AJ Agrawal has a habit of making bold calls about where business is headed. His predictions have been published in Forbes, CNN, Inc. Magazine, Fortune, Fast Company and Entrepreneur.com. He has also built and sold companies of his own, so he writes from the operator’s seat.
His newest prediction is a contrarian one. Most enterprises are racing to hand customer relationships to AI. Agrawal argues the advantage will go to companies that choose the right moments for a person to step in. Those are the moments where personal attention and accountability matter most, like an account at risk or a public complaint.
In a recent piece on LinkedIn, he calls the payoff “the human premium.” When I came across that piece, I wanted to hear more. We talked about the executives who already practice it and how companies can tell whether it’s paying off. The reply that started it Your piece opens with a reply from Google’s CEO. Why that moment?
It stopped me. Peter Steinberger created OpenClaw, an open source AI agent. He posted on X that its Android app had been stuck in Google Play review for over a week. Sundar Pichai replied 31 minutes later, on a Saturday.
He wrote four words: “Ack, will follow up.” Soon after, the update was live. That reply passed 334,000 views. After I saw it, I went down a rabbit hole. More and more executives are doing this.
And when they do, their replies get far more engagement than the same response from a company account. So I started asking why. My answer is that a human responding to a customer has become rare. That makes it one of the most overlooked growth levers enterprises have right now.
What the human premium means You call this the human premium. What do you mean by that? Most companies are handing customer relationships to AI. Support runs through chatbots, and outreach runs through automated sequences.
I get it. It’s cheaper, and on paper it scales. But when almost every touchpoint is automated, a real human interaction becomes rare. Rare things get noticed.
That’s the premium. Two things drive it. The first is scarcity. If a customer’s last ten interactions came from software, the eleventh one from a person stands out.
The second is the signal. A machine’s message costs nothing to send, so it says nothing about how much the customer matters. A person’s time is limited, and everyone knows it. The premium is highest when the person who shows up can also make the decision.
Steinberger heard from someone who could actually get his app unstuck. That authority is part of what made the reply land. Isn’t this just an argument against AI? Not at all.
I use AI every day. I build with it and research with it. I think the companies that win in customer success will find the perfect balance between AI and human interaction. AI will carry the volume.
People will carry the moments that matter most. The executives who answer themselves Are there executives who do this that you look up to? A few. Sundar is the obvious one after that OpenClaw reply.
Mark Cuban is another. Among entrepreneurs, he’s famous for going through his own email. He’s said he gets around 750 emails a day and answers at least 150 of them. Founders talk about that with something close to fanatical appreciation.
He’s at least giving you the time of day. The most inspirational one for me is David Graham, the SVP of Customer Success at Intuit. He replies to customers himself on LinkedIn. Like, he actually goes through his feed, looks for customer complaints and then responds himself.
I actually found out about it through a viral LinkedIn post from a customer. She got locked out of her QuickBooks account on a Saturday. She posted about it and asked her network for alternatives. Within the hour, David was texting her personally.
They texted for over an hour, and he listened to her entire support call. In one evening she went from almost canceling to becoming a public advocate for Intuit. Someone with that much responsibility doing this himself is leading from the front. You could see it in the comments.
One Intuit support employee wrote that she couldn’t wait to share the post with her team. The criticism Some commenters on that post pointed out that the customer has a big following. Others said an SVP can’t work every escalation. Are those fair criticisms?
Both are fair. A customer with no followers deserves the same path to a resolution. And an SVP shouldn’t handle every case. The value comes from scarcity, so these moments will always be selective.
The real job is building a system around it. The right moments need to reach the right leader fast, whoever the customer is. Then the company has to fix whatever broke, so the next customer never needs the SVP at all. What customers say they want What does the data say about how customers feel?
When Gartner surveyed 5,728 customers, 64% said they’d prefer companies didn’t use AI for customer service at all. And 53% said they’d consider switching to a competitor if a company did. A 2025 SurveyMonkey study found 79% of Americans strongly prefer dealing with a human over an AI agent. Only 8% preferred the AI.
So companies are automating the exact moments customers most want handled by people. That gap is where the premium comes from. Personal outreach, tested You’ve tested this in your own outreach. What have you seen?
I spend less time on how many emails I send and more on making each one personal. In my experience, 100 great personalized emails consistently beat 5,000 generic ones. I also make personalized videos for specific people. Honestly, they’re a grind.
Each one takes a long time, and I usually end up recording it over and over. There’s a fair argument that a cold video isn’t worth that kind of investment. But the videos work. So far they’ve booked me more than 10 meetings with Fortune 500 executives, all from cold outreach.
They also seem to land better now than they did before AI outreach took off. I can’t prove AI caused that. My hypothesis is that as AI usage goes up, human effort earns disproportionately bigger returns. The bigger change shows up in the first meeting.
When a senior VP takes a meeting from generic AI outreach, they’re not that excited. It feels like one of many meetings, and they’re probably comparing you with other vendors. When I reach out with a personal email or a short video, they can see I did the research. They got something they weren’t expecting.
That earns a level of respect and trust before we ever talk. A higher reply rate is nice, but trust that’s already there when the meeting starts is worth far more. Customers who prefer people Not every customer wants a human. How should enterprises handle that?
It varies a lot. In a 2024 Five9 survey, 66% of Gen Z said they prefer a real person for customer support. Among Baby Boomers and older customers, it was 86%. Age doesn’t explain all of it, though.
My friends make fun of me for this. Even when there’s a self checkout, I still go to the register with a person. I like asking how their day is going, and I like having someone physically help me. I grew up with technology, and I like to think of myself as forward thinking.
In those moments I’d still rather have the human. Everybody is different, and that’s the tricky part for enterprises. They have to figure out which form of communication each customer actually wants, then offer it. Some customers want a bot at midnight.
Others want a name and a phone number. How to measure it If an executive wants to try this, how should they measure whether it’s working? Carefully. Executives rarely step in at random.
They show up for the biggest accounts, or for the ones already in trouble. Compare those accounts with everyone else and you’ll mostly measure who got special treatment. So compare like with like. Match each account that got executive time with similar accounts that didn’t.
Record why the executive stepped in, because a sa
Leave a Reply