Use AI and Automation to Elevate Humans. (Not Replace Them.)

by Jean-Eric Plamondon | 13 Aug, 2025 | Software & Automation

The short answer: Use AI to support people so they can deliver an even better experience. Start with triage, intake and data gathering, where AI can support your humans, and let customers know when they’re talking to AI. Done well, that means tickets resolved faster, hours of time freed up and improved customer service.

People keep asking me about AI—how it’s changing business, what to do with it, and whether we’re all about to be replaced by robots.

From my perspective, the question shouldn’t be, “How do we use AI to replace people?”

It should be, “How do we use AI to support people so they can deliver an even better experience?”

Most businesses don’t need a robot that can do everything.

They need better systems that make it easier for the humans on the team to perform at a higher level.

Let me give you two real-life examples.

Bad Automation: Replacing an Entire Support Team with AI

I’ve been a Fido customer for 15 years.

I stuck with them because their customer service was actually good. You could get a real person on the phone, have a conversation, and walk away with your issue resolved. It was a breath of fresh air in the whole Canadian cell phone market.

Now? It’s all AI.

You literally can’t get a human anymore. No matter what you press or say, the best they’ll do is promise a call back in three to five days—on one of three time slots they give you.

That’s it. People are losing their minds. Reddit forums are full of rants. Customers are bailing.

That’s what happens when you try to automate end-to-end.

The experience goes straight into the gutter.

Good Automation: Supporting People, Not Replacing Them

Compare that to what we did with one of my clients.

They handle 60 to 100 service tickets a day, and the old way was chaos: eight emails back and forth, customers getting frustrated, missed messages, duplicate tickets, angry follow-ups.

So, we looked at where we could use AI to help the team—not replace them. Here’s what we did:

Automation to Get Feedback and Improve Training

Every time a ticket is marked “closed/complete,” we automatically send a quick NPS survey.

If the customer gives a 5/5 rating, we flag that interaction and use it to train the next agent.

Instead of having someone sit down to program responses, we’re using real conversations to train AI on what great looks like.

Off-Hours Support and Smart Triage

His team doesn’t answer phones for 14–16 hours a day.

So we built an AI layer that says, right up front, “Hey, I’m an AI agent. Our office is closed, but here’s what I can do…”

Then it gives people options:

• Leave a voicemail
• Call back tomorrow
• Or give some info now and we’ll put you at the top of the list for tomorrow

Then we collect:

• Order number
• Order date
• Customer name
• Reason for the call

And we could even go a step further: ask 2–3 follow-up questions like, “Have you checked your tracking?” or “Do you know the standard delivery time for your region?”

That AI isn’t solving the problem, but it’s gathering information.

So, the next day when a human support rep picks up the ticket, they hit the ground running.

The Payoff: Tickets Resolved Faster, Hours of Time Freed Up, Improved Customer Service

This isn’t magic. But it is a big deal.

If we reduce the average interaction from 8 emails to 3, that saves 30–60 minutes per ticket.

Multiply that by 60 to 100 tickets a day? That’s hours and hours freed up.

It also cuts down on the back-and-forth that causes customers to drop off or get angry.

The quality goes up. The frustration goes down. And the team isn’t burnt out dealing with an overflowing inbox.

4 Tips for a Successful AI Starter Project

If you’re a business owner thinking about AI, here’s what I want you to remember:

1. Don’t try to automate everything. That’s where companies go wrong. Start with triage, intake, data gathering—places where AI can support your humans, not replace them.
2. Train with real success. Use your best interactions to teach your team and your AI what “great” looks like.
3. Be transparent. Let customers know they’re talking to AI. Give them options. Don’t pretend it’s a human—just make it useful.
4. Focus on the experience. Automation should make life better—for both your team and your customers. If it doesn’t, it’s not worth it.

The real leverage doesn’t come from replacing your workforce.

It comes from giving them superpowers so they can do what they’re best at. Faster. With less friction, and a lot more consistency.

If you’re not sure where to start, just ask: What part of our customer experience is repetitive, predictable, and slowing us down?

That’s probably where your first AI win is hiding.

Common questions

What happens when a company replaces its support team with AI?

Look at Fido. You literally can’t get a human anymore. The best they’ll do is promise a call back in three to five days, and customers are bailing. That’s what happens when you try to automate end-to-end.

How can AI help a customer support team?

For one client handling 60 to 100 service tickets a day, every closed ticket triggers a quick NPS survey, and 5/5 interactions are used to train the next agent. After hours, an AI agent says up front that it’s an AI, gives callers options and collects details like order number and reason for the call, so the human rep can hit the ground running the next day.

How much time can AI-assisted support save?

If the average interaction drops from 8 emails to 3, that saves 30 to 60 minutes per ticket. Multiplied by 60 to 100 tickets a day, that’s hours and hours freed up.

Should customers know they’re talking to AI?

Yes. Be transparent: let customers know they’re talking to AI, give them options, and don’t pretend it’s a human. Just make it useful.