Product 02 · pilot module

Let shoppers ask for the car their way.

“Three rows, AWD, under $45,000” should work. DealershipMCP turns natural-language intent into inventory matches grounded in the dealer-approved feed—not a generic chatbot answer.

Current status: the Howard Bentley search API and dealer query-receipt pipeline are operating. The shopper-facing website module is in pilot development and is not yet installed on the dealer’s public website.

HOWARD BENTLEYAI INVENTORY SEARCH / PILOT
I need a full-size truck under $60k that can tow a camper.
Searching dealer-approved inventory
2026 GMC Sierra 15004WD · Price evidence available
2026 GMC Sierra 15004WD · Verify towing configuration
What matters more: maximum towing or cabin comfort?

Not another chat bubble

The inventory is the model’s ground truth.

The experience should understand shopper intent without inventing price, availability, equipment, or eligibility.

01

Natural-language discovery

Search by lifestyle, budget, body style, drivetrain, mileage, features, and combinations rigid filters make painful.

02

Retail-safe answers

Carry source time, price state, missing charges, conditional offers, and verification requirements into the answer.

03

Dealer query intelligence

Record privacy-filtered questions, result counts, matched VINs, and outcomes in the dealer-only dashboard.

04

One data layer

Use the same approved retail facts on the dealer website and through the external Agent Gateway.

The feedback loop

Every question tells the dealer what demand looks like.

SHOPPER QUESTION“Low-mileage SUV under $35k”
GROUNDED RESULTSMatched VINs + evidence
DEALER DASHBOARDDemand, gaps, outcomes

Founding dealer pilot

Turn your inventory into a conversation.

We will scope the approved feed, website placement, brand controls, and measurement plan before installation.