I'm taking a swing at solving for this right now with an open source project.
I think it goes beyond signals from inventory, CRM, ads, comms. Structure it. Push it back out so every platform reads from the same source instead of guessing on its own.
VCTRS.IO is my attempt at that. Three layers: what's declared, what's observed in the systems, and what's stuck in people's heads and never written down anywhere. That third layer is where it actually gets interesting. Most context plays stop at layer two. This industry has operated on gut for so long. That gut instinct or tacit layer might have to be a managed service.
Still early. Don't know what it becomes yet. The part I care most about is the open plugin layer. Dealers or anyone else can build their own and monetize it instead of guessing every use case for them. So far - beta users have come up with interesting things like creating employee schedulers, commission plan automations, employee onboarding, vendor digital rooms, or plug-ins as canaries in the coal mine so they dont have to have 30 windows open at a time.
Almost ready to open it up. It will be free. Bring your own keys and/or AI.
My apologies...this is me trying to understand:
You believe dealerships don't actually have a marketing problem, they have an
information problem.
A dealership has:
- Website
- CRM
- Inventory system
- Google Business Profile
- Meta Ads
- Google Ads
- Service scheduler
- Phone system
- HR software
- Accounting
- Employee handbook
- Training documents
- Slack
- Email
- DMS
- OEM portals
Every one of those knows something.
None of them know everything.
And the
most important information isn't inside any software. It's inside someone's head!
Example:
The GM knows:
"We always sell more Wranglers after the first snow."
Nobody wrote that down.
The marketing agency doesn't know.
The website doesn't know.
The CRM doesn't know.
The AI doesn't know.
And the guy that does know it just retired!
That knowledge disappears forever.
Layer 1 = Declared
Facts someone intentionally entered.
Example:
- Dealership Name
- Address
- Hours
- Phone Number
- Inventory
- Employees
- OEM Certifications
- Pricing Rules
Layer 2 = Observed
Things software notices.
Example:
- Lead response time
- Sales trends
- Inventory aging
- Ad spend
- Website traffic
- Call recordings
- Appointments
- API usage
- Search rankings
Layer 3 = Knowledge
This is what lives inside people.
Examples:
- Don't advertise lifted trucks to retirees.
- Jim closes truck buyers better than anyone.
- Avoid wholesale purchases in November.
- Customers from this county negotiate harder.
- Always feature SUVs before football season.
- The owner's wife hates blue homepage banners.
- This OEM rep answers texts faster than email.
None of this exists anywhere.
But it drives the dealership.
Your Engine
Dealership Context
Inside:
- Identity
- Goals
- Processes
- Employees
- Knowledge
- Preferences
- History
- Rules
- Performance
- Inventory
- Customers
- Market
- Culture
Every application reads from this.
Instead of
Website
↓
CRM
↓
Ads
↓
Inventory
↓
HR
↓
Accounting
↓
ChatGPT
all maintaining separate versions...
they all ask;
Context Engine
↓
Give me the truth.
Instead of having a brain, everything is connected to the nervous system.
However you don’t think you can predict every workflow, so you have plugins, kind of like WordPress.
Your not
trying to compete with AI models.
Your assuming AI will keep changing, so you allow the dealership to connect whatever AI they want to use.
Everything plugs into your system?
However you need:
- permissions
- synchronization
- conflict resolution
- security
- governance
- schemas
- APIs
- versioning
- plugins
- user management
- auditing
You're basically describing Salesforce, Notion, Zapier, LangChain, and an enterprise knowledge graph rolled into one.
Am I thinking about this correctly?
