Do you ever secretly suspect that the world has collectively agreed to stop listening to you, and that the machines we’ve built to ‘assist’ are actually just digital bouncers paid to keep you outside the club?

It is a quiet, modern horror. Helen is currently living in it. She is sitting at her kitchen table, the air still heavy with the charcoal-scented ghost of a rosemary chicken dinner I managed to incinerate forty minutes ago because I was too busy listening to her read chatbot transcripts aloud.

Helen’s mother is 74. She has a spirit that wants to see the colonial architecture of Antigua, Guatemala, but she has knees that have reached a definitive, non-negotiable treaty with gravity.

Helen has a simple, binary question: Is the hotel’s “accessible” entrance located on the side street with the 300-year-old jagged cobblestones, or does it have a paved drop-off point?

The Loop of Uselessness

For the fifth time, Helen types this into a small, floating bubble on her screen. For the fifth time, a digital avatar named “TravelBot” (represented by a generic blue-and-white icon that looks suspiciously like a smiley face with a lobotomy) responds in less than two seconds.

Neither is true. “No” doesn’t accurately describe the failure; it implies the bot tried and missed by an inch. The reality is that the bot isn’t even playing the same sport.

But if Helen clicks “No,” the bot simply cycles back to the start, offering a slightly different menu of uselessness. If she clicks “Yes,” the ticket closes, a KPI is met in a gleaming glass office in a different time zone, and the system records a successful “self-service resolution.”

The Architecture of Deflection

There are nine specific tiers of linguistic evasion documented in the Standard Taxonomy of Computational Linguistics, and the modern customer support bot has mastered every single one of them.

These systems are not designed to find the answer; they are designed to exhaust the asker. They are the digital equivalent of a labyrinth where the walls shift every time you get close to the exit, ensuring that you eventually sit down, give up, and stop costing the company money.

I used to believe that efficiency was the highest form of service; I was profoundly mistaken about the nature of human need. Several years ago, during my brief and ill-fated stint trying to optimize workflow for a logistics firm, I argued that human interaction was a “friction point” that could be engineered out of existence.

I viewed every phone call as a failure of the interface. I was wrong. I failed to understand that when a person reaches out with a specific, nuanced problem-like a mother’s mobility in a foreign city-they aren’t looking for a data point.

They are looking for a witness. They are looking for the assurance that someone on the other end understands that if the answer is wrong, the vacation is over before it begins.

Cruelty in the Cloud Forests

The core frustration of the modern consumer isn’t that technology is broken. It’s that technology is working perfectly for the wrong people. Support automation is sold to us as a way to get “faster help,” but its actual architectural goal is deflection. It is a financial firewall.

Every second a human agent spends talking to Helen costs the company roughly $1.64. Every hour a bot spends looping Helen through FAQ links costs the company less than a fraction of a penny. When you scale that across 40,000 customers, the “success” of the bot is measured by how many people simply close the window in disgust.

In the world of high-stakes travel, this deflection is particularly cruel. If you are planning a trip to a place like the Galápagos or the high Andean cloud forests, the variables are not flat. They are three-dimensional, breathing, and often unpredictable.

A bot can tell you the average temperature in Quito in October (it’s 58 degrees Fahrenheit), but it cannot tell you if the specific incline of a boat ramp will cause your mother’s breath to catch in her throat. It cannot tell you if the guide speaks the kind of English that includes the nuances of local history or just the kind of English that identifies the locations of the bathrooms.

The Radical Act of Rebellion

This is why the model at Osaviva Travel feels like such a radical act of rebellion in the current landscape.

It is a business built on the very thing that automation is trying to kill: the expensive, time-consuming, deeply human conversation. While the rest of the industry is trying to hide its humans behind layers of “AI-driven solutions,” some have realized that the human touch is the only thing that justifies the premium.

The metric of a “closed ticket” is a lie. If a problem disappears from a dashboard but remains entirely intact in the customer’s life, the system has not solved anything; it has simply outsourced the labor of the problem back to the person paying for the service.

