AI Is Incredible. Most Agents Are Using It to Become Worse at Their Jobs.
AI is incredible. Most agents are using it to become worse at their jobs.
That sentence sounds backward. It isn't.
The agents losing referrals right now are not the ones ignoring AI. They're the ones who adopted it fast, applied it everywhere, and automated the exact things that made clients want to send them business in the first place.
Two Agents. Same Tools. Different Results.
Agent A gets a deal under contract. She fires off an AI-generated "congrats" text to 200 people in her database. Schedules automated check-ins for months two, four, and six post-close. Sets up a drip campaign for anyone who "goes cold." Then moves on.
Agent B gets the same deal under contract. She uses AI to pull up every past client who bought or sold in that neighborhood. She uses AI to draft a personal note — then rewrites it in her own words. She picks up the phone and calls three of them.
Both agents used AI today. One built a referral relationship. One eroded one.
The Data Is Not Subtle
NAR's numbers are consistent year over year. 66% of sellers choose an agent they already know or were referred to. 81% contacted only one agent before signing.
Read that again. 81% didn't shop around. They already knew who they were calling.
The question is not whether AI can help you reach more people. Of course it can. The question is whether the people you're reaching still feel like you actually know them. Because that feeling is the entire business.
What AI Is Being Used For (Wrong)
Here's where most agents point their AI tools:
- Automated birthday and anniversary texts
- Drip campaigns to "stay top of mind"
- AI-generated market update emails sent to entire lists
- Chatbots that respond to inbound leads
- Social posts generated from templates
All of these replace human contact. None of them create it.
Your past clients are not stupid. They know when a message was written by software. They know when a check-in was triggered by a calendar event. They've been getting these emails from every service business they've ever used. The messages don't make them feel remembered. They make them feel processed.
What AI Should Actually Be Used For
Here's where Agent B points the same tools:
- Pulling triggers from the database: who just had a life event, who bought two years ago, who's been quiet
- Drafting a first pass of a personal message — that the agent then actually personalizes
- Researching a neighborhood's market before a call so the conversation is sharp
- Summarizing a transaction timeline so the agent can speak fluently to a referral source
- Handling every admin task so there's more time to call
The difference: AI does the logistics. The agent shows up for the relationship.
The Real Cost
The agents who automate their relationships don't lose business all at once. They lose it slowly. A past client sells and uses someone else "because they happened to be at an open house." A sphere contact refers a colleague to a different agent "who'd been in touch lately." The referrals stop coming and there's no single moment to point to.
That's the insidious part. Automated relationship management doesn't feel like failure. It feels like efficiency. Right up until the pipeline runs dry.
"AI reminded you to call. Now call."
That's the whole framework. AI is the system that tells you who needs attention, when, and why. You are the person who shows up.
The Question You Need to Answer
Think about the last 10 referrals you received. How did those people feel about you? Was it because of an automated email sequence, or because of a real conversation you'd had?
You already know the answer.
The question isn't whether to use AI. The question is what you point it at.
Point it at admin. Point it at research. Point it at logistics. Let it handle the work that doesn't require you.
Then take the time it frees up and use it to call someone.
Pick a side.