Everyone's Asking What AI Changes. Ask What Won’t Change.

People would often ask Jeff Bezos what would change in the next 10 years. Almost nobody asked what wouldn't. Bezos said that the second question was “actually the more important of the two,” because you can only build a strategy around what will stay true.

Bezos knew customers would always want things faster and cheaper, so Amazon invested in speed and price instead of guessing about the future. Twenty years later, that's still where Amazon’s value comes from.

The loud question and the useful question

In windows and doors, every conversation right now is about what AI will change. What will happen to my CSRs? Will this tool replace my team? Fair questions, but they're the loud questions, and the loud ones are rarely the useful ones.

What isn't going to change

Nobody will spend five figures on their house without letting a person into their living room.

You can generate the lead, book the appointment, even show a homeowner a rendering of their house, all online. The decision still gets made at a kitchen table, across from a person. That isn't going to change.

 The installation requires people in someone's home for days.

That's not a step in the process. It's the value. Homeowners aren't just buying windows; they're buying an install done right the first time, without their week falling apart.

 Trust is still local.

It comes from the neighbor two doors down, the yard sign, the truck people recognize. Nobody can manufacture that, and the companies that try end up sounding like everybody else.

 When it goes wrong, somebody has to own it.

Service is a relationship, not a workflow. The moment a homeowner feels like they're talking to a system instead of a person, you've lost the referral and probably the review.

The measurement has to be right.

Get it wrong and you pay twice: once in service, once in the customer's patience. Software can help you catch and fix errors fast. It can't take responsibility for them.

 Finding and keeping your talent.

This is the one nobody says out loud. AI doesn't install windows, train an apprentice, or turn down a job across town to stay with you. Your growth is still capped by how many good people you can attract and keep. That was true ten years ago and will be ten years from now.

Automate what gets in the way

First, identify your highest-impact areas and the things that won’t change anytime soon. Then automate everything that gets in their way.

Most operators do it backwards. They get excited about a tool, bolt it onto whatever's already running, and automate the parts that didn't need help while the real constraints sit untouched.

Pointed at the right things, technology improves the customer experience: fewer mistakes, better communication, faster answers. Lean on it too hard and the homeowner feels it. Keep the customer at the center of every decision, technology or otherwise, and you’ll navigate the change fine.

Where AI is earning its keep right now

1.      After hours and overflow.

Homeowners call after work, fill out forms at nine at night, and call while your CSR is on another line or at lunch. If nobody picks up, they call the next company. That's not a lead generation problem. It's a leak in leads you already paid for, and the easiest fix on this list.

2.      Follow-up that doesn't quit.

Nobody runs a consistent follow-up cadence by hand, not because your people are lazy but because they're busy. Automated follow-up and rehash on unsold appointments is unglamorous, and it works.

3.      Coaching from what reps actually said.

Ride-alongs don't scale, and most managers coach from memory and anecdote. Recording and reviewing real in-home conversations makes coaching continuous instead of something that happens when a manager has a free Tuesday. Rilla and Siro both have exceptional products here.

Realistically, this technology will save someone fifteen minutes, a few times a day. That sounds small until you compound it across a team over a year. That's the whole game right now. Headlines will make you feel behind the AI curve, but you’ve got time. Small changes made slowly beat no changes at all.

What we’d tell a friend

Three things I'd tell a friend running one of these businesses before they adopt new technology.

1.      Adoption is the hard part, not picking the tool.

Getting the tool working takes a few weeks. Getting your people to use it takes a few quarters. Budget your energy accordingly, and make clear what’s in it for them: an easier day, better performance, or both.

2.      It will expose your standards (or lack thereof) before it fixes anything.

If your qualification rules or sales process live in one person's head, or your follow-up cadence is just whatever your best rep happens to do, you'll find out fast. It may be uncomfortable. That's not a reason to skip it. Just know the first thing you get is a mirror.

3.      Build versus buy?

Buy. Start by starting. Explore a couple off-the-shelf tools, ask a friend what they use, and go.

One last thing

Don't worry about a robot showing up to install windows. Worry about the fifteen minutes a day your best people lose to work a system should be doing, and about those six things quietly slipping while everyone stares at the loud question.

People still want their packages delivered ASAP.