AI Follow-Up Writer for Automotive Sales

The writing was never the hard part. Getting the eleventh message sent on the eleventh day is, and that is a solvable problem.

Sit through four vendor demos and you will hear the same feature described four different ways. It is worth separating them, because they carry very different amounts of risk.

Template libraries with merge fields. Pre written messages that fill in a name, a vehicle and a store. Not intelligence, but genuinely useful, and this is what most of the value actually is.

Generated drafts a human sends. The system proposes wording, a salesperson reads it, edits if needed, and sends. Low risk, because the person is still the last step.

Fully autonomous conversation. The system writes and sends on its own, handles replies, and only escalates when it decides to. High risk, and the risk lands on your store rather than the vendor's.

We build the first two categories and we are not going to sell you the third. That is a position, not a limitation we are hiding. When an autonomous system misreads a customer, quotes a payment it should not have, or keeps messaging somebody who asked it to stop, the dealer is the one holding the consequence. Ask any vendor selling the third category what happens when it is wrong and who is liable. The answer, in every state, is you.

Here is the uncomfortable thing about this whole category. In most dealerships the messages that would have worked were never written badly. They were never written at all.

Look at your own numbers. A large share of leads receive one attempt and then nothing. A meaningful share receive a first response measured in hours rather than minutes. Almost nobody gets a fifth touch, and the fifth touch is where a surprising amount of business lives, because a shopper who was busy on Tuesday is not busy forever.

Better wording improves an eight percent reply rate to a slightly higher eight point something. Sending the message at all improves zero to eight. That is the entire argument for automation over authorship, and it is why we spend more engineering effort on the cadence engine than on prose generation.

So the question to bring into any evaluation is not whether the software writes well. It is whether, six months from now, on a Saturday in July when two people called in sick, the eleventh day message still goes out. Everything else is commentary. There is more on how we think about this on the AI automotive CRM page.

Named specifically, so you can check each one on a screen rather than take our word for it.

Follow up processes that define what happens on day zero, day one, day three and onward, with the step type and the owner attached. Drip leads with a drip editor for sequences that run on a schedule. Automations and automation lists for the event driven cases, like a lead arriving after hours or a customer replying to a campaign. Task automation that puts a call in front of the right person with the context attached rather than as a note somebody has to decode. Reminders and appointments so the human steps have the same reliability as the automated ones.

On the delivery side, a step can send SMS or MMS, send RCS with automatic SMS fallback so a branded message degrades gracefully on older handsets, send email through the composer, drop a prepared voicemail without ringing the phone, or route a task by distribution rule. Bulk email with recipient management handles the wider sends.

Two supporting tools matter more than they sound. An email validator and a phone validator, both built in house, check contact data before a sequence starts pushing messages at addresses that bounce and numbers that were disconnected in 2022. Deliverability is easier to protect than to repair.

Message templates sit across all of it, so the wording your best closer uses becomes the wording the whole floor uses. Our follow up templates page has examples worth stealing.

The reason most drafted follow up reads generic is that the system writing it knows almost nothing. It has a name, a vehicle of interest and a timestamp. The salesperson, meanwhile, had a nine minute phone conversation in which the customer explained that his lease ends in November, his wife wants a third row, and he is not moving until the pool is closed for the season.

None of that reaches the record in most stores, because writing notes after a call is the first thing that gets skipped on a busy day.

Call recording with transcription changes the economics of that. The conversation becomes text attached to the customer, which means a manager can read what happened in thirty seconds instead of listening to nine minutes, and the next person to touch that customer opens a record with actual content in it. Follow up written against a real transcript stops sounding like a form letter, whether a person or a template produced it. More detail on the call transcription page.

This is also the single best coaching tool in the platform, and coaching is what actually improves your follow up writing over a year. A manager who reads ten transcripts a week learns more about why deals stall than any dashboard will tell them.

Vendors sell the engine and leave you to design the sequence, which is backwards, because the sequence is where the result comes from. A workable starting structure looks like this.

Minutes, not hours, for the first touch. Text first if you have consent, because text is answered and voicemail is not. Then a call. Speed of first contact is the most controllable variable in the whole process.

Day one through day three, mixed channels. A text, a call at a different hour than the first one, and an email that contains something worth opening rather than a restatement of the first message.

Week one through week two, decreasing frequency. Every second or third day, still mixing channels, with at least one message that asks a genuine question instead of checking in.

Beyond two weeks, long tail. Monthly, with a reason: a price change on the unit, a similar vehicle arriving, a lease maturity date approaching.

Two rules matter more than the schedule. Every message should carry one clear question a customer can answer with a thumb, and the sequence must stop the instant a human being replies. A cadence that keeps firing after a customer answers is worse than no cadence, because it proves nobody is reading. Our follow up cadence page goes further into the timing.

