FOR DEALERSHIPS
Train your dealership into AI.
You stop losing deals to the dealer down the street who replies in twenty minutes. A specialized model trained on your live inventory, your financing programs, and how your best closer actually talks — answering on every surface where the buyer already is.
What it knows.
- – your live inventory — every VIN, trim, photo, mileage, price
- – your financing programs — lender tiers, down requirements, credit-union quirks, the rates you can actually quote
- – your sales process — qualification questions, test-drive scheduling, trade-in flow, F&I handoff
- – your service policies — warranty terms, recommended intervals, open recalls by VIN
- – your store voice — how your best closer talks, when to push, when to back off, when to switch to Spanish
- – your trade-in priors — what a 2022 Tacoma actually trades for in your market this week, not what the book says
Where it shows up.
- – your website chat, every hour you are not on the lot
- – inbound and outbound SMS, in English or Spanish, on the buyer's phone
- – after-hours phone, with a calm voice that knows the floor
- – service-department triage, sorting recalls from oil changes before the advisor sees the ticket
- – email reply drafts for trade-in inquiries, ready in the BDC inbox
- – internal sales-coach pings — your closer pasting a buyer text, getting the next move
A REAL TUESDAY
What a trained dealership AI actually says.
An inbound SMS to the National City rooftop, ten minutes after the showroom locks. The AI handles it in your voice and books the test drive before the buyer goes to bed.
9:47pm Tue · Buyer
9:47pm · Trained AI
9:49pm · Buyer
9:49pm · Trained AI
9:51pm · Buyer
Dialogue is illustrative. VINs are sample patterns; exact rates and trade values depend on your live programs and inspection.
Before and after.
Before Gpodz
- – after-hours SMS sits unread until 9am; the dealer down the street already replied at 8:12
- – your BDC quotes ballpark payments off a yesterday-old rate sheet
- – the closer who actually knows the store is doing rote inventory lookups instead of closing
- – service triage runs through the advisor for every recall question
After Gpodz
- – every inbound SMS, chat, and missed call answered in your voice, in minutes, around the clock
- – quotes pulled from your live program sheet — bilingual when the buyer's phone is bilingual
- – your closer gets coaching pings, not lookup chores
- – service triage sorts recalls from oil changes before the advisor opens the ticket
EARLY OUTCOMES
What we expect to measure.
Real numbers, once design-partner runs land. Placeholders below are explicit; we publish measured values, not projections.
- – inbound-lead reply time: TODO (target: minutes, not hours) — Sprint 005+ design-partner data
- – after-hours SMS conversion lift: TODO % — measured against the same store's prior 30-day baseline
- – BDC time recovered: TODO hours/week — replacing rote inventory-lookup chores
- – recall-routing accuracy: TODO % — against advisor double-check
Numbers will be replaced with measured values from the Marco Delgado design-partner run (specs/010-paid-beta) once the agreement signs and the training job completes.
A model that knows your floorplan beats a generic one every time.
A generic AI has never read your inventory feed, never seen your lender tiers, never heard the way your closer pivots when a buyer flinches at the payment. A specialized model trained on your data has. It quotes the program you can actually book, not an average rate from a finance blog. It names the VIN sitting on your lot tonight, not a stock photo. It switches to Spanish when the thread switches. The voice is yours; the math is yours; the inventory is yours — because the model was trained on yours, on shared infrastructure, for a small fraction of frontier rates at SMB volumes.
Trust, quietly.
Three guarantees underneath every reply. The mechanics live on the trust page.
- – Eval-gated replies. If the model is not confident on a trade value or a payment quote, it routes to your desk instead of inventing.
- – Verifiable billing. Every charge is Ed25519-signed and verifiable offline. Designed to start at Ready.
- – Your data stays yours. Your inventory and your transcripts never train Gpodz platform models, never get pooled across tenants, never get sold downstream.
Same shape, next door.
The dealership pattern — recurring relationships, voice matters, the moat is local knowledge — repeats one block over.
- – Realtors — lead-follow-up SMS, missed-call rescue, listing drafts that know the submarket.
- – Insurance brokers — quote SMS, renewal reminders, an internal copilot that knows your carrier appetites.
- – Contractors — SMS estimates trained on your pricing brackets and crew schedule.
Train the AI that knows your dealership.
Start with one rooftop. Ship it across every surface. Iterate from real conversations.