Skip to main content
    Case Study

    Mint & Scotties Rentals - One AI Voice Agent Across Three Branches

    Company

    Mint & Scotties Rentals

    Industry

    Car rental

    Branches

    Auckland, Christchurch and Queenstown

    What We Built

    • • An AI voice agent, Ela, on each branch's own local number
    • • One shared script with per-branch airport, hours and after-hours details
    • • Call summaries routed to each branch's own inbox
    • • A filter that only emails the calls a person needs to act on
    • • A secure call review dashboard across all three branches

    Client Overview

    Mint & Scotties Rentals hires cars from branches in Auckland, Christchurch and Queenstown. A lot of the people who ring a rental company are not at a desk: they are standing at an airport waiting for a shuttle, or pulled over on the side of the road with a car that has stopped. The phone has to work for them first.

    The Challenge

    Ela started on the Auckland line in mid-2026. Two problems showed up quickly. The first was scale: Christchurch and Queenstown needed the same help, but each branch has its own airport, its own hours and its own after-hours arrangements. The second was the inbox. At first every answered call emailed the team, and the plain information calls (opening hours, fuel policy, what documents to bring) were burying the ones that needed someone to act.

    What the team needed:

    • Airport pickups that never get missed: a customer at a terminal is waiting on a shuttle someone has to send
    • Breakdowns and accidents handled properly: clear roadside instructions on the call, and the team told every time
    • One approach, three branches: the same standard of answer everywhere, with each branch's own local details
    • An inbox worth reading: the calls that need a person, without the ones that do not

    The Solution

    Automate AI built Ela as one agent with one script, then set her up for each branch from a small block of branch settings. Each branch has its own local number and its own Ela, and all three report to one backend that works out which branch a call came in on.

    1. One script, three branches

    Everything Ela learned on the Auckland line carries across: how she handles caller ID, the order she takes an airport pickup in, and when to wrap a call up into a callback. Only the branch identity, the airport directions, the office hours and the after-hours arrangements change from branch to branch. Adding a branch means adding its settings, not writing a new agent.

    2. Airport pickups asked the right way

    Auckland has two terminals, so Ela asks whether the caller is at International or Domestic. Christchurch and Queenstown each have a single pickup point, so she does not ask. When a caller gives a flight number she records it, because that is what the team needs to send the shuttle.

    3. Only the calls that need a person

    After each call, the call is analysed and sorted. A question Ela answered completely, with nothing left for anyone to do, is not emailed. Accidents, theft, breakdowns, complaints, refunds and after-hours collections are always emailed, whatever else happened on the call, and so is any airport call with a flight number or an existing booking. When the analysis is unsure, the email goes. A spare email costs nothing; a missed shuttle leaves a customer at the terminal.

    4. Emails the team can act on from the subject line

    Each branch's summaries go to that branch's inbox, with the branch at the front of the subject. The subject itself says what the call was about and what to do, for example "Airport pickup, send shuttle" or "Booking change request, email customer". If a caller can only be reached by email or WhatsApp, the summary says so instead of telling the team to ring.

    5. A call review dashboard

    The team signs in with a one-time code sent to an approved email address and can see the last 30 days of Ela's calls across all three branches, with the transcript, the analysis and the recording for each one.

    Day to day

    What happens on a call

    A caller rings their local branch number and Ela answers as that branch. She answers from the branch's own information, takes the details for anything she cannot settle, and tells the caller when the office will be back to them.

    The script has been tuned against real calls. Callers are often somewhere noisy, so background speech is filtered out. Ela does not ask for a number she already has from caller ID, she checks a number a caller reads out before relying on it, and she does not end a call on her own unanswered question.

    An airport pickup call

    Example flow

    1. Caller rings the branch

      Ela answers as that branch and asks which terminal where it matters

    2. Call is analysed

      Sorted as a live airport pickup, flight number captured

    3. Branch inbox gets the email

      Subject leads with the branch and the action to take

    Never filtered out

    What Changed

    Three branches now share one agent that answers in the same consistent way, and each branch's inbox carries the calls it needs to act on rather than every call that was made.

    • Every branch has its own line - callers ring a local number and reach an agent that knows that branch
    • The inbox is for action - plain information calls stay out, anything that needs a person always lands
    • Airport and roadside calls come first - they can never be filtered out
    • Adding a branch is a settings change - the script and the backend are already shared

    Key Takeaways

    Mint & Scotties is a useful example for any business with more than one location on the phone:

    • Build one agent, configure it per site: improvements made on one branch reach every branch
    • Answering the call is half the job: deciding who needs to hear about it matters just as much
    • When in doubt, send: a filter should only ever hold back a call it is sure about
    • Write the subject line for the person reading it: what happened and what to do, before anyone opens the email

    Running more than one location?

    Let's map what an AI voice agent would answer at each of your sites, and which calls it should hand back to your team.

    Explore AI voice agents

    Free 30-min call · We'll show you 2–3 things to automate first · No obligation

    More Success Stories

    See how we've helped other New Zealand businesses transform with AI automation

    Retro Bowl

    Hospitality

    An AI voice agent answering the phone in a busy eight-lane bowling alley, plus front-counter texting.

    Read case study

    Packaging Products

    E-commerce

    +20% ROAS in month one; social restarted after a 12-month gap.

    Read case study

    Packaging Products Sales Reporting

    E-commerce

    Custom dashboard reveals hidden cash sales patterns and high-value clients.

    Read case study