AI Receptionist (Universal)

Local businesses that currently rely on a human receptionist, an answering service, or worst of all, letting calls roll to voicemail. Ideal for: owner-operator service businesses (the owner runs the phone and they can't anymore), small medical practices, dental offices, salons with stylists who can't answer during services, restaurants drowning in reservation calls, law firms losing leads to voicemail, real estate agents who can't pick up while showing. Any business where the owner has said "I need someone to answer the phone but I can't justify a full-time hire."

An AI voice agent (built on ElevenLabs Conversational AI) answers the business phone 24/7 in a natural human-sounding voice with sub-1.5-second response time. The agent is configured with the business's specific hours, services, pricing, FAQs, and booking rules. It can: answer hours/location/pricing questions, qualify leads, book appointments into a calendar, look up existing customers, take messages, and transfer urgent calls to the owner.

Every call is transcribed and summarized in a Base44 admin dashboard. The business owner only deals with the 10-20% of calls the AI escalates. This tool eliminates the receptionist line item from the P&L while also covering nights and weekends that no human ever would.

This is the most complex tool in the library. There are two distinct platforms that need to be configured (Base44 + ElevenLabs), and they need to talk to each other via webhooks and tool endpoints. Most non-technical students will need 2-4 hours total setup time, including the testing iteration.

Expect Base44 to build Part A in 10-15 minutes. Then the operator spends another 60-90 minutes configuring ElevenLabs in their dashboard (it's mostly form fields and copy-paste, but it's a separate platform).

Likely follow-up prompts you'll need to send to Base44:

If the tool endpoints aren't returning fast enough (ElevenLabs has timeouts): "The tool endpoints are taking more than 2 seconds to respond, which causes ElevenLabs conversation flow to break. Optimize each function: lookupCustomer should hit a single indexed query on phone_number, checkAvailability should pre-cache available slots, createBooking should respond immediately and fire the SMS confirmation asynchronously."

If the HMAC verification is blocking legitimate ElevenLabs calls: "The HMAC signature verification is blocking ElevenLabs tool calls. Check that we're using the correct shared secret from ElevenLabs dashboard, and verifying against the right header (ElevenLabs uses 'X-ElevenLabs-Signature'). If we're not sure of the exact header name, fall back to disabling verification temporarily for testing, then re-enable once it's working."

If the call summary isn't getting logged at end of calls: "The logCallSummary function isn't being called at the end of conversations. In the ElevenLabs agent system prompt, we need to explicitly instruct the agent to call this tool right before ending every call. Update the system prompt to include this instruction at the top, not just at the bottom."

If the dashboard isn't showing transcripts properly: "The transcript field is showing as plain text but should render as a chat-bubble conversation between the AI agent and the caller. Parse the transcript JSON from ElevenLabs (each turn has role: 'assistant' or 'user' and content) and render alternating chat bubbles."

The biggest gotcha: ElevenLabs configuration is a separate platform learning curve. The student will get stuck on ElevenLabs setup, not Base44 setup. Make sure the Phase 1 playbook walks through every ElevenLabs screen specifically.

Pricing reality check: ElevenLabs Conversational AI is around $0.08-0.15 per minute of conversation depending on the plan. A typical service business gets 50-200 calls/month averaging 3-5 minutes. That's roughly $25-150/month in ElevenLabs costs.

Plus Twilio (~$30/month). Build this into the retainer pricing.