Whether a template, a salesperson or a drafting tool produces the text, the same things make it work. This is craft rather than technology, and it is where a store gets the compounding advantage.

Keep it under two lines on a phone screen. Say who you are and where you are from in the first message, once, and never again. Name the specific vehicle rather than referring to your inquiry. Ask exactly one question, and make it answerable in three words. Avoid the phrases every shopper has learned to ignore, which are checking in, touching base, and just following up.

Write like a person who works at a car store, because that is what you are and shoppers can tell. A message that reads like marketing gets treated like marketing. And when a customer replies with something the template did not anticipate, the whole automated sequence should get out of the way immediately so a human can have a normal conversation.

One more thing worth saying out loud: send fewer, better messages. A store that texts a lead nine times in two days has not been persistent, it has been reported as spam, and that damage is carried by every message the store sends afterward.

Some decisions do not belong to software, and we would rather be explicit about which ones.

Anything a machine drafts should be read by a person before a customer sees it. Anything a machine infers about a customer's credit, income or intent is a guess and should be treated as one, and it must never be used to narrow an audience in a way fair lending rules forbid: age, gender, income, marital status, household size, education, language or ZIP. Anything a machine schedules must stop the moment a human replies. And numbers, meaning payments, rates and terms, come from the desking engine rather than from generated prose, because a payment quoted in a text message is a payment your store owns.

None of this is caution for its own sake. It is the difference between automation that survives its first bad week and automation that gets switched off after one furious phone call from a customer.

Automated messaging multiplies whatever your consent practices already are, in both directions. A store with clean consent handling gets more reach. A store without it gets more exposure, faster, with a written record of every single send.

The platform treats consent as a property of the customer record rather than of a campaign, which is the only arrangement that survives contact with reality. Opt outs are honored everywhere, including in a campaign somebody builds six months from now who was not in the room when the rule was set. There is a blacklist with import, so numbers you must never contact stay uncontacted. Every message is logged with a timestamp, which is exactly what you want when a complaint arrives eleven months later.

Practical habits that cost nothing: respect quiet hours in the customer's own time zone rather than yours, make the opt out language unmistakable rather than clever, and never let a sequence restart because a record was re imported. Our page on TCPA compliance for dealership texting covers the ground properly. None of this is legal advice, and your counsel should review your program.

Follow up is one of the few things in a dealership you can measure honestly, provided the communication happens inside the system rather than on personal handsets.

Watch four numbers. Time to first response, by hour of day, because the overnight number is usually the ugly one. Touches per lead before the store stops, which is almost always lower than the sales manager believes. Reply rate by channel and by message position in the sequence, which tells you which specific message to rewrite. And appointments set per hundred leads, which is the only one that pays the bills.

Three reporting layers cover it: activity reporting for what individual people did, company reporting for store performance, and management reporting for the roll up. Because a cadence is a defined object rather than a habit, you can change one message, watch the reply rate on that step, and know what caused the difference.

Pricing is month to month with no long term contract. CRM Only starts at $199 a month and includes the automations, follow up processes, drips, templates, transcription and the full messaging stack. Programs that add exclusive local leads start at $799. Details are on the pricing page. We cannot guarantee reply rates, appointment counts or sales, because the customer controls most of that and your people control the rest. What software can do is make the disciplined version the easy version, and then show you the numbers afterward.

Does LeadLocate write follow up messages for me automatically?

It ships message templates with merge fields, follow up processes, drip sequences and automations that send on a schedule or on an event. Anything a machine drafts should be read by a person before a customer sees it, which is a position we hold deliberately.

Will it hold a full conversation with a customer on its own?

No, and we will not sell that. An automated sequence stops the moment a human being replies and hands the conversation to your salesperson. Autonomous conversation puts the liability on your store rather than on the vendor.

What channels can a follow up step use?

SMS and MMS, RCS with automatic SMS fallback, email through the composer, a dropped voicemail without ringing the phone, or a task routed to a person by distribution rule. Bulk email with recipient management handles wider sends.

How does call transcription improve follow up?

It turns a phone conversation into text on the customer record, so the next message is written against what the customer actually said rather than against a name and a timestamp. It is also the fastest way for a manager to coach message quality.

How do we stay compliant with automated texting?

Consent and opt out status live on the customer record and are honored across every campaign, including ones built later. There is a blacklist with import, every message is logged with a timestamp, and quiet hours should follow the customer's time zone. This is not legal advice.

How many follow up attempts should we make?

More than most stores make. A workable structure is minutes to first contact, mixed channels through the first two weeks with decreasing frequency, then a monthly long tail with a real reason to reach out. Every message should ask one answerable question.

We will build a follow up process for your store, wire the templates and the transcription, and show you the reply rates by step. Month to month, no long term contract.