“Software is eating the world, but AI is going to eat software”, said Nvidia CEO Jensen Huang. Dealers are seeing this firsthand in many ways as the halls of NADA are filled with customer follow-up, Voice AI, Generative Engine Optimization, and more. In a world where ChatGPT reached 100 million users in two months back in 2022, every vendor suddenly adopting and pivoting into AI is actually a positive sign that Automotive contains creative entrepreneurs and vendor partners that are here to make America’s car dealerships better.
Many feel overwhelmed by the AI progress. I recently spent 3 days in San Francisco attending “Off the Record” talks by “AI-pilled” entrepreneurs predicting “nobody will have to work in 10 years” and that we will have “robots building robots to do all of our jobs.” Maybe this will happen. Maybe not. In the meantime, let’s focus on incremental progress with the goal of giving our employees and customers better experiences on a daily basis.
Cars are a necessity across most of the country. People need to drive to work, so demand for sales and service isn’t going away. The problem has always been capacity: dealers can only staff so many people, and customers don’t stop needing help just because it’s midnight or a Sunday. That’s where AI CRMs and customer follow up platforms came in. They started handling the simplest gaps first: responding to web and chat inquiries after hours, across every department, so leads and customers weren’t left waiting until the next business day. In 2025, Voice AI picked up momentum, especially in service departments. Studies show 30% of service calls go unanswered industry wide. That’s a lot of appointments and revenue slipping away. A well trained voice agent, connected to customer data and a scheduling tool, fills that gap easily. It can answer the call, pull up the customer’s history, and get them booked. Now in 2026, the competition is shifting again. Dealers are fighting for visibility not just in search results, but in AI generated answers and agentic experiences. This is Generative Engine Optimization, or GEO. Website providers are starting to build for agent readability the same way they used to build for SEO. And this isn’t a small trend: over 25% of search volume in automotive has already shifted from SEO to GEO in 2026 alone. But what’s next? The above strategies (AI customer follow-up, Voice AI, and GEO) are well-known strategies for 2026. I am interested in exploring what 2027 and beyond is likely to bring for Auto Groups and Dealers. I have two hypotheses for the near-future and will throw out a far-future hypothesis as well.
Dealers and Auto Groups are just starting to adopt enterprise grade licenses from Anthropic (Claude), OpenAI (ChatGPT), and Microsoft (CoPilot). As executives adopt these tools and want to retain privacy and shared communication records, groups are increasingly supplying their teams with these platforms. One of the most powerful use-cases of LLMs and working natively within a frontier LLM is the ability to connect a Company’s data and create super employees equipped with insane amounts of data just one query away.
Imagine your smartest executive with every live metric right at their fingertips. Or your best operator with a live dashboard just a query away. Or your most creative vendors connected to each other to inform your next marketing campaign. This is possible with Claude, OpenAI, or Copilot now.
Vendors are also beginning to offer their data and workflows in the forms of Model Context Protocol (MCP), which is similar to an API for LLMs. Lectrium for example now allows customers to query their inventory data (and proprietary EV data) directly within their LLM of choice. The question “Analyze my Electrified Inventory” will look a lot different for a dealer working with Lectrium vs a dealer who does not. Via MCP, this dealer will access their live inventory and lead data along with Lectrium’s pre-built “Dealer Skill” which organizes and surfaces the data in a clean method designed for easy consumption and immediate time to value. This model will likely permeate throughout the Automotive vendor ecosystem. Lectrium started as a pure SaaS company offering an EV/Hybrid merchandising tool, and is evolving into a data business that surfaces information in its traditional format (SaaS) but also within an MCP for dealer use (AI application).
Dealer Skills will likely take off in 2027 and beyond as recurring tasks get automated and “Workers” turn into “System Architects” who optimize their way to better results with the help of frontier models.
Most dealers don’t want to build an AI agent. They want the outcome: fewer missed follow-ups, faster rebooking, more revenue captured while nobody’s watching. That gap between wanting the result and being able to build it is exactly what ‘Managed Intelligence Providers’ are built to close. Lectrium is building agent harnesses (the technical scaffolding that lets an agent take real action, not just answer a question) for this purpose, and it’s a reasonable bet that Lectrium won’t be alone. Most software categories in automotive eventually get an AI-agent layer from more than one vendor, and this will likely be no different. The broader data backs up why this matters now: Gartner projects 40% of enterprise applications will include task-specific AI agents by the end of 2026, up from less than 5% in 2025, and an estimated 31% of enterprises now run at least one AI agent in production. A dealer doesn’t need to become an AI engineering shop to benefit from agents; they need a partner who builds the agent correctly, monitors it, and takes responsibility for how it performs once it’s live. Whether that partner ends up being Lectrium or another vendor in the space, dealers should expect this “buy it built and managed” approach to become one of the standard ways AI shows up in dealerships over the next few years.
We experienced the “build with AI” cycle the past 2 years with the rise of vibe-coding, Claude Code, and easy content generation at scale. The next phase is “building for AI.” Some large automotive marketplaces have already launched their own GPTs or Apps in LLM marketplaces where shoppers can shop natively within their LLMs. It reasons that large Auto Groups may do the same in the not so distant future. We witnessed the first agent-to-agent car deal in the USA completed with DMC-12 at Mark Miller Subaru. As AI shopping and commerce continue to grow, will agents eventually buy thousands or even millions of cars per year in the USA? Time will tell.
V20 Group LLC
P.O. Box 10066
Raleigh, NC 27605
Join the conversation
Log in to your V20 Group Marketplace account to leave a comment. We'll bring you straight back to this post.