Instant is the new baseline

A customer messages two companies the same evening with the same question. One answers in nine seconds, a real answer, not a ticket number, with the right price and the next step. The other sends an auto-reply: thanks, we’ll get back to you within one business day. By morning the first company has the deal half-closed and the second one is still in a queue. Nobody decided this race. The customer just noticed who showed up.

That gap used to be invisible. Both companies “responded,” after all, one fast, one slow, both inside what we politely called normal. The thing that changed isn’t that fast got faster. It’s that slow started to feel broken.

When the agent on the other side never sleeps, “we’ll get back to you” stops being polite and starts being a confession. This post is about why the customer’s clock just reset to now, and what that does to every company still batching its replies.

The expectation didn’t drift. It snapped.

Here’s the part that’s easy to miss if you only watch your own company. Customer patience doesn’t degrade slowly, a little each year, the way we like to imagine. It resets the moment one provider in someone’s life answers instantly, and then it never goes back.

A person doesn’t have a separate patience budget for each company they deal with. They have one. The bank that approves a card in two seconds, the food app that re-routes a late order before they complain, the airline that texts the gate change before the board updates, every one of those teaches the same lesson, and the lesson transfers. By the time that person reaches your contact form, they are not grading you against your competitors. They are grading you against the fastest thing that happened to them all day.

When the agent on the other side never sleeps, “we’ll get back to you” stops being polite and starts being a confession. The customer hears: we are not staffed for you right now. They are not wrong. The auto-reply is true. It just used to be acceptable and isn’t anymore, because the comparison set changed under your feet.

The naive read of this is “people got more impatient.” That’s the version that lets you off the hook, it makes the customer the problem, and the fix is to ask them to wait nicely. The honest read is the opposite. People didn’t get impatient. Some companies got instant, and instant, once tasted anywhere, becomes the floor everywhere. The baseline moved. You’re being measured against it whether or not you opted in.

Why “we’ll get back to you” was always a staffing artifact

Step back and ask why the delay existed in the first place. Not why it’s bad, why it was ever there.

The naive story is that good answers take time. Thoughtful work can’t be rushed; a careful reply is worth the wait. That story flatters us, and for the genuinely hard questions it’s even true. But it’s not why most replies are slow. Most replies are slow for a far less noble reason: there is a human, the human is busy, the human is asleep, the human is in a meeting, the human has forty other messages, and yours is number forty-one.

The delay was never about the difficulty of the answer. It was about the availability of the answerer.

That’s the whole mechanism. “We’ll get back to you” is a queue, and a queue is what you build when demand arrives faster than humans can serve it. The phrase isn’t a service standard. It’s a confession about throughput dressed up as courtesy. We batched responses because batching was the only way a finite team could survive an infinite inbox, answer in clumps, twice a day, when someone finally had a free half-hour.

For decades that was simply the cost of doing business, and everyone paid it, so nobody lost by paying it. The customer waited because every company made them wait. The queue was universal, which made it invisible. You can’t feel slow when the entire market is the same speed.

The instant company breaks that symmetry. It doesn’t have a human at position one of the queue, because it doesn’t have a queue. It has a system that read the question, knew the context, and answered while the customer was still on the page. Now the wait isn’t universal anymore, and the moment it isn’t universal, it’s a disadvantage with your name on it.

The trap: speed without the company behind it

Now the obvious objection, and it’s a good one. If instant is the new baseline, can’t anyone just bolt a chatbot onto the website and call it solved?

This is where most “we added AI” projects quietly fail, and it’s worth being precise about why. Fast and wrong is not a small improvement over slow and right. It’s often worse. A reply that arrives in two seconds and quotes the wrong price, promises a feature you don’t have, or cheerfully invents a policy is not a faster service. It’s a faster way to lose trust. The customer wanted now, yes, but they wanted now from you, with your facts, your prices, your commitments. Not now from a stranger wearing your logo.

