I can’t share the client project that inspired this story, so I’ve created a fictional account to explore a similar situation. The business, conversations, and results below are illustrative.
The owner of a small flower and garden shop two towns over asked me to come in as an AI consultant. When I arrived, she turned her laptop toward me and pointed at a string of customer emails. “My team has done a great job with AI. We’re answering people faster than ever,” she said. “Now I’d like them to come back.”
“Come back?” I asked.
She shared how the team had AI handle common questions and follow-up after a sale. This began soon after the exhausted owners realized how many evening hours they spent catching up on emails.
AI suggested ready‑made replies, links to care guides, and a schedule of sales messages. The team checked the plant advice and put the plan to work. Answers went out faster, and fewer messages sat waiting overnight.
But six months later, the owners noticed that some familiar customers had stopped coming in. When they dug into the “why,” the answers hurt more than a few empty parking spaces.
One woman said she used to feel comfortable asking questions, even the ones she thought were silly. Lately, the answers were matter-of-fact and sent her toward another purchase. She realized the shop’s local, personal connection had been replaced by commodity sales. So, she focused on costs and found lower prices elsewhere.
What Customers Came For
The owner told me about a man who arrived each spring convinced he could kill a plastic fern. Her staff helped him grow his first tomatoes. He brought in a photograph of them as if they’d just graduated from college.
Then there was the woman who wanted roses for a shady corner. A staff member walked her through other choices, even though the roses cost more. She returned the following week with a neighbor.
As the owner talked, her face changed. These were the parts of the business she loved. She could tell you who grew those tomatoes, where they planted them, and which staff member had helped.
“Did any of that make it into your conversation with AI?” I asked. She looked back at the screen and shook her head.
The team had explained the work they wanted to finish. They had left out why customers trusted them to do it.
An accurate care guide answered a question about a plant. Yet the person asking might also need to hear that one yellow leaf didn’t mean they had failed at gardening. The old conversations had made room for knowledge transfer, confidence-building, and the making of green thumbs.
Give AI Your Perspective
I shared a simple framework I call “Give AI Your Perspective.” Before asking AI to solve a project, explain what matters, describe what must remain, and define success.
We started with what matters. The owner wanted evenings back, but she also wanted people to enjoy growing things. A faster reply mattered because it could help a worried customer sooner and leave her staff free to give someone in the shop their full attention.
“We help people feel they can do this,” she said. That gave us a much better place to begin.
Next came what must remain. “Friendly service” sounded fine until I asked what that looks like on a Tuesday afternoon. She laughed, then gave me the examples we needed.
Staff asked about light and space before suggesting a plant. They welcomed beginner questions and sometimes told people they didn’t need to buy anything. A customer with a sick plant could reach a person instead of working through a pile of links.
Finally, we defined success. We asked AI to draft three short follow‑up emails and a simple way to sort incoming questions. The emails needed to offer useful care advice without pushing a sale, welcome a reply, and sound like something her staff would say, not a botanist.
The sorting process placed uncertain answers in the hands of a staff member. We also asked AI to flag any idea that might weaken the customer’s connection with the shop.
Now AI had a purpose, examples, and a clear job. The owner had a way to judge its answer beyond whether it sounded polished or overly professional.
The Humanizing Tweak
Two weeks later, I was called back in to help refine the AI’s process. The emails were much closer to being human, but the last lines seemed to be that used car salesman’s pitch to buy more. It told the reader they weren’t cared for, but used to make the shop money.
The owner asked me to address the issue, holding up a sample email. “We’ve just asked how their plant is doing. And a couple of lines later, we’re pushing for another sale. Could we wait for the answer before selling them something?”
We made a simple adjustment. The second version gave the person the answer they needed, plus a bonus tip. The sales pitch was gone.
I also created an AI avatar who understood the owner’s love for plants. It quickly sorted the emails against sample questions and answers. A simple request for store hours could get a simple, quick answer. A message saying, “I followed your advice, and my plant is dying,” needed someone to ask more personal questions and provide specific help.
We shifted from clear answers to focused answers.
After a small trial, the revised process proved fully functional and effective. Staff spent less time repeating basic information and had more room for the conversations customers valued. Helpful replies came back with photos of plants doing well and their happy owners.
Final Tips for Your Business
Before your next AI project, try finishing three thoughts: “This matters because…” “We must keep…” “A good result will look like…” Give each one enough detail that someone outside your world could understand it.
Then explore the answer. Picture the person who will receive that email, use that plan, or live with that decision. What might they need that your first request left out?
AI is a great tool when humans bring passion, purpose, and meaning to the work they do together. Those qualities need a place in the conversation before we let a tidy answer change the way we serve people.
The garden shop owner had a photograph of a proud man holding two tomatoes. That picture told her more about success than the number of emails her team could clear before lunch.
Copyright 2026 by CJ Powers
