← All HVAC articles Case Studies / Demos (15)

How AI Chat Handled a 2 AM Emergency Call

The 2 AM Call

It was 2:14 AM on a Saturday in August. A homeowner in San Antonio woke up to a sweat — their AC had stopped completely. The indoor temperature was already 87 degrees and climbing.

They grabbed their phone and searched "emergency AC repair San Antonio." The first three results were all franchise shops with 24/7 call centers. They picked the third one and called.

Voicemail.

They called the second one.

Voicemail.

By the time they got to the fourth result — a local 3-tech shop with a CoolerDesk site — they were frustrated and desperate.

What Happened Next

The fourth result had a chatbot. It opened with: "I see you're looking for emergency AC repair. Can you tell me what's happening?"

The homeowner typed: "My AC is dead. It's 87 degrees inside. I have a baby in the house."

The chatbot responded within 2 seconds:

> "I'm sorry to hear that. Let me get a few details and send help immediately."

2:16 AM — The bot collected the address, phone number, and a quick description of the issue.

2:16 AM — An SMS alert fired to the shop owner's phone: "URGENT: AC completely dead. Temp inside 87°F. Infant in home. Address: 123 Main St, San Antonio."

2:17 AM — The owner called the homeowner back.

2:18 AM — He confirmed the job and said "I'll have a tech there in 30 minutes."

2:50 AM — The tech arrived. The capacitor had failed. He replaced it in 20 minutes.

3:15 AM — The AC was running. The house was cooling down. The homeowner was relieved.

The Full Story

This wasn't a one-off. Over a 6-month period, this San Antonio shop handled 23 after-hours emergency calls through their chatbot. Each one followed the same pattern:

1. Customer visits the site after hours (average time: 11 PM - 4 AM)

2. Chatbot engages immediately and identifies urgency

3. SMS alert goes to the owner (not the whole team)

4. Owner calls back within minutes

5. Tech dispatched within 30-60 minutes

The results:

| Metric | Value |

|---|---|

| After-hours emergencies handled | 23 in 6 months |

| Average response time (first contact) | Immediate (chat) |

| Average callback time by owner | 4 minutes |

| Jobs booked from after-hours emergencies | 18 of 23 |

| Revenue from those jobs | $7,560 |

| Average job ticket | $420 |

| Competitors who answered before the bot | 0 |

What Made the Difference

The shop was a small 3-truck operation. They didn't have a call center. They didn't have a 24/7 dispatcher. They had a chatbot that worked while they slept.

The owner: "I keep my phone on silent when I sleep. But SMS alerts from the bot come through as urgent. I wake up, read the summary, and decide if I need to go. Most of the time, I send a tech and go back to sleep."

Why Customers Choose the Bot

When surveyed, the emergency customers said:

• "It answered immediately." No ring, no voicemail, no hold.
• "It asked the right questions." The bot got their info quickly without friction.
• "The owner called back fast." Within minutes, not hours.
• "I felt like someone cared." Compared to the voicemail graveyard at the franchise shops.

The Takeaway

An AI chatbot doesn't replace a human in emergencies. It replaces the voicemail. When a customer's AC dies at 2 AM, they don't need a conversation — they need someone to show up. The bot makes sure that happens.