but it’s real, and it compounds.

PMM is supposed to be the connective tissue between Product, Sales, and Marketing. Connective tissue that’s never touched the product doesn’t hold much together.

Confidence is a byproduct, not the goal

Something changes in a PMM once they can walk into a room, any room, any time, and drive the console themselves. It’s about what happens to your posture once you’re no longer dependent on someone else to prove your product works.

I’ve seen early-career PMMs transform after this, not because they read a better book on storytelling, but because they stopped needing permission to demonstrate value, and because they have the swagger of knowing how shit works. That’s worth more than another framework.

The console is Zero-Trust in action

I talk a lot about Zero-Trust fact-checking when it comes to AI output: never publish a claim without verifying it yourself. Turns out the same discipline applies to your own product.

The spec is somebody’s summary of what should be true. The MSD is your best articulation of what you want to be true. The slides are what looked true in the last review. None of those are the actual product. They’re all abstractions, and abstractions drift from reality the moment nobody’s checking.

The console reality doesn’t drift. It’s the ground truth. If you’re only ever working from what’s written down, you’re trusting somebody else’s word for what your own product does. Ask me how I know that particular truism over a beer one day.

You become a better storyteller AND a better critic

Here’s the two-way trade. Once you’ve felt the friction yourself … the confusing click, the load spinner that runs a beat too long, the feature that’s 80% built and everyone’s been quietly saying is “basically done” … two things happen at once.

Your stories get sharper, because you’re not repeating a customer’s complaint secondhand, you’re describing something you lived. And your product feedback gets taken seriously, because you’re not relaying feedback, you’re reporting firsthand experience. “A customer mentioned friction in onboarding” gets a nod. “I hit this exact wall myself, here’s the click that broke it” gets a ticket.

Where AI runs out of road

AI can write your demo script. It can build a sandbox flow, anticipate objections, even draft the recovery line for when something breaks. It is, as I like to say, a vibe engine: excellent at simulating tone and structure at scale.

What it can’t do is stand on a stage when the staging environment resets itself and read the room in real time. It can’t feel the shift in energy when a few thousand people go quiet at once. That recovery, that improvisation under pressure, is entirely human. It’s the 20% AI is never going to take from you, and it’s exactly the part that makes people remember you afterward.

Back in the System Center days, we did EVERY demo live. We shipped literal racks of servers to events to run private cloud infra. We had shadows backstage mimicking mouse clicks and keyboard strokes and could switch over. And we did post-keynote sessions where we showed our technical audiences what we built and how it was running. That brings street cred and confidence.

If you’re hiring, make them show you

Founders, this part’s for you. If you’re evaluating a PMM candidate, or wondering whether the person you already have can do this, skip the portfolio review and skip “walk me through your favorite launch.” Words are cheap in an interview. Everyone’s favorite launch sounds great in the retelling.

Here’s the test: ask them to prepare a demo. Doesn’t have to be your product … their last company’s product, a tool they love, a hobby they can explain in five minutes. Give them real time to prep, a day, an hour, whatever’s fair. Then ask them to open a screen share and deliver it live, in front of you.

This isn’t a surprise-attack interview trick. Let them prep, that’s the point, not the loophole. What you’re testing is whether they can turn preparation into something told live, in the room, instead of something read off a script. Almost anyone can claim they’re a strong demoer or a sharp storyteller on a resume. Only the good ones can prove it live, on command, with you watching.

Ten minutes of that screen share will tell you more than an hour of talking about their process. Same instinct as Zero-Trust fact-checking: don’t trust the claim, verify the output.

How you build this

Not with a workshop. With reps.

Get sandbox access on day one, before you’ve even finished onboarding. Shadow SEs on real calls before you ever present solo, and pay attention to what they do when something breaks, not just what they say when it works. Demo internally, to the most skeptical audience in the building, before you ever demo external. And when something goes sideways in front of a real customer, treat it as reps, not failure.

The PMMs who can drive the console with total comfort aren’t the ones with the best slides. They’re the ones who’ve broken things enough times, in low-stakes rooms, that nothing left in the console can rattle them anymore.

Get uncomfortable early. It’s a lot cheaper than getting uncomfortable on stage.

Adam