AI BDC Implementation Guide for Dealerships

The order of operations that works, written for a store that has to keep selling cars while the change happens.

Most failed BDC projects fail at this step, months before any software is involved. A store buys a phone system and a set of automations, then discovers that nobody agreed on what the department was supposed to produce. Is the BDC setting appointments and handing off? Is it working leads all the way to a sold unit? Is it covering service, sales, or both? Those are three different departments with three different comp plans, and automation built for one of them actively gets in the way of the others.

Write the answer down in one sentence and get the general manager to sign off on it. Then write the second sentence: what happens to a lead the moment it arrives, by name, by minute. If you cannot describe the first ten minutes of a lead's life at your store without saying "it depends," you have a process problem, and layering automation on top of a process problem produces faster chaos rather than more appointments.

The honest test is whether a new hire could read your one page and work a lead correctly on day two. Until that is true, hold off on the technology purchase. The good news is that this exercise costs nothing and usually surfaces two or three obvious leaks on its own.

Every store believes it responds quickly. Very few actually measure it, and the gap between belief and reality is where most of the money sits. Before you change any tooling, spend two weeks capturing one number per lead: minutes from arrival to a genuine human attempt, not an auto responder firing.

Do it manually if you have to. Pull twenty leads a day at random, look at the timestamps, and write the number on a whiteboard where the sales floor can see it. You want the median and the worst case, because the median flatters you and the worst case is what a shopper who came in at 7:40pm on a Thursday actually experienced.

This baseline matters for two reasons. It tells you whether your problem is speed, depth of follow up, or lead volume, and those need different fixes. And it gives you the before number that will justify the project to whoever signs the invoice. A department that cannot show its starting point cannot prove improvement later, which is how good BDC programs get cut in a bad month. Our page on what to expect from car sales leads covers why the depth number is usually the more damaging of the two.

The phone is still where the highest intent contacts land, and it is the part of the BDC that stores most often leave to chance. A shopper who calls about a vehicle and reaches a full voicemail box is not a lead you failed to convert. It is a lead you never received.

AutoMail, the voice side of the platform, gives you an IVR with call routing, number management, call forwarding and full call history. Build the tree deliberately and keep it shallow. Two levels maximum. Sales, service, parts, and a person. Route sales calls to the BDC first, with a defined rollover to the floor and a defined rollover after that, so no path in your tree ever ends in silence.

Then test it yourself from a cell phone, at 8am, at noon, and at 6:45pm on a Saturday. Call your own store from the number a customer would actually dial. Most managers who do this exercise honestly find at least one dead end in the first week. Fixing dead ends is the cheapest lead generation available to you, because those callers were already trying to give you their business.

Now automation earns its keep. The work that should be automated is the work that is mechanical, repetitive and easy to skip on a busy day: the scheduled touch, the reminder, the record keeping, the routing decision.

Inside LeadLocate that means follow up processes, automations and automation lists, task automation, drip campaigns with a drip editor, reminders and appointments. A lead arrives, an automation assigns it by rule, a process starts the cadence, and the tasks show up on the right person's list at the right hour whether or not anyone remembered. Voicemail drop lets an agent leave a prepared message and move to the next call instead of listening to the greeting forty times a day. RCS messaging with SMS fallback means richer messages reach devices that support them and plain text reaches everything else.

What should stay human is the conversation itself. A machine writing a first draft is a productivity tool. A machine pretending to be your salesperson while a customer believes they are talking to a person is a trust problem, and it will cost you more than it saves. The human BDC versus AI BDC comparison works through where each is genuinely better.

The single most common implementation mistake is switching on automation with default templates in it. Vendor defaults are written to be inoffensive to every store in the country, which means they sound like nothing and convert like nothing.

Block two hours with your best closer and write the library yourself. You need a first response for a lead with a specific vehicle attached and a different one for an open shopper who described what they want in general terms, because leads arrive both ways. You need a no answer voicemail script, three text messages spaced across the first week, a two week reactivation, and an appointment confirmation. That is nine pieces of copy and it will carry eighty percent of your volume.

Keep them short, name the store and the person, and never send anything you would be embarrassed to have read aloud in a small claims hearing. Then get consent handling right from the start, including honoring opt outs everywhere at once. Our texting compliance page covers the rules that apply, and they are not optional.

Call recording has existed in dealership software for twenty years and almost nobody listens to the recordings, because listening is expensive. A manager has eleven minutes of audio and four minutes of attention. So the recordings pile up and the coaching never happens.