The naive version of going instant is a generic model with a friendly tone and no idea who it works for. It’s quick because it’s empty. Ask it the one question that actually matters, can you do X for my specific situation by Friday, and it either guesses or stalls, and now you’ve spent your one shot at speed on an answer the customer can’t act on. Speed without grounding doesn’t beat the slow human. The slow human at least knew the price.

The version that works is slower to build and instant to use, because the speed comes from a place the chatbot doesn’t have: the company’s own knowledge, already assembled, already current. The answer is fast because the system already knew the price, already knew this customer’s last order, already knew which of three plans they’re on and what’s promised in the contract. It isn’t generating a plausible reply. It’s reading the company and telling the truth in nine seconds instead of nine hours.

Instant only counts when the instant answer is also the right one.

That’s the distinction the market is about to enforce, hard. The first wave of “we got fast” will be full of confident wrong answers, and customers will learn, quickly, because they learn everything quickly now, to distrust the fast company that’s fast and hollow. The durable advantage isn’t a chatbot. It’s a system grounded in what the company actually knows, answering at the speed the customer now expects, without making things up to hit the clock.

The asymmetry: nights, weekends, and the question at 11 p.m.

There’s a second edge to this that’s easy to underrate, because it doesn’t show up in your average response time. It shows up in the responses you never got the chance to make.

Think about when customers actually reach out. Not at 10 a.m. on a Tuesday when your team is staffed and sharp. They reach out when they’re thinking about you, which is often the evening, the weekend, the quiet hour after dinner when they finally have time to consider the purchase or stew on the problem. The message lands at 11 p.m. The human who could answer it is, correctly, asleep. So the question waits until morning, and by morning the moment has cooled. The customer who was ready to buy at 11 is browsing a competitor at noon.

The naive fix is to staff the nights, a graveyard shift, an offshore team, someone always awake to catch the late questions. It’s expensive, it’s hard on people, and it still has a queue. You’ve moved the staffing artifact, not removed it. The 11 p.m. question still waits behind the 10:58 question, and the person answering at 3 a.m. is, understandably, not at their best.

A system that never sleeps doesn’t have a graveyard shift, because it doesn’t have shifts. The 11 p.m. question gets the same nine-second, grounded, correct answer as the 11 a.m. one. Say a meaningful share of your inbound arrives outside business hours, for many companies it’s most of it, because that’s when humans are free to be customers. Every one of those messages used to age overnight. Now none of them do. That’s not a small efficiency. That’s the difference between catching a customer at the peak of their intent and catching them after it’s passed.

This is the asymmetry that compounds. The instant company isn’t just faster during the day. It’s present during the hours its competitors are dark, and those hours are exactly when the customer was paying the most attention.

The turn: this is about the customer’s time, not yours

Here’s the part that isn’t about software at all.

We keep framing speed as an operational metric, response time, SLA, tickets-per-hour, as if it were about us, our efficiency, our throughput. It was never about us. The customer’s clock is the only clock that matters, and what the customer is really telling you when they expect an instant answer is something more human than impatience. They’re telling you their time is worth respecting. The wait was always a small tax we levied on them so that we could batch our lives more comfortably. They paid it because they had no choice. Now they have a choice.

When you answer in nine seconds with the right answer, you’re not showing off your stack. You’re saying: your question mattered enough that we were ready for it. That’s the thing the customer actually feels, not “this company is fast” but “this company was waiting for me.” It’s the same feeling Donna gives Harvey when she knows the thing before he asks. It reads as care, because the work of being ready was done before it was needed.

The companies that win the next few years won’t win because their model is smarter. They’ll win because they stopped making the customer wait for a human to come free, and started letting the customer’s question reach something that already knew the answer and never had to be woken up. Instant isn’t a feature you add. It’s what respect for someone’s time looks like when the technology finally allows it.

---

That’s what we’re building at Apollo, not a faster way to send the auto-reply, but a company that’s already ready when the customer arrives, grounded in what the business actually knows, awake at 11 p.m. so a person doesn’t have to be. “We’ll get back to you” was always a promise about our schedule. The customer was only ever asking about theirs.