By Sagar Shankaran, Founder of CallSphere
Which jobs on an independent used-car lot change shape when AI drafts the work, what a new title clerk learns in week one, and what leaves the job posting.
Key takeaways
That is the objection, and it is a fair one. Two-thirds of US small businesses are using AI in some form now — 66 percent, up from 55 percent a year ago — and around 70 percent of owners say their people need more training to actually use it. On an independent lot with nine employees, "training" sounds like a day off the floor you cannot spare during tax season.
But look at what that statistic is really saying. The tools stopped being the problem. Nobody on your lot is blocked because the software does not exist; they are blocked because nobody showed them what to hand over and what to check. That is a training problem with a specific answer, and it takes about six hours in week one, not a semester.
So here is the useful version of the question: on a store selling forty-five to seventy units a month, which jobs actually change shape, what does a new hire need to be taught in the first week, and what quietly stops being a hiring requirement?
Your chart probably reads: owner or dealer principal, a sales manager who also desks deals, one or two salespeople, a finance manager, a title clerk, a collections rep if you carry paper, a recon or lot porter, and somebody answering internet leads whether or not you call it a BDC.
The title clerk changes the most, and nobody expects that. Her week is the state title portal, odometer disclosures, lien perfection, temporary tags, out-of-state transfers, and chasing a payoff letter from a credit union that answers the phone twice a day. An assistant can now draft the whole packet from the deal jacket, check that the buyer's name matches the title exactly, flag the mileage discrepancy before it becomes a brand, and tell her which of thirty pending titles are past the state's filing deadline this week. Her job shifts from typing the packet to verifying it and working the exceptions. That is a better job and it is also a harder one.
The internet lead person changes second-most. Two hundred to four hundred leads a month, most of them "is this still available," a handful that are real buyers. Drafting the response stops being the work. Deciding which twelve leads get a phone call today becomes the work.
The finance manager changes least in shape and most in speed. He still owns the credit decision, the disclosures and the lender relationship. What he stops doing is memorising nine program guides.
Hear it before you finish reading
Talk to a live CallSphere AI voice agent for auto shop in your browser — 60 seconds, no signup.
The salesperson barely changes at all, which is worth saying out loud, because most of what a good used-car salesperson does happens in a walk-around and a test drive in a parking lot in February.
flowchart TD
A["New hire, day two"] --> B["Watch the assistant draft a title packet"]
B --> C["Check every field against the title and odometer statement"]
C --> D{"Did it get anything wrong?"}
D -->|Yes| E["Correct it and write down exactly why"]
E --> F["Note goes into the lot's house rules file"]
F --> B
D -->|No| G["Clerk signs and files in the state portal"]
Do not start with the tool. Start with the paper. A new title clerk or lead coordinator should spend her first morning with a completed deal jacket open on the desk — retail installment contract, Buyers Guide, odometer disclosure, credit application, GPS disclosure if you use one, we-owe — and learn what each document is for and whose signature makes it binding. If she does not know what a correct odometer statement looks like, she cannot catch a wrong one.
Afternoon of day one: the two lines she may never cross. One, no customer Social Security number, no credit report, no bank statement goes into any tool that has not been approved by you, because your store is a financial institution under the FTC Safeguards Rule and that is your license and your bond on the line. Two, she never signs or files anything she has not read field by field against the source document.
Day two through four: the loop in the chart above. She watches the assistant draft, she checks it against the actual title, she corrects it, and every correction gets written into one shared file — the lot's house rules. "Our stock numbers start with the acquisition month." "Never abbreviate the buyer's middle name." "If mileage is over 100,000 and the title says exempt, stop and ask." After two weeks that file is worth more than any vendor's manual, because it is about your store.
Day five: teach her to spot a confident wrong answer. Give her three drafted packets, one of which you have deliberately broken — a transposed VIN, a payoff figure that is a week stale, a lienholder address for a branch that closed. If she finds all three, she is ready. If she finds two, she does day five again.
Look at the ad you ran last time you hired. Some of it is now noise. "Fast, accurate data entry" is not a differentiator anymore. Neither is "experience with our specific dealer software," because the parts of that job that were software-specific are the parts being drafted for her. Neither is "strong written communication for internet leads," at least not as a filter.
What replaces them: attention to detail on documents, willingness to say "this doesn't look right," and comfort making a phone call to a stranger. Those have always been the qualities of a good title clerk. Now they are the whole job description instead of half of it. Practically, that widens your hiring pool — you can hire the meticulous person from a bank branch or a doctor's front office who never touched dealer software, which in a tight labor market in a small town matters more than any software feature.