Transcription changes the economics. When a call becomes text, a manager can read what happened in thirty seconds, search across calls for the word "payment" or the phrase "call me back Monday," and pull three real examples for a Tuesday morning meeting instead of one remembered anecdote. That is the difference between a department that improves and one that repeats itself.

Build the habit into the implementation rather than adding it later. Pick a standing thirty minute slot each week. Read four calls: two that went well and two that did not. Discuss what the agent said in the first fifteen seconds, because that is where most calls are won or lost. There is more detail on how transcription works in the platform, but the tool is the easy part. The calendar invite is the hard part.

The question every dealer asks is how many agents they still need. Automation reduces the number of low value touches per agent, so throughput per person rises. It does not remove the requirement for people, and any vendor telling you otherwise has not worked a Saturday in a dealership.

Work it backwards from volume. Take your monthly lead count, add your inbound sales calls, and be honest about how many meaningful conversations one agent can carry in a shift once the busywork is automated. Then check the coverage question separately, because coverage is what actually kills stores. A department with enough total capacity can still miss every evening and every Sunday if the schedule is built around a nine to five week.

The pattern that works for most single point stores is a small dedicated team with genuine evening and weekend coverage, supported by automation for the touches that fall outside staffed hours. Groups usually get better economics from a centralized team covering several rooftops. Either way, decide the coverage map before you decide the headcount.

Resist the temptation to build a twenty metric dashboard. Four numbers, reviewed weekly, will tell you almost everything about a BDC's health.

Median time to first genuine human response. Not the auto reply. This is the number you baselined in phase one, and it should move within two weeks of the phone routing being fixed.

Contact rate. The share of leads where a real two way conversation happened. If speed improves and contact rate does not, the problem is your message library or your phone numbers, not your cadence.

Follow up depth. The average number of attempts before a lead is abandoned. In most stores this number is lower than management believes, and raising it from three to seven is usually the single largest available gain.

Appointments set and, separately, appointments shown. Tracking only the first one teaches your team to set appointments that nobody keeps.

The platform's three reporting layers cover activity, company and management level views, and because the communication happens inside the system rather than on personal cell phones, the underlying data is real rather than self reported.

Sequence beats ambition. Here is the plan that survives contact with a real store.

Days 1 to 14. Write the one page purpose statement. Baseline response time by hand. Test your own phone tree from outside and document every dead end. Buy nothing yet.

Days 15 to 30. Stand up the phone routing and the number plan. Import contacts, run the phone validator and the email validator across the list, and load the blacklist so opt outs are honored from day one. Write the nine pieces of copy.

Days 31 to 60. Turn on lead routing rules and the first follow up process. One process, not six. Watch it for two weeks and fix what it gets wrong before adding a second. Start the weekly transcription review.

Days 61 to 90. Add drip campaigns for aged leads, appointment reminders and the reactivation sequence. Compare all four numbers to your baseline. Then decide what to change, with evidence rather than opinion.

If you want to see the pieces running before you commit to the plan, the AI BDC software page covers the toolset itself, and the platform is month to month so a ninety day trial is a real option rather than a sales line.

How long does an AI BDC implementation actually take?

Ninety days is realistic for a single rooftop that commits management attention to it. The software configuration is a matter of days. The parts that take time are writing your own message library, fixing phone routing, and building the weekly coaching habit that makes the data useful.

Can automation replace BDC agents entirely?

No, and we would not sell it that way. Automation handles routing, scheduled touches, reminders and record keeping. Conversations that involve trade values, payment expectations or a customer who is upset still need a person. Expect higher throughput per agent, not zero agents.

What should we automate first?

Lead routing and the first follow up process, in that order. Routing removes the delay caused by nobody being sure whose lead it is. The follow up process removes the touches that get skipped on a busy Saturday. Add drip campaigns and reactivation only after those two are stable.

Do we need a DMS integration or an inventory feed to start?

Neither is required. That is why individual salespeople and brokers can run on the platform. If you do have an inventory feed, Inventory Link can ingest it and advertise your actual vehicles alongside everything else.

How do we keep automated texting compliant?

Capture consent at the point of contact, keep opt out handling centralized so an unsubscribe at one point applies everywhere, and keep records. The blacklist and blacklist import exist for exactly this. Treat the compliance page as required reading before you turn on any sequence.

What if our numbers do not improve?

Then you have learned something specific rather than something vague, which is the point of baselining first. We cannot guarantee results, because lead outcomes depend on your market, your inventory and your team. What we can do is make the four numbers visible so the conversation is about evidence.

We will walk your current lead flow, show you the phone stack and the follow up processes running, and give you a straight ninety day plan. Month to month, no long term contract.