I Tried 4 Chatbot Tools and My Customers Hated All of Them
I own a property management company. We manage about 85 residential units. Tenants call and message about everything. Maintenance requests, lease questions, lockouts, noise complaints, payment issues. The volume is constant and i cant answer every call in real time.
So two years ago i started looking into chatbots. Everyone was telling me it was the future of customer service. Install a chatbot on your website, put one on Facebook Messenger, let it handle the basics. Sounded great.
I tried four different tools over 18 months. My tenants hated every single one. Not disliked. Hated. I got more complaints about the chatbots than I did about maintenance delays. And maintenance delays are my tenants number one complaint.
What went wrong (four times)
Tool 1: A basic decision-tree chatbot. You know the kind. "What can I help you with? Select from: Maintenance, Payments, Lease Questions, Other." Then it branches into more menus. Then more. By the time the tenant gets to the right category they've clicked through 5 screens and still end up at "please email us at support@."
Tenants told me it felt like a phone tree but worse. At least with a phone tree you can mash 0 to get a human. The chatbot just kept looping.
Tool 2: A slightly smarter NLP chatbot. This one could understand typed messages instead of just menus. But its understanding was, lets say, limited. A tenant typed "my toilet is running all night" and the bot responded with "I see you're asking about running. Here are some nearby running trails!" I wish i was making that up.
Tool 3: A well-known customer service platform chatbot. Better tech, better responses. But it took 40+ hours to set up properly, required constant training on new questions, and still fell apart on anything that wasnt in its training data. Which was most things, because tenants don't ask questions in predictable ways.
Tool 4: An AI-powered chat widget from a startup. The best of the four by far. Actually understood context, could hold a real conversation. But it was text only and my tenants wanted to call. Most of my tenants are between 28-55 and when they have a problem (especially an urgent one like a leak or lockout), they pick up the phone. They don't want to type.
The data on chatbot satisfaction is grim
Turns out my experience is not unusual. A 2024 survey from Salesforce found that 46% of consumers have a negative perception of chatbots in customer service. Nearly half.
The complaints are consistent across studies:
- "It doesn't understand my question"
- "I can never get to a human"
- "It gives generic answers that don't help"
- "It's frustrating when I have an urgent issue"
A Gartner study found that while businesses are deploying chatbots at increasing rates, customer satisfaction with chatbot interactions has actually declined over the past three years. More chatbots, worse experience. Thats not the trajectory you want.
And the most damning stat: according to CGS research, 86% of consumers say they prefer to interact with a human over a chatbot. When given the choice, people overwhelmingly choose a real voice or at minimum a channel that feels human.
Why chatbots fail for small businesses specifically
Big companies can make chatbots work because they have the resources to build, train, and maintain them. Amazon's chatbot works reasonably well because they've poured millions into it and they handle the same 20 questions millions of times.
Small businesses don't have that luxury. Your question set is diverse, your customers are unique, and you dont have a team of engineers tweaking the bot every week. So what happens is:
- You spend hours setting up the chatbot
- It works okay for the first month
- Customers start asking questions it can't handle
- Frustration builds
- You stop maintaining it because you're busy running the business
- It becomes a worse experience than having nothing at all
The other problem specific to small businesses is that your customers often have a personal relationship with you. They're used to calling and talking to a person they know. Suddenly they're talking to a robot that can barely understand them. It feels like a downgrade. Like you stopped caring about service quality.
One of my tenants literally said "I pay $1,800 a month in rent and I cant even talk to a person anymore?" That one stung.
What customers actually want
After the chatbot fiasco I did something old fashioned. I asked my tenants what they actually wanted. Not what technology I should deploy. What experience they wanted when they had a problem.
The answers were really consistent:
- Answer the phone. Dont make me navigate menus or type into a box. Let me talk.
- Understand my problem quickly. Dont ask me 12 qualifying questions. Just listen and respond.
- Give me a real answer or a real timeline. "Someone will get back to you" is not an answer.
- Be available outside business hours. Problems dont only happen 9-5.
Notice what's missing from that list? Nobody said "I want a chatbot." Nobody said "text-based AI on your website would be great." They wanted a phone answered by something that could actually help.
The phone-first approach
After failing with chat-based tools I switched strategies entirely. Instead of trying to deflect customers to a chatbot, I focused on actually being reachable by phone.
Thats actually why I built AgentErgon. Its a voice AI phone system, not a text chat. When a tenant calls and I cant pick up, the AI answers. The tenant talks, the AI listens, asks clarifying questions, and either provides an answer from the knowledge base or creates a ticket and tells the tenant what to expect.
The reception was completely different. Same AI technology, fundamentally different experience. Tenants who hated the chatbot were fine talking to the AI on the phone. A couple didn't even realize it wasn't human until I told them.
I think the difference is that phone conversations feel more natural and less "techy." People talk on the phone every day. Chatting with a bot on a website feels like interacting with a vending machine.
What I'd tell another small business owner
If your thinking about chatbots, ask yourself these questions first:
How do your customers prefer to reach you? If they mostly call, a chatbot is solving the wrong problem. If they mostly message or use your website, a chatbot might make sense.
How complex are their questions? If you get the same 5-10 questions repeatedly, a simple FAQ page might be more effective than a chatbot. If questions are diverse and contextual, you need something smarter than a decision tree.
Do you have time to maintain it? A chatbot that isnt regularly updated becomes worse over time, not better. If you cant commit to reviewing conversations and improving the bot monthly, it will eventually become a liability.
What happens when the bot fails? Every chatbot fails sometimes. Whats the escalation path? If "sorry, I cant help with that" is a dead end with no way to reach a human, your customers will be furious. The off-ramp matters more than the bot itself.
For most small service businesses, honestly, the answer isnt a chatbot. Its figuring out how to be reachable and responsive through whatever channel your customers already prefer. Usually thats the phone. And there are better solutions for phone reachability than there are for chat.
The 46% of consumers who hate chatbots arent wrong. They're just telling you what they want. And what they want is pretty simple: talk to someone who can help, right now.
Top comments (0)