Colin Kelton got his first job at Vanguard in 1990. He stayed for 36 years. In an industry where people job-hop every three or four years, his entire career is a study in consistency — and so is the brand he was protecting.
Vanguard manages $13 trillion in other people’s money. It’s the second-largest asset manager in the world, but what makes it different isn’t the size. It’s the structure. Vanguard is owned by its investors. There are no outside shareholders demanding quarterly profits. No Wall Street pressure to push higher-margin products. Just Jack Bogle’s original principles, unchanged for 50 years — think about goals, stay balanced, keep costs low, be disciplined.
That consistency shows up in the numbers. In 2025, Vanguard ranked No. 1 in brand trust in the J.D. Power Investor Satisfaction Study. 73% of people give Vanguard a 9 or 10 out of 10 for reputation. Apple sits at 39%.
So when Colin rose through the ranks to become Vanguard’s first Global CMO in 2018, he wasn’t just managing a brand. He was stewarding one of the most trusted names in finance. And he didn’t walk into a perfectly oiled machine. He walked into a fragmented function with outdated, expensive infrastructure.
“When you say I want somebody to build a new website, when you’re done, you see the new website. When we were done, we were still sending emails and producing ads, and people are like, well, I don’t quite understand what I’m getting for this.” – Colin Kelton
His conversation on Humans of AI is the story of what happened next. Colin rebuilt the tech stack. He standardized the brand. He killed good work that wasn’t on brand. And he did all of it years before AI was a conversation. So when AI arrived, Vanguard was ready. Not because anyone predicted it, but because the hard work was already done.
- AI doesn’t create what makes your brand different — it amplifies whatever’s already there. If your brand has never spoken in a way that’s distinctive, AI won’t fix it. It will take what you’ve said before, put a spin on it, and reproduce your lack of distinctiveness many times over.
- The work that makes AI work isn’t AI work. Vanguard rebuilt its marketing tech stack in 2018–2019 and standardized its brand across divisions and regions. When AI arrived, the foundation was already in place. Companies that haven’t done this work will have to do it and join the AI boom simultaneously.
- LLMs are scanning third-party signals, not just your own messaging. Search engines now look at Reddit, press, social media, and endorsements to determine whether your brand should be served as an answer. This means brand investment matters more now, not less.
- The efficiency trap causes companies to miss revenue growth. The first instinct with AI is to cut costs and reduce headcount. Colin’s CEO reframed it: “If you can add a zero to the revenue line, that would be a lot better.” AI can deliver higher quality, faster speed, and better targeting — not just savings.
- Change management starts small. Trying to convince a 600-person organization to turn on a dime doesn’t work. Start with a small group that’s genuinely interested, let them innovate, and expand outward.
- Writing skills become more valuable, not less. When asked which marketing skill becomes more valuable because of AI, Colin’s answer was immediate: writing skills. Not because you’ll write more by hand — but because someone still has to look at AI output and judge whether it’s well-crafted or AI slop. That judgment requires the reps of writing itself.
Framework: The distinctiveness mandate
Colin Kelton’s approach to preparing Vanguard for AI rests on a framework built over 36 years. It has three components, and none of them are AI work:
1. Rebuild the foundation before you need it.
Colin convinced Vanguard to rip out its fragmented collection of mini tech stacks and rebuild a single, unified stack across the firm. The ROI wasn’t visible on day one — the emails still went out, the ads still ran. But when AI arrived, the infrastructure was ready. Companies that haven’t done this work will have to rebuild and adopt AI at the same time.
“We had three of this and two of that and five of those. One person liked one analytic system, another person liked another. I convinced the organization to give me money to rip it all out and redo the entire tech stack and have a single tech stack throughout the firm.” – Colin Kelton
“We obviously didn’t know anything about AI per se when we started that in 2018, 2019. But had we not done that, we would not have been able to take advantage of what happened with AI.” – Colin Kelton
2. Define and standardize what makes you different.
Vanguard had distinctive divisions whose messages sometimes played against each other. Colin’s job was to standardize the brand across every division, region, and channel — and to say no to good work that wasn’t on brand. As Diego Lomanto, WRITER’s CMO, noted in the conversation: one of the hardest things a marketing leader does is kill something that’s good and creative but isn’t on brand.
“We are seen as a serious brand. You can’t be silly in your messaging, you can’t be goofy. Jack Bogle was not a tongue-in-cheek kind of guy.” – Colin Kelton
“Jack would stop me in the hallway when he was still alive and ask me why we’re spending all this money on marketing. I’d say, well, Jack, you were the best marketer we ever had. We’re just spending all this money to replace you. Because his messages were what cut through everything.” – Colin Kelton
3. Spend more at the top of the funnel, not less.
LLMs are scanning third-party signals to decide which brands to surface. That means high-level brand investment (the kind CMOs have struggled to justify to CFOs for years) now has a measurable argument. The search engines are rewarding exactly what brand marketing was always trying to do.
“The engines are frankly smarter. They’re looking for more evidence that your brand should be served up as an answer — third parties, social media, Reddit, press, endorsements. You’ve got to do that in order to show up in these LLM search results, or you’re just going to be invisible.” – Colin Kelton
FAQ
How do enterprise marketing teams stand out when AI makes content infinitely scalable?
You stand out by knowing what makes your brand different before AI touches your marketing. AI amplifies whatever’s already there. If your brand is generic, AI scales the genericness. The companies that will differentiate are the ones who’ve done the hard work of defining their distinctiveness first.
What should marketing leaders do to prepare their organization for AI?
Start with the unglamorous infrastructure work: rebuild your tech stack, standardize your brand across divisions and channels, and define what makes you different. Then start small with AI adoption: with a group that’s genuinely interested, rather than with a mandate to the entire organization.. Trying to turn a 600-person team on a dime doesn’t work.
Why does brand investment matter more in the age of LLM search?
LLMs don’t just read what you say about yourself. They scan Reddit, press, social media, and third-party endorsements to determine whether your brand should be served as an answer to a buying question. This gives CMOs a new argument for brand budget that didn’t exist three years ago: the search engines are rewarding exactly what brand marketing was always trying to do.
Should companies use AI for efficiency or revenue growth?
Both, but don’t stop at efficiency. The first instinct is to cut costs and reduce headcount. Colin Kelton’s CEO reframed the conversation: “If you can add a zero to the revenue line, that would be a lot better.” AI can deliver higher quality, faster speed, and better targeting — not just savings. Companies that only focus on cost cutting miss the chance to grow.
“That would be great if you could do that. On the other hand, if you can add a zero to the revenue line, that would be a lot better.” – Colin Kelton, recalling his CEO’s reframe
How do you manage the human side of AI adoption?
Start small. Let a group that’s genuinely interested innovate first, then expand. The hype cycle around AI feels existential, but Colin compares it to the arrival of the iPhone: people at Vanguard resisted building an app, and it didn’t take anyone’s job. AI will be a slower burn than people think, and a person will always be there judging whether the output sounds like the brand.
“We don’t know the answer yet. We know where it’s headed, we know it’s going to be big, we know this is real. But we don’t know how it’s going to wind up. The most successful ones are out there in it, making some mistakes, maybe using something that doesn’t work. But you need to be in the mix.” – Colin Kelton