DMS Integration Checklist Before Deploying a Dealership AI Voice Agent
A checklist for fixed ops and IT stakeholders evaluating an AI voice agent: DMS ownership, data access, escalation rules and pilot rollout.
Most vendor pitches for an AI voice agent focus on the phone call itself: how natural it sounds, how quickly it books an appointment, how it handles an angry customer. Those things matter, but they're not where deployments actually go wrong. The failure point is almost always upstream of the call, in the handful of decisions about DMS access, data ownership, and escalation that nobody wrote down before go-live.
If you're the person who has to answer for this system once it's live, whether that's a fixed ops director signing off on the rollout or an IT manager fielding the DMS vendor's questions, it helps to have a checklist you can work through before any contract gets signed. This is that checklist. It's not a sales pitch for a particular architecture. It's the set of questions worth asking, regardless of who you end up choosing.
Start with what DMS you're actually running, and who owns it
This sounds obvious, but it's worth stating plainly because the answer is rarely as simple as 'we run Revolution' or 'we're on Tune.' Many dealership groups run different platforms at different stores, sometimes because of an acquisition, sometimes because a store negotiated its own contract years ago and nobody's revisited it. Before you talk to a voice agent vendor, get a straight answer on which DMS or scheduling system each location actually uses today, not what the group standardized on in theory.
Just as important: who has the authority to approve a new system touching that DMS. In a lot of stores, that's not the GM. It might be a regional IT contact, a third-party DMS administrator, or a vendor relationship the dealership doesn't manage directly. Figure out who that person is early, because they will need to be in the room, or at least on the email thread, before any access gets granted. Trying to route around them later costs more time than it saves.
Get specific about what data actually needs to be accessible
It's tempting to think of DMS access as one big switch: on or off. In practice, a voice agent only needs a narrow slice of what the DMS holds, and it's worth naming that slice explicitly rather than letting a vendor assume broad access by default. At minimum, ask what the agent needs to see to do its job: open appointment slots and scheduling rules, the list of service types the shop actually offers (and which ones need special handling, like certain repairs that require more lead time or a specific advisor), and enough customer and vehicle lookup to recognize a returning caller without starting every conversation from zero.
Push back on anything broader than that. A voice agent doesn't need visibility into repair order history it isn't using, technician pay data, or parts inventory, and asking a vendor to justify each piece of access they're requesting is a reasonable thing to do, not a sign of distrust. The tighter the scope, the easier this is to explain to your DMS vendor, and the less there is to worry about if something ever goes wrong.
Decide how escalation should actually work, before the first call comes in
Every AI voice agent will eventually run into a call it shouldn't handle on its own: a pricing dispute, a customer who's upset about a prior visit, a repair question that needs a technician's judgment, or simply someone who wants to talk to a person. What happens next is a decision your dealership needs to make deliberately, not something to leave to a vendor's default settings.
Work through the specifics before launch. Which calls should transfer live to a person, and who is that person during business hours versus after hours? What information does that person need handed to them so the customer doesn't have to repeat themselves? Is there a fallback if nobody picks up the transfer, and does the customer get a callback commitment or a voicemail path? None of this needs to be complicated, but it does need to be written down and agreed to by whoever manages your service drive and BDC, so the escalation path matches how your team actually operates rather than a generic default.
Ask hard questions about call recordings and data handling
Every call the agent handles will generate a recording and a transcript, and that data has to live somewhere, be retained for some period, and eventually be deleted or archived under some policy. Before you deploy, get clear answers on where that data is stored, who inside the dealership can access it, how long it's retained, and what happens to it if you ever switch vendors or shut the system down.
This isn't just a compliance checkbox. Call recordings involving customer names, phone numbers, and vehicle information are sensitive by nature, and your dealership is still the one accountable for how that data is handled, even if a vendor is processing it on your behalf. Ask whether recordings are used to improve the underlying models, whether that's opt-out or opt-in, and whether you can pull an export of your own data on request. If a vendor can't answer these questions clearly, that's worth treating as a signal on its own.
Pilot on a subset of lines before you cut over everything
Whatever the sales conversation implies about how ready this is for full deployment, don't cut every phone line over on day one. Start with a defined subset: maybe just after-hours calls, or a single line dedicated to a specific campaign, or one location out of a multi-store group. That gives you a controlled environment to check that appointments booked by the agent actually match what your service drive can deliver, that escalations are landing with the right person, and that the transcripts read the way you'd expect a conversation with your dealership to read.
Set a real timeline for the pilot, not an open-ended one, and decide up front what you're checking for: booking accuracy against actual DMS availability, how often calls get escalated and whether those escalations were the right call, and whether customers are getting through the conversation without getting stuck. Only widen the rollout once those numbers look right on the subset you can watch closely. Cutting over all at once because a demo went well is how a small configuration issue turns into a dealership-wide problem before anyone notices.
None of this is complicated, but it's the kind of groundwork that's easy to skip when a demo looks good and everyone's eager to move fast. Get clear on your DMS setup and who owns it, scope the data access tightly, write down your escalation rules before the first call comes in, ask direct questions about how recordings and customer data are handled, and pilot on a subset of lines before you commit to a full rollout. If you're evaluating this for your dealership or group and want to walk through what this looks like against your specific DMS setup, it's worth a short conversation with the Zenith CC team before you sign anything.
To discuss how Zenith CC could fit your dealership, book a conversation.