Who Gets a Person?

Somewhere inside every digital transformation, somebody writes a quiet rule that never appears in the deck, never gets debated in the offsite, and never shows up on an org chart. Every customer feels it within thirty seconds of contact anyway: how much attention you receive depends on how much you are worth.

That is the real meaning hiding inside the phrase “AI-first service.” Not a better answer. A cheaper first responder. The chatbot on the lender’s website. The SMS auto-responder. The phone tree that quotes your own policy back at you in a voice trained to sound patient. And behind it all stand the humans — real ones, with judgment, authority, discretion — reserved for whoever pays enough to reach them.

This is not an airline story anymore. That war ended and everyone lost. This is mortgage, title, insurance, healthcare, banking: industries where the person typing into the chat window may be standing inside the largest financial decision of their life.

Watch what happens when a borrower asks a question that is not standard. Can I buy my mother-in-law’s house with a mortgage that already exists between family members? My income is real, it just does not have the shape your portal asks for. The bot has a face for this, and the face is “I'm sorry, I didn't understand that.” And the applicant learns something the onboarding flow never states out loud: their situation is too complicated to be worth escalating. They have been sorted.

Here is the part the vendor demo never mentions: being served by a machine is not neutral. It is a message. When the cheapest thing a company owns sits between the customer and the institution, what the customer hears is “you are not worth the cost of a person.” Nobody says it that way. They don’t need to. People know the feeling of being politely processed.

And watch the market invert the old pyramid. Human contact is becoming the luxury tier. Private banks keep branches with people in them; mass-market banks have apps. When service gets automated to the floor, people stop paying for service and start paying for a human to be present to it. “Access to a real person” becomes a plan feature, like priority boarding — monetized beautifully, because the ability to reach someone with authority when your closing is melting is not a feature. It was always the product.

Meanwhile, companies have built themselves a grotesque little lock. When the only key that opens the door to a human is escalation, the company trains every customer to be angry. Outrage becomes the priority queue. The complaint becomes the only phone number that connects. And the organization ends up surrounding its humans with nothing but the worst five minutes of each relationship — then concludes from the data that customers are difficult. The customers are not difficult. The architecture is an insult, and the customers are just reading the blueprint back out loud.

The trust math is brutal. A machine can answer. Only a person can vouch. The sorted customer does not learn to hate the bot; they learn to never trust it again — which means they start checking its claims the way you check a contract written by a party you have already decided not to trust. The institution pays twice: once for the software, and once for the reputation of making the customer do the institution’s own verification work. A customer who has been told by the interface that they are a second-class human does not recommend you. They warn other people about you, in language far better than your marketing team could ever write.

To be clear, because I have said the other half of this sentence for years: automate relentlessly. Automate the extraction, the formatting, the follow-up, the document chasing, the thousand small machines that eat administrative drag and hand back hours. If you do not, your competitor will, and they will buy back the time to spend it on your customers.

The mistake is spending the bought-back hours building a thinner wall between the customer and your firm. The right move is the opposite: let the machine answer the routine question instantly, so the human can afford to show up for the unusual one, the messy one, the frightened one. Automation should raise the number of moments in which a person can afford to be present. It should not become a cheap skin stretched over an absence.

So run the test in your own business, tonight. What does it cost your customer to reach a human being with the authority to change an outcome? If the honest answer is “a series of polite emails plus a meltdown on social media,” you do not have a service design. You have a class system wearing a chatbot as a uniform.

Then notice the opening, because it is wide open and it is underpriced. In nearly every industry where everyone has installed a machine in the front door, a firm can win on a single unfashionable, expensive promise: when it matters, a named human will show up. Not a better chatbot. A person who knows your file, your family, your deal. That promise cannot be white-labeled, and the competitor who spent their budget on surfaces cannot easily copy it — because the cost of the promise is exactly the cost they just eliminated.

Every automated surface you ship answers a question your customers instantly understand: who gets a person? The answer is either a promise or an insult. They will know which one within thirty seconds — and they will remember it for the rest of their lives.