> ## Documentation Index
> Fetch the complete documentation index at: https://docs.apollospace.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Dealerships & automotive

> Dealerships and service shops: qualify leads, schedule test drives and service, bring back overdue-service customers, and follow up on parts — with pricing, trade-ins, and financing always staying with the person who decides.

It's 9am on a Monday. Thirty messages came in over the weekend: someone saw
the SUV listing and wants to know if "it comes in silver"; two asked for a
test drive but nobody confirmed a time; a customer is a month past due on her
30k service and no one called; and the part another customer has been waiting
on for days finally arrived — except he was never told. It isn't a lack of
traffic — it's too much traffic for one front desk.

**What leaves your plate:** the first reply to every prospect, lead
qualification, scheduling test drives and service with reminders, bringing
back whoever is due for maintenance, and the "your part is in" notice all
start happening on their own — with context, within budget, and with pricing,
trade-in, and financing always waiting on the person who decides.

## How it works, in one line

In automotive the bottleneck isn't having cars on the lot — it's **keeping up
with many prospects and service customers at once without letting any of them
go cold.** The agent doesn't replace the salesperson or the service advisor:
it makes sure every lead is answered, qualified, and followed up, and hands
the person what truly needs judgment — price, trade-in valuation, financing,
and contract.

```mermaid theme={null}
flowchart LR
    A[Trigger<br/>new lead, test drive,<br/>service due, parts] --> B[Agent pulls context<br/>from Company Brain + CRM]
    B --> C[Decides + uses tools<br/>CRM, Outbound, Routines, Docs]
    C --> D{Sensitive?<br/>price, trade-in,<br/>financing, contract}
    D -->|No| E[Replies / qualifies /<br/>schedules / reminds]
    D -->|Yes| F[Proposes + calls the decider<br/>with full context]
    E --> G[Becomes memory<br/>the next step already knows]
    F --> G
```

## Concrete flows

### 1. Qualify the vehicle lead — and capture preferences into the CRM

**Trigger:** a prospect arrives by [WhatsApp](/en/integrations/whatsapp),
email, or the vehicle listing form.

The agent replies on the spot and runs a short qualification, one question at
a time: which model they're after (new or used, trim, transmission, color),
price range, when they'd trade, whether a car is coming in on the deal,
whether financing is on the table. As answers come in, it creates the
[lead in the CRM](/en/features/crm), fills the preference fields, and logs the
conversation as activity — nothing gets lost in a WhatsApp screenshot. If the
profile matches a vehicle in stock that's already in the
[Company Brain](/en/features/brain), it notes the suggestion on the card.

The result: when the salesperson opens the [pipeline](/en/features/crm) in the
morning, every new lead arrives **qualified and sorted by stage** — not as a
queue of unread messages to triage from scratch.

<Note>
  Qualification pulls from the Brain what's worth asking for *this* kind of
  sale — a financed new car raises different questions than a cash used one.
  The script lives in the Brain, not in a fixed form.
</Note>

### 2. Schedule test drives and service — and the reminder that cuts no-shows

**Trigger:** the lead wants a test drive, or a customer wants to book a service
appointment.

The agent matches the agreed availability against the calendar
([Google Calendar via Composio](/en/integrations/composio)) and proposes
times. Once confirmed, it logs the appointment as activity in the
[CRM](/en/features/crm), creates the [task](/en/features/boards) on the
responsible salesperson's or advisor's board, and **schedules an automatic
reminder** on the customer's channel ([WhatsApp or email](/en/features/outbound))
— the day before and a few hours ahead. If you want to keep a hand on the
messaging, each send can wait for **your approval before it goes out**.

```mermaid theme={null}
flowchart TD
    A[Wants test drive or service] --> B[Agent proposes times]
    B --> C{Customer confirms?}
    C -->|Yes| D[Logs appointment + creates task<br/>+ schedules reminder]
    C -->|No reply| E[Routine re-engages the customer<br/>in a few days]
    D --> F[Reminder day-before + day-of]
    F --> G[Becomes memory]
    E --> G
```

The win isn't "a bot that books slots." It's that **the agreed test drive or
service doesn't vanish** among dozens of parallel conversations, and the right
reminder the day before knocks down the number of people who don't show — and
frees a bay or a demo car that would otherwise sit waiting.

