AI & Marketing in 2026 with Neil Patel ($100m agency owner)

Neil Patel is not impressed by AI adoption theatre. The co-founder of NP Digital, a global ad agency in 28 countries, says he asks rooms of marketers around the world whether AI has increased revenue, and usually only 2% to 3% raise their hands.

His view is grounded in 24 years of hands-on marketing rather than contrarian performance. He still tears apart lead funnels line by line, but spends much of his week diagnosing business problems, making deals and challenging the work that looks impressive but does not move revenue.

This conversation is useful because Neil names the vanity traps senior marketers are under pressure to defend: reach without buyers, Super Bowl ads without revenue, AI content without advantage, and headcount cuts dressed up as efficiency.

<iframe width="100%" height="180" frameborder="no" scrolling="no" seamless src="https://share.transistor.fm/s/ece82006"></iframe>

This interview has been edited for clarity and concision.

Tom: After nearly 25 years in marketing, why do you still get into the detail of funnels instead of staying at strategy level?

Neil: We’re a global ad agency in around 28 countries, and I’ve been doing marketing for 24 years, about to hit 25. I love growing traffic, generating leads and sales online. For me, it’s an addiction.

If you ask where I spend a lot of my time today, it is entrepreneurship and doing deals. I still know marketing extremely well compared with many people, and I still get my hands dirty. Yesterday I clicked leads through a few funnels on our website, tore them apart, and put every change I wanted into a Word document: the APIs I wanted called, the text I wanted, and the lead form changes.

On a weekly basis, I pinpoint problems inside the organisation and solve them one by one. If you asked me to manage employees and run a big organisation, I would fail miserably. I am much better as an executor, strategist and deal maker, so I focus on what I am good at.

Tom: Your LinkedIn following is around 801,000, but you say the personal brand was an accident. What were you actually trying to build?

Neil: I did not even know I had 801,000 followers on LinkedIn until you pointed it out. I assumed it was closer to 300,000. I do not look at following and clout the way many people do.

When I started, I did not have money to spend on ads, so I spoke at conferences and blogged when blogging was not popular. I hoped it would help generate business. The intention was never to build a personal brand. The personal brand happened by accident.

I appreciate people coming up to me, but I did not do it for recognition. I log into TikTok close to zero times a week, and Instagram maybe two or three times to send leads to my sales team. I am unplugged from a lot of what marketers spend time on because much of it does not affect revenue.

Tom: Where do marketers get fooled by reach, and what kind of content has actually brought in serious revenue for you?

Neil: Everyone thought I was blogging to get tons of visitors. We ranked for things like “Instagram followers” and got around 600,000 visitors a month from terms such as how to get more Instagram followers and how to be popular on Instagram. It drove no revenue. It only created calls from people saying, “my daughter wants to be Instagram famous, can you help her?”

At one point we ranked number one for “digital marketing”, which sounds like a prime keyword for an agency. It drove little to no revenue. The content that worked was about specific edge cases for large enterprise businesses, such as how to structure a blog when you have multiple products and multiple languages.

That kind of piece does not get much traffic, but the person reading it may be in a company operating in more than 100 countries with 32 products and internal competition across divisions. That client can pay seven figures a year. The content that got attention, like what I learned from flying first class for a year, got views. The specific content drove revenue.

At NP Digital, the first year was around $5 million or $6 million in revenue, and almost all of it came from my personal brand. The second year was around $18 million, with about $10 million from my personal brand. By year three we were somewhere in the $30 million to $40 million range, and the personal brand was still about $10 million. The growth after that came from referrals, word of mouth, awards, good employees and very specific work for ideal customers.

Tom: If reach is not the same as revenue, why do companies still spend on Super Bowl ads, Olympic spots and Formula One sponsorships?

Neil: We have customers that run Super Bowl commercials, World Cup commercials and Olympic commercials. When I look at the numbers, very few see an increase in revenue from those ads. A Super Bowl ad might be $6 million or $7 million, but they also expect you to buy other spots, then you pay for production, YouTube and other promotional activity. You can quickly be well over $20 million.

The money in marketing is in the boring and ugly, not the sexy. Every once in a while you get a company like GoDaddy where Super Bowl ads help put them on the map, but that is a needle in a haystack.

Companies still do it because they can afford it and because it is branding. It is the same with Formula One. Some clients sponsor Formula One, see no direct ROI, but think it is fun and take top customers. When a company makes 10 or 11 figures in profit, the way it looks at money is different.

