Revolution AgencyRevolution Agency

Delivered by our team through its European experience (Evolution Agency, France).

Case study: SaaS for dealerships, Canada

A dealership agent that qualifies every visitor, built to be replicated.

For Autoroot, a Quebec startup selling a turnkey SaaS to car dealerships, our team built an agent that presents new and used vehicles, qualifies the visitor and passes the lead to the CRM. It was designed to be replicated from one dealership to the next.

10+

criteria to search the vehicle inventory (mileage, engine, price, type...)

2

departments kept apart: sales and service

30 words

at most per answer, unless the visitor asks for more

Methodology: the results on this page are qualitative. They come from operating the tools and from the testing done before each launch. We publish client-specific percentages only once they are measured and approved.

Context

Autoroot is a Quebec startup building a turnkey SaaS for car dealerships. On a dealership website, visitors look for a vehicle, want a test drive, ask about financing or a trade-in, or book a service appointment.

Autoroot needed a replicable agent that could present new and used vehicles and generate qualified appointments with precise tracking. The starting point:

  • A vehicle inventory described by more than ten criteria (mileage, engine, price, type...)
  • Two departments not to be mixed up: sales (purchase, test drive, financing, trade-in) and service (maintenance, tires, bodywork, parts)
  • A product sold to several dealerships: each new installation had to be fast

Challenge

A dealership agent has one job that matters: turn a visit into a contact a salesperson can call back. Without inventing prices, and without pushing the visitor.

1

Get the contact without pushing

Name and phone number come first, but visitors arrive with a question. The agent answers it, asks again later, and never requests contact details more than three times in a row.

2

New or used, sales or service

Each request follows a different logic: a carousel of vehicles for used cars, information for new ones, a booking link for the service department. The agent first has to know which one it is.

3

Replicate without rebuilding

Autoroot sells the same product to several dealerships. Each new dealership had to be equipped quickly, not run as a new project.

Solution: a dealership agent connected to the inventory and the CRM

An agent built by our team for a first dealership, then replicated. It answers from the dealership's own data, and offers a callback from a specialist as soon as it does not know.

  1. 1

    An inventory searchable by criteria

    The vehicle databases were vectorized with more than ten criteria (mileage, engine, price, type...) and sync automatically from Google Sheets. For used vehicles, the agent searches the inventory before showing a carousel, then offers to refine by budget, mileage, type, fuel or color. When nothing matches, it suggests close alternatives.

  2. 2

    One source per type of question

    Beyond used vehicles, the agent queries separate sources for new vehicles, monthly promotions, financing and trade-ins. On new vehicles it gives no exact price, rebate or government incentive amount: it offers to have a specialist explain the current offers.

  3. 3

    The lead goes to the CRM right away

    The agent asks for name and phone number in the same sentence. A first name is enough, and it asks for no confirmation. As soon as it has both, or an email address, the lead is sent to the CRM through the API, without waiting for the conversation to end.

  4. 4

    Appointments that respect opening hours

    Before confirming an appointment, the agent checks the opening hours and turns down slots in the last thirty minutes of the day. Service requests (maintenance, tires, bodywork, parts) go to the service department's online booking, sales requests to a salesperson.

  5. 5

    Short answers, Quebec vocabulary

    The agent addresses visitors formally, keeps answers short (thirty words at most unless the visitor asks for more), never asks two questions in one message and uses Quebec terms: VUS, not SUV. When it does not know, it apologizes and invents nothing.

  6. 6

    Built to be replicated

    The architecture was prepared for fast duplication per dealership: the inventory, promotions and opening hours change, the agent's logic stays the same.

Results

No percentages on this page: we only publish a client rate once it is measured and approved. The results below describe how the delivered agent behaves.

Service requests go to the online booking, sales requests to a salesperson

No exact price, rebate or incentive amount given on new vehicles: a specialist takes over

No appointment confirmed outside opening hours or in the last thirty minutes of the day

Leads sent to the CRM as soon as name and phone number are known, without waiting for the end of the conversation

The honest framing: the agent does not sell cars. It answers, qualifies, and hands the salesperson a contact that already carries the project, budget and timeline when they are known.

Stack

Chat agent on the dealership websiteVehicle database vectorized on 10+ criteriaAutomatic sync from Google SheetsDedicated sources: new vehicles, promotions, financing, trade-inUsed-vehicle carouselLeads sent to the CRM by API

Your website shows the inventory. Your agent should bring the visitor to the dealership.

Evaluate your project free with our interactive tools, or book a call: we look at your real requests and tell you what an agent can handle, and what it should leave to your salespeople.

Book a call
NovaVoice guide
Ready