One grew revenue 50x after half his team quit over the strategy. One brought in 50K signups in a single day with no paid budget. One generated 100M+ views from a stunt that took 50 hours to conceive. One asked every prospect to demo the product themselves instead of demoing it for them.

None of them followed the safe playbook. They treated GTM like an experiment, moved before they had proof, and made bets most founders would never get approved.

HubSpot for Startups documented all 6 stories in the free Bold Bets Playbook. The risks they took, why it was risky, and what it returned.

Read the free Bold Bets Playbook

Dennis Pekel quit GumGum to sit at home and think. No company, no product, no number. Just the sharpest read on the AI bubble I have heard all year, and a phrase you are going to steal: the alienation tax. The cost of selling to people whose worldview you never checked. He cannot tell you what it costs. We spent an hour finding out anyway. This is The ADOTAT Show.

Look, I have been doing this long enough to recognize the sound, and the sound right now is every vendor in adtech discovering artificial intelligence at the exact same moment, like a middle school that just got Wi-Fi. Suddenly everyone is an AI company. The retargeting outfit is an AI company. The viewability vendor is an AI company. The guy who sold you a dashboard in 2018 is, I regret to inform you, also now an AI company. Nobody changed the product. They changed the deck. They went into the slides, found and replaced "machine learning" with "AI," bumped the price, and called it a roadmap.

And here is what makes it funny, in the way that only very expensive things are funny. Your vendor said AI. Your board heard AI. Your CMO repeated it on the earnings call with the serene confidence of a man who has not read his own contract. And now, somewhere in row three of that board, there is one person who actually reads, the one with the law degree and the irritating habit of follow-up questions, and that person is going to lean into the microphone and ask you the thing you cannot put on a slide. What did the AI do. What did it return. Show me. And the room is going to get very quiet, because the honest answer is a shrug in a good blazer.

So let us have the numbers ready, because the numbers are a crime scene. MIT studied enterprise generative AI pilots and found 95 percent of them never made it out of the lab. Ninety-five. That is not a rough patch, that is a genre. PwC went and asked CEOs directly in its 2026 survey, and 56 percent said they got nothing. Not "modest returns." Not "early innings." Nothing, the word people reach for when they are being polite about a fire. And then there is the stat your vendor will wave in your face like a hall pass, the McKinsey line that 88 percent of organizations now use AI in at least one business function, which sounds like a tidal wave right up until you notice that "in at least one function" is doing the work of nine grown adults. Somebody, somewhere in the building, ran a chatbot once, and the whole company now counts as having adopted AI. By that standard I have adopted Pilates. I own the pants. I have been twice.

Enter Dennis Pekel, who has the good manners to diagnose the disease before anyone gets near the symptoms. He thinks we are living through a crisis of truth, and I want to be clear, he means it flatly, as a fact about the room, not as a flourish for a panel. His words. "People don't have similar views on the world. They don't have similar facts. So we are speaking about a very disjointed society." He watched two educated former colleagues argue themselves into a wall and could not figure out why, so he did the operator thing and asked what media each of them actually consumed. Different shows. Different stations. Different websites, different podcasts. Functionally different planets, sharing a Wi-Fi password. "As a brand," he told me, "you need to deal with the fractured reality," the kind the platforms make worse by feeding everyone more of what they already believe. And out of all that comes the sentence the whole thesis hangs on, the one you will be quoting in a meeting by Thursday whether you credit him or not. "If the worldview of the consumer doesn't align with the worldview of the brand or the message, there is an alienation happening between the brand and the consumer."

That alienation, he says, has a price, and he calls it the alienation tax. Genuinely great phrase. So naturally I did the obvious, rude thing and asked him to put a number on it, because a tax implies a figure, and a figure is the difference between a budget and a TED talk. This is where the beautiful machine drove into the wall at full speed. "Philosophies don't get budgets," he agreed, cheerfully, right before failing to produce one. He tried. He reached for the back of the napkin and gave me his math out loud. Global ad spend, he said, is somewhere around 850 billion dollars. Only about 10 percent of it lands in news, politics, or genuinely opinionated environments, the places where attention actually runs hot, which by his arithmetic leaves something like 85 billion dollars that is, in his word, underutilized. Which is a real number attached to a real instinct attached to, as of this recording, no actual invoice. You can see the tax even when nobody will itemize it. Ask Google, which pulled its "Dear Sydney" Gemini ad off the Paris Olympics in 2024 the instant the internet decided that coaching a little girl to outsource her feelings to a chatbot was a bridge too far. That is the alienation tax arriving as a PR crisis instead of a line item. It always shows up. It just refuses to come with a receipt.