### 3. Bring back the overdue-service customer — on its own, on the right day

**Trigger:** a [routine](/en/features/rotinas) the agent schedules for itself
sweeps the customer base for anyone with maintenance due or overdue — by
mileage, by time since the last visit, or by the manufacturer's recommended
interval.

The routine pulls from the [CRM](/en/features/crm) who's entered the service
window, cross-references the service history in the
[Company Brain](/en/features/brain), and fires — via
[Marcus](/en/agents/marcus) — a reminder by
[email or WhatsApp](/en/features/outbound): "your 30k service is coming up —
want me to hold a slot?". Anyone who replies drops straight into the scheduling
flow from flow 2; anyone who doesn't is marked for a fresh touch later and
becomes a [proactive notification](/en/features/rotinas) in the advisor's bell.

<Info>
  **The service base is a dealership's most forgotten asset.** A routine has no
  hectic week — it circles back to whoever is due for service on the right day,
  every time, and turns a stalled spreadsheet into a full shop and a loyal
  customer.
</Info>

### 4. Follow up on the ordered part — and keep the customer informed

**Trigger:** a part was ordered for a service and the customer is waiting, or
the order status changed.

The agent tracks the order and, on every meaningful change — "shipped from the
supplier," "arrived at the shop," "ready to install" — logs the update as
activity in the [CRM](/en/features/crm) and prepares a note to the customer on
their channel. When the part lands, it proposes **booking the car back in**,
already matched against the shop's calendar. Anything that's status information
flows freely; any message that touches the **service amount, the approved
estimate, or a charge** is **proposed for the advisor to approve** — it never
goes out on its own.

<Info>
  **Why this matters:** a customer left in the dark calls three times asking
  about the part and walks away annoyed even when the work was done well.
  Keeping them informed at every step, with nobody having to remember, is half
  of after-sales satisfaction.
</Info>

## What stays human

<Warning>
  The agent **proposes**, never decides alone, whenever the subject is:

  * **Vehicle price and negotiation** — discount, terms, any value
    conversation with the customer.
  * **Trade-in valuation** — what the used car coming into the deal is worth.
  * **Financing** — an approved simulation, credit terms, the installment, any
    decision involving credit.
  * **Service estimate and charges** — labor and parts amounts, estimate
    approval, refund, any adjustment to amounts.
  * **Contract and signature** — a formal proposal, purchase agreement, a work
    order that creates an obligation.
  * **Heavy external, legal, or regulatory communication**, governance
    decisions, and destructive actions.

  In those cases the agent prepares everything — the draft, the customer's
  context, the negotiation or service history — and **waits for the decider to
  approve.** Every action lands in the auditable trail, with author and
  justification.
</Warning>

## Where to start

<Steps>
  <Step title="Upload stock and service history to the Brain">
    Ingest your vehicle sheets, the service-interval table per model, and your
    customers' service history into the [Company Brain](/en/features/brain).
    That's where the agent pulls its first response, its qualification, and the
    service window from — without it, there's nothing to consult.
  </Step>

  <Step title="Wire a channel and let the agent take first line">
    Connect [WhatsApp](/en/integrations/whatsapp) and let the agent qualify and
    log every lead in the [CRM](/en/features/crm). Start with human approval
    before each external message and loosen it as it earns trust.
  </Step>

  <Step title="Schedule the overdue-service routine">
    Create a [routine](/en/features/rotinas) that sweeps the service base and
    proposes bringing back whoever's due for maintenance. One agent, one
    cadence, one channel — and autonomy grows like a ratchet, not a leap.
  </Step>
</Steps>

## Next steps

<CardGroup cols={2}>
  <Card title="Commercial & sales" icon="handshake" href="/en/use-cases/commercial">
    The full prospecting, qualification, and follow-up pattern — of which
    selling vehicles is a high-volume case.
  </Card>

  <Card title="People & after-sales" icon="users" href="/en/use-cases/people">
    The relationship and retention side — the customer who comes back for
    service and refers the dealership.
  </Card>

  <Card title="Native CRM" icon="address-book" href="/en/features/crm">
    Where the lead, the preferences, and the service history stay alive — no
    human memory required.
  </Card>

  <Card title="Routines" icon="rotate" href="/en/features/rotinas">
    The routines that bring back the overdue service and follow up on the part
    — on the right day, every time.
  </Card>
</CardGroup>