There is a fortune to be made in automotive AI, and plenty of ways to burn it.
McKinsey estimates that agentic AI could add $450 to $650 billion in annual revenue across advanced industries by 2030, with automotive among the sectors set to benefit. Then there is the other number. Gartner expects more than 40 percent of agentic AI projects to be canceled by the end of 2027. The problem is rising costs, unclear business value, and weak controls, rather than any failure of the technology.
$450–650bn
Annual revenue agentic AI could add across advanced industries by 2030, with automotive among the sectors set to benefit
McKinsey
40%+
Share of agentic AI projects expected to be canceled by the end of 2027
Gartner
Reading those two figures together brings the real question into focus. The debate over whether automotive should adopt AI is settled. The open question is whether an organization is building the foundation to make AI useful at scale, or simply buying more tools and hoping they add up to something.
I have rolled AI out across 25 markets at Seez, and one lesson has held in every one of them. The technology itself is rarely the hard part. The hard part is everything around it.
The market has an AI fatigue problem
When I talk to dealer principals today, I hear a familiar exhaustion. Every vendor now has an AI product, and most are chatbots with a new label or a thin layer of automation on top of something that already existed. Many leaders interpret that fatigue as a reaction to too much AI, which misses the problem entirely.
Dealers aren’t tired of AI. They’re tired of disconnected technology that promises intelligence and just gives them one more system to manage.
Consider what happens on the ground. A dealership buys a chatbot from one vendor, a lead nurturing tool from another, a service reminder system from a third, and an engagement app from a fourth. Every one solves a narrow problem and brings its own dashboard, data silo, training, and quarterly invoice. Teams wanted automated support to free up their time. Instead, they got four more systems to oversee and less time to serve customers.
The fatigue comes from the fragmentation itself. When AI arrives as one more disconnected point solution, it adds friction instead of taking it away. The catch is that each of those tools can only see what its own system holds. A chatbot talking to a customer has no idea which car they are asking about, while a tool inside the CRM makes calls with no view of inventory, finance, or service history. Every one of them is working from a fraction of the picture.
If your customer has to repeat their story every time they move from a chatbot to a service advisor, the AI isn’t helping. It’s just creating noise.
AI That Helps vs. AI That Does
The real dividing line separates AI that assists people from AI that does the job on its own.
Generative AI
Writes the follow-up email and waits for someone to send it.
Operational AI
Qualifies the lead, updates the record, books the appointment, and only pulls in a human when the situation genuinely needs one.
One approach adds a task to someone’s list, and the other takes it off.
What lets operational AI act is the thing the fragmented tools never had: a connected view of the business. The dealerships getting real value rarely have the fanciest model. What they have is AI that can see across sales, service, parts, finance, and customer history, and act on all of it. Data connectivity isn’t an IT detail. It decides whether your AI keeps getting smarter or quietly hits a wall.
Leaders weighing where to spend over the next two years can test any product against three things:
- VisibilityThe tool has to actually see the data behind a decision, not start blind every time.
- GovernanceSomeone has to decide what the AI can touch, and where a person steps back in.
- Real payoffIt should save time or make money in the first month of use, not just look good in a demo.
Most projects that stall never fail because of the AI model. The cracks show up in one of these three long before anyone blames the technology.
How this looks in practice
Rather than adding another standalone feature, we designed CARS at Seez as the native AI layer built directly into the Pinewood.AI Platform and core dealership systems. It deploys four specialized agents (Connect, Assist, Reach, and Serve) to drive a continuous workflow across the customer journey. The number of agents matters less than the fact that they share the same data and context. The one handling a fresh lead is working from the same picture as the one who books a service visit months later.
Here is how that operates day to day:
- In SalesWhen a lead enters the Lead Management System (LMS), the AI agents process the inquiry while respecting structured Deal File rules. Mandatory documentation, finance requirements, and signature controls are applied automatically based on the parameters of the transaction.
- In AftersalesWhen the AI agents send maintenance communications, customer history and live workshop records align instantly. Customers can review details and select relevant options through an Online Check-In Portal, updating the core platform before they even set foot in the dealership.
Pulling from connected data and clear business rules gives the AI real context to work with and far less reason to guess, though no integration makes AI perfect.
Our aim was simple. We didn’t want to give dealership teams one more thing to manage. We wanted the AI to carry the repetitive work inside the systems they already use.
Architecture Beats Speed
Many leaders treat AI as a sprint and assume early adopters automatically win, but the evidence suggests otherwise. The dealerships that succeed this decade will be those that build a proper foundation, treating AI as an operational capability built on integrated systems, fluid workflows, and strict governance. These fundamentals determine whether investments compound or stall.
If I had to reduce the whole test to one line, it would be this.
Your AI is only as good as the data it can see, the workflows it can access, and the actions it can take.
The $650 billion is real, and so is the 40 percent that will never reach it. What separates the two is the quality of the foundation underneath the ambition.
Sources: Revenue figures from McKinsey, “Empowering advanced industries with agentic AI.” Cancellation forecast from Gartner press release, “Gartner Predicts Over 40% of Agentic AI Projects Will Be Canceled by End of 2027,” 25 June 2025.
Ready to build the foundation, not just buy more tools?
Tell us where your dealership is today, and our team will show you what connected, operational AI could look like across your sales, service and customer journey.
