My family was about to miss a connection in Atlanta. Tight layover, delayed inbound, our entire vacation to St. Thomas evaporating in real time as we sat on the taxiway watching the minutes drain. So I did what any reasonable person in 2019 would do: I tweeted at Delta.
What happened next is the reason I've told this story in every keynote I've given for the past seven years.
They met us at the gate. Took us down to the tarmac. And drove us between concourses in a Porsche Macan. My kids thought we were secret agents. My wife thought we'd been upgraded to a reality show. I thought: this is what unified customer data actually feels like when it works.
Somewhere in Delta's system, the data points converged: high-value loyalty member, tight connection, family traveling with kids, vacation booking, previous history of rebooking issues. The system — or more likely, an empowered frontline employee with access to the right data — made a decision in minutes that no amount of marketing spend could replicate. Save the vacation. Make it memorable. Turn a service failure into a brand-defining moment.
What was it worth? I may never book another airline again.
Data enables. Experience delivers. That's been my one-liner for years. But the agentic era forces a harder question: what happens to the experience layer when an algorithm stands between the brand and the human?
The Loyalty Encoding Problem
Here's what makes the Porsche story complicated in the agentic context. My loyalty to Delta wasn't earned by the best price, the most convenient schedule, or the highest on-time percentage. It was earned by a moment — an unpredictable, emotionally resonant, deeply human interaction that happened because the right data met the right employee at the right time.
When I tell my agent to book flights, how do I encode that?
"Always book Delta" is the simple version. But an agent that blindly follows that instruction without evaluating alternatives isn't a good agent — it's a vending machine. A good agent should push back: "United has a direct flight that's $200 cheaper and arrives two hours earlier. Delta requires a connection through Atlanta. Are you sure?" And at that point, I have to articulate why I prefer Delta in terms the agent can weight.
Price? Delta often isn't cheapest. Schedule? Varies by route. On-time performance? Competitive but not category-leading. Loyalty points? Quantifiable, but the agent can calculate the equivalent value of switching to another program. The Porsche on the tarmac? How do you encode a feeling?
This is the loyalty encoding problem, and it's the central challenge for every brand that's built its competitive advantage on experience rather than commodity attributes. The moment that made me a Delta customer for life is exactly the kind of thing that's hardest to translate into agent-readable terms.
Two Kinds of Loyalty
The agentic era is going to split brand loyalty into two distinct categories, and most companies haven't recognized the split yet.
The first is structural loyalty — loyalty that can be quantified, compared, and evaluated algorithmically. Points balances. Tier status. Accumulated benefits. Contractual advantages. Switching costs. This is the loyalty that agents handle well. "You have 180,000 SkyMiles and Diamond status. Switching to United would cost you $4,200 in equivalent benefits over the next twelve months." An agent can compute that, weight it against price and schedule advantages, and make a recommendation.
The second is emotional loyalty — loyalty built on moments, trust, narrative, and felt experience. The Porsche on the tarmac. The gate agent who went above and beyond. The brand that feels like it gets you. This is the loyalty that agents can't evaluate because it exists in the space between data points, in the human memory of a moment that transcended the transaction.
Most brands have some of both. The question is how much of your loyalty moat survives the agent's evaluation, and how much evaporates when the decision-maker changes from a human who remembers the Porsche to an algorithm that sees a Delta flight with a connection versus a United nonstop.
How Delta Wins the Agent
Delta's strategic advantage in the agentic era isn't the Porsche. It's the ability to create Porsche moments and translate them into structural loyalty that an agent can evaluate.
Consider what Delta could build. An agent-facing service reliability API that exposes real-time and historical data: connection success rates by airport pair, rebooking responsiveness metrics, average delay resolution time, customer satisfaction scores segmented by scenario type. When a customer's agent evaluates whether to book a tight connection through Atlanta, it can query Delta's API and get: "For this specific connection, Delta has a 94% success rate. In the 6% of cases where the connection is missed, average rebooking time is 23 minutes. For Diamond members, a dedicated recovery protocol activates automatically."
That's the Porsche moment, translated into data. Not the emotional experience of being driven across the tarmac in a luxury SUV — that's irreducible, and no API will ever capture it. But the capability that made the Porsche moment possible — the unified data, the empowered employees, the recovery infrastructure — can be quantified and exposed to agents.
The brand that can say "we will save your vacation, and here's the data that proves it" wins the agent. The brand that can only say "we have a nice loyalty program and sometimes we do cool things" loses to whoever has the better algorithm.
The Service Layer as Competitive Moat
There's a broader lesson here about where competitive advantage lives in the agentic era. For decades, airlines competed on a combination of price, schedule, and brand perception. Price and schedule are pure commodity attributes — easily compared, easily optimized by an agent. Brand perception is the soft moat that justified premium pricing and customer retention.
In the agentic era, the competitive moat shifts to the service layer — the demonstrated, verifiable, agent-evaluable ability to handle complexity, recover from failure, and create positive outcomes in unpredictable situations. This is measurable. It's structured. And it's incredibly hard to replicate, because it depends on organizational culture, employee empowerment, data infrastructure, and operational discipline — all of which take years to build.
Delta's advantage isn't that they have a Porsche at the Atlanta airport. It's that they have the organizational capability to identify a high-value customer with a tight connection and make a real-time decision to save their trip. That capability — measured, quantified, and exposed through an agent-facing API — becomes a legitimate competitive moat in a world where the agent is doing the choosing.
The irony is beautiful: the very thing that makes the Porsche moment magical — the human judgment, the empowered employee, the institutional culture of service — is the thing that, properly measured, can differentiate Delta in an algorithmic evaluation. The moment is human. The moat is data about the moment.
What the Agent Can't See
I want to be honest about the limits of this argument. There are things about the Porsche experience that no API will ever capture and no agent will ever evaluate.
The look on my kids' faces. The feeling of going from anxiety to delight in thirty seconds. The story I've told a hundred times at dinner parties. The way it changed my default answer from "I'm flexible on airlines" to "I fly Delta." These are human things, and they matter — not because they're quantifiable, but because they shape the preferences that humans encode into their agents.
Brand marketing's job in the agentic era isn't to influence the agent directly. It's to create the moments and the stories and the associations that cause humans to say "always book Delta" when they're setting up their agent's travel preferences. The Porsche on the tarmac is still doing marketing work — it's just doing it upstream of the agent, in the preference-setting conversation rather than the transaction.
And there's something poignant about that. The most powerful brand moment I've ever experienced — the one I've built keynotes around, the one that's probably sold millions of dollars' worth of Delta flights through word of mouth — will eventually be reduced to a parameter in an agent's configuration file. "Preferred airline: Delta. Override threshold: $400."
The Porsche still matters. It just matters differently now.