We work with one company making double digit billions in profit. They have planes for executives and buy new ones every few years because maintenance and downtime are a headache. If you are making more than $10 billion in profit, a Super Bowl ad or a plane becomes a tiny fraction of a fraction.

Word of mouth is better built through targeted marketing to ideal customers, then those customers spread it to more ideal customers. You also need a good product or service. If the product is bad, word of mouth will not work no matter what the marketing does.

Tom: You speak to marketers around the world. Why are so many using AI without seeing revenue impact?

Neil: When I ask rooms of marketers who uses AI, everyone raises their hand. I have asked this in Brazil, Colombia, Hong Kong, the US and the UK. When I ask how many see an increase in revenue from AI, it is usually 2% or 3% at most.

At HubSpot’s Inbound conference, with maybe 4,000, 5,000 or 6,000 people in the audience, less than 1% raised their hand for ROI. I asked other questions to check whether people were just not raising their hands, and they were raising them for other things.

AI is not bad. People use it in the wrong ways, and they spend time on things that are cool rather than things that drive ROI. The ROI we see is more often from cost efficiencies than revenue growth. You can do more with fewer people, so the saving is mainly on labour.

The strongest use cases we see are in product, engineering and support. In Ubersuggest, support tickets and emails can be parsed, Claude Code can help fix issues, another agent can QA, and then a human approves the release. Bugs get fixed faster. If you delight customers faster, word of mouth improves and we have seen that affect revenue.

For marketing, data and analytics is one of the strongest AI use cases, but people do not like talking about it. Marketing data is like a thousand-piece puzzle where the pieces are in different boxes in different rooms. AI can help bring the pieces together so you see what is actually working.

What I do not believe in is everyone using the same technology to create the same content and strategies. The internet has a lot of mediocre content, AI trains on that, and the output tends to be average. Average marketing does not work when everyone else is average too.

Tom: Does AI give startups a real advantage over larger companies, or do bigger companies catch up once they organise themselves?

Neil: AI levels up smaller businesses and gives them a better chance to compete for now. In a few years, large organisations will adapt, because AI creates efficiencies and they have large budgets. If AI helps them move faster, they have even more money to spend on marketing and advertising.

Technology is not the main reason startups beat big companies. Big corporations have access to the same technology. The real reason startups can beat them is bureaucracy.

At NP Digital we talked about offering customers help setting up AI agents and modernising their marketing stack. A startup would just do it. At our size, we need to think about legal, privacy, data, what happens if there is a breach, and what happens if we get sued. A startup says it will deal with that later.

That same complexity is also why marketers cannot simply pick one channel forever. Buyers are on Instagram, TikTok, Facebook, YouTube, LinkedIn, Google, Bing, Reddit, ChatGPT, Perplexity, Claude, email, SMS, push notifications, reviews, Trustpilot, Capterra and G2. Start with the main channels where your ideal customers already are, scale them, get them close to maintenance mode, then add the next channel.

Tom: When leadership wants AI to cut marketing headcount, what argument should marketers make, and what career lesson keeps you cautious about shifts like this?

Neil: I got this question in London. A marketer said their CFO wanted them to cut a lot of headcount, become more efficient and do more in marketing with AI. I said first, explain that AI produces a lot of mediocre work if you use it that way, so you will not get the results you want.

Second, ask the CFO why they still need all the bookkeepers, accountants and finance people if AI can automate so much. In that room, the answer was that the accounting team had grown. That is what we are seeing. AI has created more channels and more things to do. Companies are not cutting headcount as much as trying to get more done with the same people.

The lesson from my own career is still to adapt quickly. At Kissmetrics, we were in talks to sell to Meta and Microsoft, then a class action lawsuit and FTC investigation over data privacy hurt us, even though the FTC looked at the code and we passed. We had raised around $4 million, still owned the majority of the company between me and my co-founder, and a $60 million or $70 million exit would have been great at that point.

The bigger mistake was that we did not adapt to mobile fast enough. We were too focused on the lawsuit while the web moved to smart devices. If we had moved the product from desktop to mobile earlier, I think Kissmetrics could have been a multibillion-dollar company.

My strongest personal skill has been making deals, which I would describe as persuasion. One of my favourite books is Principles by Ray Dalio. And if anyone wants marketing help, check out NP Digital.