COLE'S NOTES VAULT 2026/ BARTLETT BRIEFCHELS.AI ↗
Vault Conference · September 1, 2026

The audience
you own.

Steven Bartlett’s operating system for winning in an AI-saturated world: own the relationship, stay irreplaceably human, kill guesswork, and care more about the tiny details.

6:26–7:59 PM EDTINTERVIEW + Q&ALIMITLESS TRANSCRIPT
“I’ve not said this publicly before” · 6:51 PM

Follower counts are losing their power.

“The most important thing as a business owner who has an audience on the internet is sovereignty. Do you own the relationship with your audience?”

AI becomes the matchmakerPlatforms understand each post and route it to whoever is most likely to stay—not merely to followers.
Supply goes infiniteAgents can produce content while creators sleep. Fixed attention divided by unlimited supply means each unit of content loses expected reach.
Human + owned winsMove people toward email, transactions, events, addresses and tactile experiences. Pair sovereignty with lived human experience.
Explain this whole idea in plain English

Today: A social app may show your new post to your followers. This can make the follower count feel like a list you own. It is not.

What may change: AI can read each post and pick who may like it. A person may see your post even if they do not follow you. This can help a small account. It also means a big account may not get big reach.

The risk: The app still picks who sees you. It can change the rules in one night. AI can also fill the app with far more posts. Yet people still have the same amount of time to watch them.

His answer: Meet people on social apps. Then ask them to join a list or place you can reach on your own. That could be email, an event, a paid group, a sale or even mail at home. He calls this sovereignty. It means the app is not the only door to your people.

Simple example: 100,000 Instagram followers can disappear from your reach after an algorithm change. A list of 10,000 customers who asked for your emails is smaller, but you can still contact them directly.

No buzzwords left unexplained

The plain-English guide.

Tap any phrase. Each explanation tells you what it means, why it matters, and what it looks like in real life.

What is “audience sovereignty”?

It means you have your own way to reach people who care about your work. You do not need a social app to let you speak to them.

Email, phone, an event ticket or a home address can give you a direct path. A follower is not the same. The app still gets to pick what that person sees.

Think of it this way: Followers are people in someone else’s building. An email list is a door you have the key to.

What does “social is the top of the funnel” mean?

A funnel is just the path from first meeting someone to earning their trust or business.

The top is where strangers discover you. The bottom is where they take a useful step, such as joining your email list, booking a call, attending an event or buying something.

Steven’s point is that a like or follow is no longer the finish line. It is only the first hello.

What is “AI matchmaking”?

AI can look at a video and work out what it is about. It can then guess who may like it. The app shows the video to those people.

A good post from a small page can now take off. A big page can no longer count on all its fans seeing each post.

What is an “irreplaceably human” story?

AI can repeat facts and make a list of tips. It cannot live through a loss, build a firm, raise a child or feel fear.

A human story tells what took place. It also tells how it felt, what went wrong and how the person changed. Steven thinks we will want more of this as plain facts get cheap.

Instead of: “Five tips for hiring.” Try: “The hire I got wrong, what it cost us, and the warning sign I missed.”

What does “kill guesswork” mean?

When two people argue about which title, design or offer is best, do not keep arguing. Run a small test and let real behavior answer.

Change one thing at a time. Say what you expect. Pick the number you will measure. Record what happened. A failed test is still useful because it removes one bad answer.

What is the “1% principle”?

A small fix can look like it did not help. Keep making good small fixes and they add up over time.

Steven’s team tests room air, guest music, titles and art. One change may not save the show. Many good changes can.

Save one minute on a task done 500 times. That is no longer a small gain.

What is a “paper wall”?

A paper wall is a rule that looks solid but is really only a habit, old choice or tradition.

When someone says, “It always takes eight weeks,” ask why. The real problem may be one supplier, an old computer or a step nobody has questioned. Some walls are real. The job is to test them before accepting them.

What does “stay unromantic” mean?

Care deeply about the result, but do not fall in love with one method.

If readers dislike your favorite chapter, cut it. If AI edits better, use it. If a winning process stops working, rebuild it. Steven’s point is not to become cold. It is to protect the mission from your own pride.

What is “long-termism”?

Your deadline changes your choices. With sixty seconds to build a tower, you stack bricks fast. With a year, you build a wide base first.

Long-termism means choosing a longer time frame so you can build stronger systems, better relationships and assets you own. It may look slower today but survive longer.

What is “value per minute”?

It asks: how useful, moving or entertaining is each minute for the audience?

A long interview can work if most minutes earn attention. A short video can fail if it wastes time. The goal is not “make everything short.” The goal is “remove minutes that do not pay off the promise.”

Words worth keeping

The lines that carry the room.

Swipe or use the arrows. Quotes are lightly cleaned for punctuation from an automatic transcript; timestamps point back to the source.

01 · EXPERIMENTATION
“If you incentivize the outcome, you disincentivize the behavior—because people get scared of being wrong.”
Full context

A test means you do not know the answer yet. If a worker gets blamed when a fair test fails, the worker will stop testing. Steven tracks how much the team learns. He does not demand that each guess be right.

6:38 PM · Reward the rate of trying, not lucky outcomes
02 · SOVEREIGNTY
“Social media should be seen as the top of the funnel. Your owned environments have to be the bottom.”
Full context

Use social apps to help new people find you. Then give them a good reason to join an email list, event, customer account or community you can reach directly. A follow is the first hello, not the final goal.

6:52 PM · The take he had not shared publicly
03 · HUMAN MOAT
“If there’s not a layer of lived experience, you should assume the value of the thing is going to decay.”
Full context

AI can give people facts. It cannot tell your true story. Steven thinks plain facts will lose value. Work with real fear, hope, pain and change will mean more.

