labs@atheron:~/industries/automotive/ai$ train agent --grounded

AI for dealerships

Service advisor agents for calls, texts and chat.

Service departments lose customers to unanswered phones. An agent can take bookings, give repair status and check recalls at any hour, and pass the rest to an advisor. A frontier model through its API is the usual start, because conversation quality matters most and the data is routine.

Start a project See the example projects

// what matters here

What is different in this industry.

Bookings that fit the shop

The agent books into your scheduler with the time each job really takes, so the shop is not overbooked by a polite machine.

Status from the DMS, not a guess

Repair order status comes from the dealer management system, so the agent says what the advisor would say.

Recalls checked by VIN

Open recalls are checked against manufacturer data when the customer gives a VIN, and the booking includes them.

Advisors keep the hard calls

Complaints, warranty disputes and anything the agent is unsure of go to an advisor with the conversation summarised.

// example projects

Priced examples for this service.

Each one is hypothetical, labelled as such, and priced live from our rate card. Open any of them in the estimator and make it yours.

Example projectWeb appAIFrontier model

Advisor agent for calls and chat

Hypothetical. Not a client, not a result.

// The problem

A busy dealership misses calls about bookings, repair status and recalls, and every missed call is a customer who may book somewhere else.

// What we would build

An agent that answers chat, texts and, through a voice service, phone calls: it books appointments in the scheduler, gives repair order status from the DMS and checks open recalls by VIN, handing anything else to an advisor. It runs on a frontier model through its API, which handles conversation well and starts fastest; only what each answer needs is sent.

// What is in it

Bookings made in your service scheduler, with the right time for the job
Repair order status read from the DMS
Open recall checks by VIN against manufacturer data
Hand-off to an advisor with the conversation summarised
Conversations in English and French
Evaluation set of real conversations, run before every change

// Stack

Frontier model API
Twilio voice gateway
Python
Next.js
DMS and scheduler APIs
DigitalOcean

// estimate

Build: ≈ US$102,000 to US$156,000, delivered within 21 weeksCAD 144,900 to 221,600
Hosting: ≈ US$558 a monthCAD 795 a month
Support: ≈ US$2,500 a monthCAD 3,565 a month

Prices in your currency are estimates from today's Bank of Canada rate. All invoicing is in CAD or USD.

AI route: Frontier model
Model running cost: ≈ US$383 a monthCAD 545 a month

Timeline by milestone

Discovery: 2.9 to 3.4 weeks
Specification and evaluation plan: 0.4 weeks
Design approved: 2.4 to 3.4 weeks
Core features: 1.4 to 1.9 weeks
Full build: 0.9 to 1.4 weeks
Working pilot: 3.9 to 6.4 weeks
Testing and fixes: 0.9 to 1.9 weeks
Launch: 0.9 to 1.4 weeks
Production: 0.9 to 1.4 weeks

Outside our hands, and added to the calendar

API access approved by your DMS vendor: 1 to 4 weeks

How it is paid

Deposit 20%: CAD 28,980 to 44,320
Discovery 2%: CAD 2,898 to 4,432
Specification and evaluation plan 10.2%: CAD 14,779.80 to 22,603.20
Design approved 2%: CAD 2,898 to 4,432
Core features 6.2%: CAD 8,983.80 to 13,739.20
Full build 4.1%: CAD 5,940.90 to 9,085.60
Working pilot 30.7%: CAD 44,484.30 to 68,031.20
Testing and fixes 2%: CAD 2,898 to 4,432
Launch 2%: CAD 2,898 to 4,432
Production 10.8%: CAD 15,649.20 to 23,932.80
Holdback, 30 days after launch (10%): CAD 14,490 to 22,160

Start from this example

// questions

Questions we get about this.

Can the agent answer the phone too?

Yes, through a voice service that turns speech into text and back. It answers in English and French and transfers to a person whenever asked.

Which customer data does the model provider see?

Only what each answer needs, such as the booking details or a repair order status. If that is too much, an open-source model in your own cloud is priced as an option.

Will it quote repair prices to customers?

Only prices you publish, such as fixed-price maintenance menus. Anything else goes to an advisor.

Back to the industry: Automotive and mobility About the service: AI agents and models

// next

Not quite your project?

Tell us what you have in mind. We will come back to you with a range and the questions that would narrow it. Or book a call and talk it through.

Start a project Book a call