A Developer Built an AI Receptionist for Her Brother's Mechanic Shop to Recover 100+ Missed Calls a Week
A software engineer built a RAG-powered voice receptionist for her brother's luxury mechanic shop after realizing hundreds of missed calls a week meant hundreds of lost jobs worth up to $2,000 each
Continue exploring this workflow
The Strategy
Kedasha Kerr, known online as That LadyDev, noticed her brother's luxury mechanic shop was losing serious money to a problem that had nothing to do with the quality of the mechanical work: the phone simply rang too much for anyone to answer consistently. With jobs worth anywhere from $450 to $2,000 each, every unanswered call was a meaningful and quantifiable loss, not a minor inconvenience. Her response, a voice receptionist she named Axle, is built with a level of engineering rigor that goes beyond most of the productized voice-agent offers in this directory. Rather than a shallow FAQ script, Axle runs on a retrieval-augmented generation pipeline with more than 21 vectorized documents covering the shop's services, pricing, and policies, stored in MongoDB Atlas's vector search and embedded using Voyage AI. This lets the agent answer pricing questions accurately based on real shop data instead of guessing or hallucinating numbers. Claude Sonnet 4.6 generates the actual conversation responses, while Vapi handles the voice layer itself, using Deepgram for speech-to-text and ElevenLabs for text-to-speech, with a FastAPI backend tying the whole system together. This is a from-scratch, code-first build rather than a no-code automation stack, reflecting Kerr's background as a developer rather than an automation agency operator. As the first post in a planned series, the piece is explicit that these are projected numbers rather than confirmed results: the goal is recovering over 100 missed calls a week, based on the shop's actual call volume, not a number already measured after deployment.
How It Works
Quantify the actual cost of the problem first: estimate missed call volume and the average value of a job to justify building a dedicated system.
Gather and vectorize the shop's real operational knowledge, in this case more than 21 documents covering services, pricing, and policies.
Store the vectorized knowledge in MongoDB Atlas using its vector search capability, with Voyage AI generating the embeddings.
Build the conversational logic using Claude Sonnet 4.6, grounded in retrieval from the vectorized knowledge base rather than relying on the model's general knowledge alone.
Handle the voice interface with Vapi, using Deepgram for speech-to-text transcription and ElevenLabs for natural-sounding text-to-speech output.
Tie the pipeline together with a FastAPI backend that routes calls, retrieves relevant context, and manages the conversation flow.
Test the system against realistic caller scenarios, including pricing questions that require accurate retrieval from the knowledge base rather than approximate answers.
Deploy as a first phase, with plans to track actual missed-call recovery over time as later posts in the series.
Results
The post projects the system recovering more than 100 missed calls a week for the shop, based on documented call volume and average job values between $450 and $2,000. These are explicitly framed by the author as projections at the time of writing, part one of a series, rather than confirmed post-deployment results, and have not been independently verified.
Our Take
This is one of the more technically credible builds in this directory, and it shows: using retrieval-augmented generation with real vectorized pricing and policy data is a meaningfully more robust approach than the scripted FAQ flows common in most productized voice-agent offers, particularly for a business like auto repair where inaccurate pricing quotes can directly damage trust. The honesty about these being projected rather than measured numbers is also refreshing in a space full of confident, unverifiable claims. The tradeoff is complexity: this is a code-first, multi-service architecture (vector database, dedicated embeddings, custom backend) that requires real engineering skill to build and maintain, not something assembled from no-code building blocks in an afternoon. Best suited as a reference architecture for developers building voice agents that need to answer detailed, accurate pricing questions rather than generic scheduling.
Frequently Asked Questions
The practical questions a builder or operator is likely to ask before trying a strategy like this.
What does this auto repair customer service AI agent actually do?
This auto repair customer service AI agent is a real workflow where the agent takes on an operational job, not just a brainstorming task. A Developer Built an AI Receptionist for Her Brother's Mechanic Shop to Recover 100+ Missed Calls a Week shows what that looks like in practice. A software engineer built a RAG-powered voice receptionist for her brother's luxury mechanic shop after realizing hundreds of missed calls a week meant hundreds of lost jobs worth up to $2,000 each The practical value comes from the agent handling repeatable business work with enough autonomy that a human only steps in after context has already been gathered.
Who should use a auto repair customer service AI agent like this?
This example is most relevant for auto repair operators. It is especially relevant for businesses where speed to lead, after-hours coverage, or consistent intake quality directly affects revenue. The category here is Customer Service, which means the best fit is a team looking to turn a manual bottleneck into a repeatable system with a auto repair customer service AI agent.
Which tools are used in this auto repair customer service AI agent setup?
The source names Vapi, Claude, MongoDB, ElevenLabs, Deepgram. That matters because one of the strongest signals in this directory is whether the operator shared the actual stack. Named tools make a auto repair customer service AI agent strategy far more useful than vague claims about “an AI system” doing the work.
How hard is it to implement a auto repair customer service AI agent like this?
Advanced difficulty is the current read. The listing suggests a launch window of weeks. Startup cost is listed as $50-200/mo. We were able to extract 8 concrete workflow steps from the source. We would treat a auto repair customer service AI agent like this as a workflow that needs real business context, testing, and exception handling rather than something you should copy blindly from one prompt.
What results can a auto repair customer service AI agent produce?
The post projects the system recovering more than 100 missed calls a week for the shop, based on documented call volume and average job values between $450 and $2,000. These are explicitly framed by the author as projections at the time of writing, part one of a series, rather than confirmed post-deployment results, and have not been independently verified.
How credible is this auto repair customer service AI agent case study?
Right now the evidence comes from an article from itsthatlady.dev. That is enough for us to study and curate the workflow, but not enough on its own to treat this auto repair customer service AI agent like an audited case study. We look for named tools, concrete results, and enough workflow detail to understand what was actually deployed, then we add our own editorial judgment on top.
Related Strategies
More AI agent strategies you might find useful
Building a $3,500 AI Customer Service Chatbot Live From Scratch
A full live build of a customer service chatbot with knowledge base retrieval, l…
A $10K Healthcare Voice Agent That Handles Patient Intake and Appointment Booking
A voice AI agent for healthcare clinics that qualifies patients, verifies insura…
A $40K MRR AI Customer Support Tool Built by Two Founders After a Failed VC Startup
Two founders pivoted from a failed VC backed startup to bootstrap an AI customer…
A Voice AI Receptionist for Veterinary Clinics That Books Appointments Automatically
A voice AI agent handles incoming calls at a vet clinic, checks calendar availab…
A Plumbing AI Receptionist That Books Emergency Calls and Logs Everything to a CRM
An AI voice receptionist for a plumbing company that books emergency service cal…
AI Voice Receptionist With Calendar Booking Using Vapi and n8n
An AI voice agent that answers pest emergency calls 24/7, qualifies the infestat…
Want more strategies like this?
Get weekly AI agent case studies in your inbox.