6:55 PM · Information is cheap; experience is not
04 · DETAILS
“The most important thing is to care more than your competitors do about small details—and try more things.”
Full context

He does not believe one secret trick makes a great show. The advantage is a habit: notice details other people ignore, test them, keep what helps, and repeat for years.

7:25 PM · The simplest version of his operating system
05 · LONG GAME
“The game here is to go long. And going long requires me to love my life.”
Full context

His team may offer a huge guest, but he says no if he does not care about the conversation. He wants to make the show for fifty years. A plan that makes him hate each day will not last that long.

7:13 PM · Why he declines famous guests he does not care about
06 · LISTENING
“Great interviewers don’t lead; they follow… people will subconsciously take you where they want to go if you just listen.”
Full context

Steven learns enough to care about the guest. He says he has never made a list of questions for the show. He hears the answer and follows what the guest cares about. This can find a better story than a fixed list.

7:14–7:15 PM · Curiosity over a written question list
The operating system

Six ideas to steal.

Not motivational wallpaper—each one changes what a team measures, builds or stops doing.

01 · KILL GUESSWORK

Increase the rate of learning.

Hold variables steady, change one thing, state the hypothesis, define the measurement, and publish the result—even when it fails.

What did he mean?

Most teams reward people only when an idea works. That makes people hide risky ideas. Steven rewards the act of running a fair test, because a clear “no” teaches the team too.

Test two real titles with a small audience before choosing one by taste. Keep the rest of the post the same so you know the title caused the change.

6:34–6:39 PM
02 · THE 1% PRINCIPLE

Tiny gains are invisible—until they aren’t.

The show stayed flat for years, then compounded. No experiment is too small when hundreds of small wins stack.

What did he mean?

One small fix may not move a chart today. Many small fixes can build on one another. Steven’s team even studies room air, guest music and tiny packaging choices.

Improve the hook, then the thumbnail, then the checkout page. Each gain helps the next one instead of waiting for one magic idea.

6:39–6:43 PM
03 · AUDIENCE SOVEREIGNTY

Followers are rented. Relationships are owned.

Measure email addresses, transactions, event attendance and direct contact—not just platform followers that an algorithm can stop serving tomorrow.

What did he mean?

This was the idea he said he had not shared publicly. AI will choose who sees each post, so follower totals will promise less. Social apps should help you meet people, not be the only place you can reach them.

Offer a useful email guide after a video. The viewer chooses to join, and you can still reach them if Instagram changes its feed.

6:51–7:01 PM
04 · IRREPLACEABLY HUMAN

Wrap tactics in lived experience.

AI can deliver facts. Humans still carry struggle, emotion, risk, relationships and the context that makes information matter.

What did he mean?

If a piece only repeats facts, AI can make a similar version. A human story has stakes: fear, sacrifice, conflict, mistakes and change. Other people can feel themselves inside it.

Do not only explain how a founder raised money. Show the rejected pitch, the doubt at home and the moment someone finally said yes.

6:53–6:57 PM
05 · PAPER WALLS

Ask which constraints are merely tradition.

Zara’s nine-month cycle, the four-minute mile and wheel-less luggage all looked fixed—until someone asked “why?”

What did he mean?

A deadline can sound like a law when it is really an old process. Steven told a story about an animator who needed nine days. When asked how to do it in two, the answer was simple: buy a faster laptop.

Ask “Which exact step takes eight weeks?” before moving the whole project. The real wall may be one approval or supplier.

7:05–7:12 PM
06 · STAY UNROMANTIC

Love the impact, not the method.

Delete favorite chapters, rebuild a winning swing, and disrupt your own product before success turns into a cage.

What did he mean?

Success makes people protect the thing that worked before. Steven warns that the old winner can block the next better answer. Keep the goal, but be willing to change the tool, format or plan.

He removed three favorite book chapters after reader tests showed they weakened the book.

7:20–7:50 PM
Signals beneath the soundbites

What the talk actually implies.

01

Distribution becomes the durable moat.

AI collapses production advantages. Trust, direct audience access and the ability to reliably reach people become more defensible than tooling.

02

Content should create an owned next step.

Every post needs a path toward a newsletter, purchase, event, community, physical delivery or another relationship the platform cannot revoke.

03

Testing is an organizational capability.

A/B testing one title is not the point. The advantage comes from a company-wide system that turns assumptions into inexpensive experiments.

04

AI changes hiring, not the need for people.

Automate research and operations; expand roles where trust and relationships matter. Bartlett said every employee now needs AI proficiency.

Turn insight into motion

A seven-day action list.

A practical first pass for an operator, creator or media business.

Audit audience ownershipCount reachable email addresses, customers, members, event attendees and postal addresses separately from followers.
Add an owned conversion to every content seriesChoose one useful newsletter, downloadable resource, event, consultation or transaction—not a vague “follow for more.”
Launch a “kill guesswork” logRecord one-variable experiments with hypothesis, metric, result and next move. Celebrate valid failures.
Create a 1% channelGive the team a place to post tiny improvements. Review the strongest compounding wins each week.
Test the packaging before polishingTest titles, hooks, thumbnails and promises cheaply. Make the opening immediately deliver what the winning package promises.
Find one paper wallAsk what would need to be true to make a “fixed” timeline, cost or process radically smaller—before accepting the constraint.
Produce one irreplaceably human storyCapture the tension, mistakes, emotion and lived experience—not merely the lesson or information.
Keep thinking muscles aliveWrite a memo before asking AI. Use the model to challenge or extend your thinking, not to replace forming the question.