The usual pitch is "cut a position." That is not what happens on a lot this size. What happens is you grow volume without adding the next person. Suppose you go from forty-five to seventy deals a month over a year — the growth that normally forces a second title clerk.
Still reading? Stop comparing — try CallSphere live.
See the auto shop AI agent handle a real call — complete, industry-specific, and live in your browser. No signup.
| Item | Assumption | Figure |
| Title work today | 38 minutes per deal, 45 deals | 28.5 hours a month |
| At 70 deals, unchanged | 38 minutes per deal | 44.3 hours a month |
| With drafting and exception flagging | 22 minutes per deal, 70 deals | 25.7 hours a month |
| Second title clerk avoided | Part-time, loaded | $28,000 a year |
| Week-one training cost | 3 staff x 6 hours at $22 | $396 once |
| Weekly 30-minute review for 8 weeks | 3 staff | $264 once |
Six hundred and sixty dollars of training against a position you did not have to fill. The number that would kill this is turnover: if the person you trained leaves in month four, you pay the six hundred and sixty again and you lose the house rules file if it lived in her head. Which is exactly why it goes in a shared document from day one.
Anything with a signature and a regulator behind it. The odometer disclosure, the retail installment contract, the Buyers Guide, the adverse action notice under the Equal Credit Opportunity Act when you decline someone — a person reads and a person signs. The assistant drafts.
Anything where the customer is upset. A collections call about a car that broke down two weeks after delivery is not a workflow. It is a conversation about whether this person keeps paying you for three more years, and it belongs to your collections manager.
And your own judgment about people. Do not let a tool score your applicants or your customers into categories you cannot explain. Several states now have their own AI statutes — Texas and California rules took effect at the start of 2026, with Colorado, New York, Utah, Nevada, Maine and Illinois passing their own — and federal preemption is unsettled as of this July. If you cannot explain in plain English why a decision was made, do not automate the decision.
Start with the thing she complains about most, which is almost never the typing — it is chasing payoff letters and catching state filing deadlines. Let the assistant produce a Monday list of every pending title, how many days old it is, and what is missing. She will use that on day one because it makes her look good, and everything else follows from there.
No. If you hire a technical person to run this on a nine-person lot you have made an expensive mistake. The person who should own it is whoever already owns your deal jackets, because the value is in knowing what a correct packet looks like, not in knowing how the software works.
Leave him alone if he sells cars. The seats where this matters are title, finance and internet leads. A salesperson who closes on the walk-around does not need a tool, and forcing one on him is how you lose him to the lot down the road.
The broken-packet test from day five, run again at ninety days with new mistakes planted. If your clerk still catches the planted errors after three months of the assistant being right most of the time, the checking habit survived. That habit is the whole point of the training.
One seat worth mentioning: the phone. On most independent lots the internet lead person is also whoever happens to be near the phone, and the calls that matter arrive at 7pm and on Sunday when the lot is closed. CallSphere builds AI voice and chat agents that answer the line and the website chat around the clock, answer the availability questions, and book the appointment or capture the lead so your people start Monday with a call list instead of a voicemail box. It does not do title work or write your contracts. It keeps the seat covered at the hours you were never going to staff.

Written by
Sagar Shankaran· Founder, CallSphere
LinkedInSagar Shankaran is the founder of CallSphere, where he builds production AI voice and chat agents deployed across healthcare, hospitality, real estate, and home services. He writes about agentic AI, LLM engineering, and shipping voice agents that handle real calls in production.
See how AI voice agents work for your industry. Live demo available -- no signup required.
Independent lots key each auction VIN into six systems. In 2026 assistants write inside the dealer management system itself. What it saves on a 60-car lot.
Ten-key speed stopped being the job. A week-one training plan for a logging office hire, what to drop from the ad, and what AI teaches nobody in the woods.
US small-business AI adoption hit 66% but 70% of owners say staff need training. What changes on an optometry org chart and what week one must now cover.
Independent dealers shotgun credit apps because nobody can memorise nine program guides. In 2026 the whole binder fits in one question. The cash-flow math.
A new client service rep used to need six weeks on Cornerstone. Here is what to teach in week one now, what stops being a hiring requirement, and the cost math.
A cloned voice can get a collections clerk to release a starter-interrupt code or reroute a payoff wire. The three-step callback rule that stops it, priced out.
© 2026 CallSphere Inc. All rights reserved.
Made within San Francisco
Watch how CallSphere handles real customer calls, schedules appointments, and processes payments — live.
Try Live DemoBook a DemoCalculate Your ROI