With artificial intelligence (AI) becoming one of the most talked-about topics in automotive retail, dealers are being bombarded with promises of next-gen tools that claim to revolutionize operations. But not all AI solutions are created equal. On this episode of Inside Automotive, CDK’s Product Marketing Lead Amber Good explains what truly defines a generative AI-powered solution, where it’s driving real value for dealers today, and how to vet providers before signing on the dotted line.
Good begins today’s conversation by noting that AI in dealerships is nothing new – tools like mining have used machine learning for years – but what’s changing is how AI interacts with customers. The shift from decision-tree automation to generative AI, she explains, is a game changer because it enables natural, human-like conversations that can recognize emotion, nuance, and intent.
“Generative AI is super powerful... understanding not only the emotion but the nuance and truly just the human language that we would have in those conversations.”
One of the clearest applications of this new AI capability is lead follow-up. According to Good, salespeople have always struggled with timely and consistent responses to incoming leads. Generative AI can now handle the bulk of repetitive follow-ups, allowing staff to focus on high-value tasks like closing sales and creating customer videos. It’s especially impactful for internet leads, where customers are often left frustrated by generic replies or delayed responses.
Further, Good emphasized that poor lead handling not only wastes expensive ad spend but also damages the customer experience. “AI is going to do what AI is trained to do, which is answer the questions and have an intelligent conversation,” she said.
To avoid being misled by AI buzzwords, Good offered a list of questions dealers should ask vendors before signing any contracts:
- What language model is being used? Was it built in-house or adapted from OpenAI, and are there safeguards in place to prevent hallucinations or off-brand messaging?
- Was it written for automotive? Can it understand intent specific to buying or servicing a car?
- How often is the model retrained? And what data is being used—your dealership’s or public data?
- What happens in the demo? If it’s only answering basic questions like “where are you located,” it might not be true generative AI.
She stresses that dealers should test for more complex, vehicle-specific inquiries, like model trims, to ensure the system’s true intelligence and value.
As the industry heads into NADA and a crowded marketplace of AI vendors, Good’s advice is clear: ask thoughtful, smart questions, test thoroughly, and prioritize tools that protect the brand while truly enhancing the customer journey.