2 min read

How Our AI Receptionist Handles 200 Daily Inquiries Without a Human at the Desk

How Our AI Receptionist Handles 200 Daily Inquiries Without a Human at the Desk

Hook

It was 9 AM on a Monday. My phone showed 14 missed calls and 32 unread Telegram messages from a single client's property listing page. The real estate office had no receptionist that day.

Background

I run a small IT consultancy that also manages the tech side for a few local businesses. One of them is a busy real estate agency with a constant flow of inquiries — people asking about open houses, square footage, parking, pets allowed, and schedule a viewing.

Their admin staff was drowning. They tried email auto-responders, but most people used Telegram or phone. They needed someone — or something — to stand at the front desk 24/7.

Problem

The pain was simple: every unanswered inquiry meant a missed lead. Staff had to answer the same questions dozens of times a day. Calls went to voicemail. Telegram messages piled up. By the time someone replied, the prospect had already contacted three other agencies.

Why it matters

Lost leads translate directly to lost commissions. But the cost is deeper: clients remember the slow reply. They tell friends. The agency's reputation suffers. And the staff burns out from repetitive Q&A, leaving less energy for actual sales.

Doing nothing was not an option.

Solution

I built an AI receptionist using AgentForge. The core is a Hermes agent that lives inside a Docker container, deployed with Compose on a small VPS. I configured the agent's behavior through the web UI — defined conversation flows, FAQ answers, and routing rules.

The agent connects to the agency's Telegram group. When a new message arrives, it handles it automatically:

  • 🤖 Answers common questions like "Is the unit still available?" or "What's the pet policy?" from a knowledge base
  • 📞 Routes complex requests (e.g., "I want to negotiate the price") to a specific staff member via a private chat alert
  • 🔄 Keeps the conversation context — remembers what was discussed earlier in the thread

If the agent cannot answer, it hands off gracefully: "Let me connect you with Elena, she'll be right with you."

Result

After two weeks, the agent was handling about 200 inquiries per day. Most were resolved without human involvement. Staff only saw messages that required a personal touch — maybe 20 a day.

The agent doesn't replace the human, it buys them time to focus on what matters.

Response time dropped from hours to under 30 seconds. The agency started capturing leads that previously slipped through. The admin used their saved time to follow up on high-intent buyers instead of repeating square footage numbers.

The entire setup took me two hours: one to deploy the stack via Docker Compose, one to tune the prompts in the web UI. The hardware is a $10/month VPS. The agent runs unattended.

AgentForge

AI agents for real work. If you want to see how this works in practice, visit our product site and blog.

Share
Serhii

Serhii