Stop Optimizing. Start Committing.
If this made you feel seen, challenged, or called you’re not alone. Lights On is for those who crave more than noise, trends, and surface takes. Your presence here means something. Subscribe and stay with the signal.
The Signal Everyone Is Missing
The pattern keeps repeating, across markets and formats, and the logic is always the same.
In the UK, Steven Bartlett launched a long-form interview podcast where business leaders talked about what they actually learned, including where they failed and what it cost them. The Diary of a CEO became Europe’s number one podcast and the first UK-produced show to reach one billion views and listens across Apple, Spotify, and YouTube. It does 50 million downloads and views per month and was crowned the UK’s most popular podcast of 2024 by Apple Podcasts for the third consecutive year. Off the back of that trust infrastructure, Bartlett turned down $100 million in acquisition offers to keep creative control, a decision that paid off.
In Egypt, Enterprise’s Making It series does the same thing at a regional scale, interviewing CEOs and entrepreneurs about building businesses in the Arab world, going deep on vision, strategy, and the conflict that both drives and hinders them. In the Gulf, ABtalks by Anas Bukhash built a global audience, now available on Netflix by discovering the human behind the title, shining a light on the raw side of achievers and leaders.
Three formats. Three markets. One mechanism: a real person on camera saying things they cannot take back.
None of them built audiences by producing more content faster. All of them built audiences by putting practitioners in front of a camera and letting them speak from genuine experience failure included. The trust that transferred became the business asset. Clients came not from outreach but from having already watched, already decided, already trusted before the first conversation began.
No landing page produces that outcome. No AI-generated campaign comes close.
The question worth asking isn’t why this format works. It’s what it tells us about what trust actually is and why the current trajectory of AI deployment is, for most businesses, moving directly away from it.
What Trust Actually Is
Trust is not sentiment. It is not likability, familiarity, or brand recall. Those are proxies useful, measurable, and entirely insufficient.
Trust, in the economic sense that matters for business, is the reduction of perceived risk in a decision. When a buyer trusts you, they are making a specific calculation: the probability that you will behave consistently with their interests, even under conditions they cannot fully observe or control. It is, at its core, a bet on your behavior in the dark.
This is why optimized content cannot manufacture it.
Optimization works by maximizing signal strength, sharper hooks, cleaner design, higher engagement rates. But trust is not built by signal strength. It is built by signal consistency over time, and more specifically, by signals that are costly to fake. A polished carousel proves nothing about how you behave when a project goes wrong. A testimonial system producing 1,000 five-star reviews is evidence of process discipline, not character. These things are useful. They are not sufficient.
What Bartlett, Bukhash, and the Making It team produced was something different: costly, irreversible vulnerability. Practitioners exposing failure, doubt, and the specific texture of their worst professional moments, on camera, permanently, for anyone to watch. That kind of signal is almost impossible to optimize because optimization requires control. And vulnerability, by definition, means releasing it.
The market read the signal correctly. And it responded accordingly.
The AI Paradox
Here is the structural problem that most businesses are not yet looking at directly.
AI is an amplifier. It takes what you already have your argument, your voice, your positioning and multiplies its reach and velocity. This is genuinely valuable. But amplification is not neutral. It amplifies the signal and the noise in equal proportion.
In a market where every competitor has access to the same amplification infrastructure, the output quality gap closes rapidly. What remains is the question buyers are now asking before they ask anything else: is there a real person behind this, and do I trust them?
The data is instructive. Bartlett’s podcast draws 50 million listeners a month, not because the production is exceptional, but because the format is structurally difficult to fake. People show up for the same reason they open email from sources they already trust and scroll past content, however polished, from sources they haven’t. The gap is not algorithmic. It is relational.
Get FARIDA KHALAF’s stories in your inbox
Join Medium for free to get updates from this writer.
AI can help you produce the content. It cannot generate the reason people come back for it.
This is the paradox: the faster AI accelerates content production, the more trust becomes the scarce variable. Speed without trust does not produce compound returns. It produces compound noise, more output, less signal, diminishing attention from the audience that actually matters.
Vulnerability as Infrastructure
These formats work because they are structurally incapable of being optimized.
This is not a romantic argument about authenticity. It is a cold observation about signal economics. In a market saturated with polished, controlled, low-cost content, the production of something genuinely costly, financially, emotionally, reputationally, is differentiating precisely because most competitors will not do it.
Vulnerability is not the willingness to appear imperfect. It is the willingness to incur real cost in the act of communicating. The production budget is not the point. The irreversibility is. A CEO who describes on camera the moment their business nearly collapsed cannot un-say it. That commitment, visible to every viewer, is what makes the trust transfer real. It is why a stranger who has never met you will enter a sales conversation already decided.
This has implications beyond podcasting and documentary series.
Every business interaction is either accumulating trust or depleting it. The mechanisms vary, a decision made against short-term interest, a public acknowledgment of error, a piece of writing that takes a position rather than optimizes for engagement but the underlying logic is identical. Trust is built by costly signals, sustained by consistency, and destroyed instantly by a single act that reveals the signal was manufactured.
AI cannot manufacture the signal. It can only amplify what is already there.
The Strategic Implication
The businesses that will compound through this period are not the ones deploying AI most aggressively. They are the ones who understand that AI deployment is only as valuable as the trust infrastructure underneath it.
An AI-generated email from a trusted source gets opened, read, and acted on. An AI-generated email from a source the reader doesn’t trust gets filtered, not by spam algorithms, but by the reader themselves, before the first line.
The equation is not complicated. More AI without trust produces more noise. More AI with trust produces compounding advantage because each piece of content benefits from the credibility already built, and each interaction adds to the reserve.
The firms that understood this earliest are not the ones with the best AI stack. They are the ones who built the trust infrastructure first, and are now multiplying it.
Bartlett didn’t build one of the five most-listened-to podcasts on Spotify globally by optimizing content. He built it by putting real people on camera to say things they couldn’t take back and when offered $100 million to hand over control of that trust infrastructure, he walked away. Bukhash built a regional audience that reached Netflix by the same mechanism. The Making It team built credibility across the Arab business world the same way. None of them were making a marketing decision.
They were making an infrastructure decision.
The difference, in the age of AI, is everything.
Closing
Trust cannot be accelerated. It cannot be compressed. It cannot be delegated to a model, however capable.
What it can be is built, systematically, deliberately, at real cost while everyone else is optimizing for speed.
AI is making that speed available to everyone. Which means the only sustainable edge left is the one it cannot replicate.
The businesses that build that now are not just ahead of their competitors.
They are building something those competitors will not be able to buy.
Every interaction you have today is either building the asset or depleting it. There is no neutral.
The algorithm can be copied. The commitment cannot.
Originally published at https://fafi25.substack.com.