Now here is where Pekel stops sounding like a philosopher and starts sounding like a man trying to build something, which is the part that should worry the holding companies and excite the rest of you. He does not want the alienation tax to live in keynote slides. He wants it to live in the bid stream. His framework, the actual machine, runs on two axes. One, conservative mindset against progressive mindset. Two, high trust in institutions against low trust in institutions. Cross them and you get four quadrants, four populations, four ways of receiving the same message, and his ambition, stated plainly, is to "label very clearly the sentiments" of content and brand alike and "match that with worldviews," riding inside the OpenRTB protocol "as a custom signal, custom field." Read that twice. He wants cultural fit to be a thing that moves the bid, not just a thing that shapes the copy. Whether that is the cure for the fracture or a more sophisticated strain of the disease is the question he has not fully answered, and to his credit he does not pretend otherwise. The FTC, as it happens, now has very strong opinions about who gets to apply that kind of filter and with whose permission. We will get there. Hold the thought.

He also refuses the easy version of the America-versus-Europe fight, which I appreciated, because the easy version is boring and he is not. The US, he says, is "high velocity, high risk, high rewards, very much focused on profit." Europe is "way more a guided collective responsibility," governments and institutions in the room balancing interests whether you invited them or not. And then the line that should be embroidered on a pillow in every holding company lobby. What Europe can learn from America, he said, is the velocity, "being more bold and go all in." What America can learn from Europe is "to be a little bit more mindful and take into account perspectives and the interest of different stakeholders." Translation. One side ships before it thinks. The other thinks until the moment to ship has politely left the building.

Then he handed me the frame, the one nobody says into a live microphone on an earnings call. America is the AI factory that does not use its own machine. It builds the engine, pours in the billions, throws the party, and quietly assumes the rest of the planet wants to drive the same car off the same lot. "That's of course what went wrong with the auto industry," he said, "and that's why I took it as an example." Detroit, the 1970s, enormous and certain and building entirely for itself, right up until it wasn't. He pushed it further than I expected. The first-mover advantage, the capital, the head start, all real, all temporary. "As time goes on," he said, it becomes "very possible to also build models way cheaper, and then the capital doesn't matter that much anymore, and your competitive advantage will reduce." Meanwhile the Gulf states are sitting at 84 percent AI adoption and India is moving faster on a deeper and cheaper engineering bench, and the punchline writes itself. The factory is American. The driving is happening somewhere else.

And underneath all of it, the thing the sticker exists to cover. Most of what gets sold to you as AI is machine learning in a fresh coat of paint. The optimization that was "our algorithm" in 2019 is "our AI" in 2026. Same math. Same model. New invoice, bigger number, and a vibe. When I put it to him that the real adoption rate is close to nothing and the rest is just machine learning with an "AI" suffix bolted on, he did not argue. "Yeah, yeah," he said, the way an operator agrees when you have just described his entire industry back to him. Ninety percent of the business says it has adopted AI. Under ten percent mean it. And the gap between those two numbers is precisely where your budget goes to quietly die. Pekel even put his finger on the real prize hiding behind the theater, the question of whether anyone can "totally redo the plumbing of the industry," the way media gets transacted between supply and demand, the pipes "almost fully owned by the DSPs and the SSPs." That is the genuine game. Everything else is a sticker and a smile.

So here is your week. You are going to sit across a table from someone who says AI constantly and proves it never. Your board is going to ask what it returned. You need the kit, and you need it before the meeting, not in the back of the cab afterward drafting the apology email with one thumb.

But meet the man first, because honestly that is the better story, and I do not say that about many people in this business. Dennis Pekel built a machine for sorting the entire world by worldview, by who still trusts the institutions and who has stopped, by which stranger across the table sees reality the way you do and which one is a problem to be managed. Nobody builds that by accident. Part Two is why he built it, the part the LinkedIn bio quietly deletes. A desert island he would stock with five executives and a barbecue chef and not one member of his own family. A framework that turns out to be autobiography. A man who swears up and down that he does not do politics while sketching, in fine detail, the political fracture of the entire West. Then Part Three is the deliverable: the eight questions, the relabeling tells, the scorecard, the three numbers for your board, and the one worldview question to ask now that the FTC changed what your agency is legally allowed to do with your money. Plus the downloadable one-pager. Join ADOTAT+