"I have not failed. I’ve just found 10,000 ways that won’t work."

—Thomas Edison, Inventor

Ford's new AI assistant is rolling into the Ford and Lincoln apps, with plans to reach up to 8M customers.

It already knows the owner's vehicle well enough to answer questions about tire pressure, oil life, features, troubleshooting, and even whether that questionable Facebook Marketplace furniture haul will fit in the back. Ford plans to bring the assistant directly into select vehicles in 2027.

That owner's manual is starting to look a little intimidating.

Google is also pushing their AI platform Gemini deeper into Android Auto and Google built-in, which now spans more than 100 models across 16 brands. GM says roughly 4M newer vehicles are eligible for its Gemini rollout, while it works on its own vehicle-specific AI.

Apple keeps moving further into the cockpit too. CarPlay is expanding what developers can build for the vehicle, and Ford recently announced that Apple Maps will be baked directly into its upcoming UEV platform in 2027, with mapping data also supporting future hands-free driving tech.

This all shifts the dealer opportunity toward what happens next: onboarding customers into these systems, teaching the tech at delivery, and making sure service pathways stay connected when the vehicle starts diagnosing, explaining, and recommending on its own.

The dashboard is becoming part assistant, part support desk, and part storefront.

Dealers should know who's sitting behind it.

Meta just released Muse Glimmer, a 30-billion-parameter open-weight model built to run locally on a Mac or PC with a consumer GPU.

That means the model can handle agent-style work without every task depending on a cloud connection. It can call tools, work through multi-step requests, review documents, interpret screenshots and images, assist with coding, and recover when a tool call fails.

The technical unlock is compression. At full precision, a model this size would need more than 55 GB of memory. Meta says Muse Glimmer can be compressed to roughly 4-bit precision, bringing the language model under 20 GB so it can run inside a 24 GB or 32 GB hardware setup.

For dealers, the reason to pay attention is data proximity.

A useful agent needs context. Customer notes, inventory files, pricing data, service history, call summaries, internal process docs, and store-specific rules. The more context a tool needs, the more important it becomes to know where that information is going, who can access it, and what the system is allowed to do with it.

Muse Glimmer will not be plug-and-play for most stores tomorrow. It is more developer-facing today, with support coming through the technical tools developers use to run models locally, customize them, connect them to apps, and deploy them at scale (such as GitHub, Unsloth, and Hugging Face).

The market is moving toward agents that can run closer to the business, closer to the workflow, and closer to the data.

For dealerships, that could eventually mean faster internal tools, more private automation, and less dependence on sending sensitive store context into a black box.

Your customers are going to show up with more questions about AI assistants, apps, subscriptions, software updates, digital keys, driver-assistance features, and whatever gets added in the next OTA update.

This prompt turns that into a training plan before the questions hit the showroom.

Your best content may already be walking around the dealership.

This episode breaks down how AI can pull useful ideas from everyday conversations with service writers and product specialists, then turn that knowledge into content people can actually find.

You’ll also hear where automation can get sloppy, why human review still matters, and how stores can use AI without piling another task onto the team.

Watch the full episode and steal a few ideas for your own content process.

Thanks for reading, Friend! Remember that there’s rarely one magic move. So, keep testing, keep learning, and keep a healthy suspicion of anyone selling guarantees.