It is a shell game where the pea is your own peace of mind.

Thomas H.L., a man I know who spends his days as a sand sculptor on the coast of Oregon, once told me that you cannot automate the “feel” of the sand. You cannot build a sensor that replaces the intuition of a man who has spent 30 years watching the tide.

Planning a meaningful journey is a lot like sand sculpting. It requires a specific “moisture content” of human connection. When you remove the human hand from the process, the structure might look fine on a screen, but the moment you put any weight on it-the weight of a specific mobility need, the weight of a once-in-a-lifetime anniversary, the weight of a complex family dynamic-it crumbles.

The bot tells Helen that the hotel is “ADA compliant,” a term that has a very specific legal meaning in the United States but is often used as a vague marketing tag in international booking platforms. It means nothing when applied to a 400-year-old convent turned boutique hotel in the heart of a UNESCO World Heritage site.

Helen needs to know about the two steps leading into the breakfast room. She needs to know if the elevator is large enough for a modern motorized chair or if it was built in 1974 for a single person and a suitcase.

The bot cannot see those two steps. The bot doesn’t know that those two steps represent the difference between Helen’s mother feeling like a guest and feeling like a burden.

The Great Deflection

We are currently living through a Great Deflection. We see it in banking, in healthcare, and most visibly in the way we move across the world. We are being trained to accept “good enough” as the default.

We are being told that a link to a generic PDF is the same thing as an answer. But the more we automate the “easy” parts of life, the more we realize that the only parts that actually matter are the “hard” parts-the parts that require judgment, empathy, and local expertise.

The ticket is marked resolved the moment the mother’s wheelchair hits the first insurmountable step.

There is a profound arrogance in the belief that we can replace the complexity of human experience with a decision tree. When I burned that chicken dinner, it wasn’t because the stove failed.

The stove did exactly what it was programmed to do: it provided heat. It was a perfect machine. But it lacked the “judgment” to know that the meat was drying out or that the garlic was turning bitter. It didn’t care that the meal was intended to be a gesture of care for a friend who was stressed. It just followed the settings.

The Cost of Erosion

An automated travel platform is that stove. It will “cook” your itinerary. It will provide the flights, the rooms, and the dates. But it has no capacity to care if the result is bitter.

It has no way to sense that the pacing of the trip is too fast for a 74-year-old, or that the “boutique” hotel is actually located directly above a late-night salsa club.

The true cost of the chatbot isn’t the frustration of the moment; it is the erosion of trust. Every time Helen is told that her question has been “resolved” when it hasn’t, she loses a little more faith in the possibility of a seamless experience. She starts to expect the friction. She starts to plan for the failure.

In a world of high-speed, low-touch transactions, the highest form of value is proximity. Proximity to the truth, proximity to the destination, and proximity to a person who has the authority to say, “I will find out the exact height of those steps for you.”

That single sentence is worth more than a thousand “optimized” FAQ pages. It is the sound of a human being standing in the gap between a corporate dashboard and a real-world problem.

Helen eventually closed the laptop. She didn’t click “Yes” or “No.” She simply sat in the quiet of the kitchen, the smell of burnt rosemary still lingering, and realized that she was trying to buy a memory from a machine that doesn’t know what a memory is.

We have designed a world that is very good at closing tickets but very bad at solving problems. We have built systems that prioritize the silence of the customer over the satisfaction of the human.

And as we continue to push the “human cost” off the balance sheet, we are finding that the remaining product is a hollow shell-efficient, scalable, and utterly devoid of the one thing that makes travel worth the effort: the feeling that you are being looked after by someone who knows the way.

If the goal of the bot is to make you give up, then the only way to win is to refuse the premise. To seek out the “expensive” conversation. To find the people who haven’t yet outsourced their brains to a decision tree.

Because at the end of the day, when you are standing on a street corner in a city you don’t know, you don’t need an algorithm. You need a person. And you deserve to find one before the ticket is closed.