An Artificial Intelligence tsunami is on the way. The people that understand how to use the power behind it are going to have a new lease on their careers.
The depth of value offered by artificial intelligence to a dealership’s operations is not yet fully known. What is understood can be compared to the tip of an iceberg, with AI’s full potential still hidden below the surface.
To many providers, including those of conversational and agentic AI services, the tip of the iceberg has already proven valuable. For example, Matador.ai, a communication platform at heart, has replaced its chat product (essentially just a form) with an identical one that uses AI to chat — and consumers are biting.
Conversational AI
Conversational AI attempts to understand and generate human-like text in response to prompts from users. Examples include chatbots, virtual assistants, information retrieval, and customer service, among other things.
In an interview with Canadian auto dealer, Kevin Esmezyan, Head of Product at Matador.ai, said the company uses AI to generate communications and handle conversations on behalf of a dealership.
“We’ve implemented AI large language models (known as LLMs) and machine learning in various places throughout the customer journey to improve the customer experience for the end user, while also generating more value for the dealership,” said Esmezyan.
Regarding its AI chat, he said the initial presumption when replacing it with the new product was that they would see a decrease in overall conversions. They thought consumers would balk at actually communicating with someone (or something), but the outcome was “quite surprising.”
Esmezyan said the opportunities generated from interactions and conversations with the AI chat were almost double fold across the board.
“People were actually more inclined to have a conversation. And what mattered the most was the quality of that conversation and how much value the AI could deliver within it,” said Esmezyan. “And so, what we’ve been laser-focused on for a while is how to build an AI that’s not just a chatbot, but is more human-like and can deliver that experience that consumers are now expecting 24/7.”
To arrive at this conclusion, Esmezyan said they looked at the context and then identified various key performance indicators that were then used to evaluate the situation.
He also listened to audio recordings of consumers talking to the AI. One included an elderly lady that spoke to it for several minutes and then said, “Oh, you’re so nice.” Another was a man who started off apprehensive and aggressive towards the AI, but when it said “Have a nice day,” the man responded with, “Thank you. You too.”
Esmezyan said the opportunities generated from interactions and conversations with the AI chat were almost double fold across the board.
Agentic AI
Agentic AI extends conversational AI by adding the ability to autonomously make decisions. It is used to manage complex data and processes, troubleshoot problems, automate workflows, and more.
If conversational AI is seen as simple chatbots, agentic AI could be considered intelligent AI copilots. It can help dealerships become more efficient, profitable, and insightful in many areas.
“It’s delivering value (in that it’s) allowing individuals in every department to truly optimize what they’re doing within their own individual jobs,” said Barry Hillier, Co-Founder & CEO of Auto Agentic. “Where a lot of people are worried that AI is going to replace them, I and many others have a very different view.”
One dealer Hillier spoke to echoed the fear of job replacement. So he asked her about her workload. The dealer said she had month-end reports to do for 10 dealerships and that it would take her a week and a half to complete. “I said, imagine if you could do those in one to three days. What are you going to do with the other week?” (The answer: she had a lot of other things to get done.)
Hillier said AI helps remove the things that are interfering with someone being able to perform their job, “because you’re doing the tasks that are taking up time — but they’re not creating value.”
He also noted that if 10 people are now capable of doing more work within their timeframe because they’re using AI, the benefits not only fall on them but also the dealership and the overall dealer group. “Ideally what ends up happening is these groups continue to expand and grow, (and) those same 10 people can manage that workload.”
Auto Agentic has created 10 AI agents that are deeply trained in the auto sector and built for specific tasks. (Think of AI that can act like a digital assistant for roles like sales manager or data or finance analyst.) Their agents are trained between 60,000 and 100,000 words (more than ChatGPT’s non-fixed 500-1500 output) on top of all the training the LLMs have provided.
“This is the other aspect: we’re LLM agnostic,” said Hillier. “So we use OpenAI, we use Claude, we use Gemini.”
The company plans to expand to 30 individual agents within the next few months, and aims to eventually have full agentic AI.
“So instead of you going and using three or four of our agents to do your month-end report, we will be at a point where they will know and have been instructed by you to pull all the info, generate the report, do the analysis, and give it to you,” said Hillier. “That is going to be coming up within the next seven months.”
“As long as we tell it to keep it within the voice of Mississauga Toyota, I just think that’s massive. It’s not recreating something, it’s just, it’s enhancing who you are as a business and how you represent yourself,” said Gubasta.
Susan Gubasta, Dealer Principal at Mississauga Toyota, said she’s been piloting Auto Agentic in her dealership, and she and her team are always finding more ways to make use of it.
“We have so much data that within our own businesses for each department, and it’s now a whole new world because we can take that data and we can enhance our business in days by coming up with a game plan, figuring out what we’re looking to do and then implementing this stuff quickly,” said Gubasta. “So it’s not the days of old, it’s a new world of the ‘here and now’ and the instant.”
Gubasta said that after some initial training, people were able to easily adapt to using the tool. “It was simple. There are other things historically that we’ve done and they weren’t simple. But this, it’s easy. And shame on you if you don’t use it,” she said.
“I think that dealers who will get on the AI bandwagon, we’ll gain so much in short order. It’s about being a little bit better, getting things done a little bit faster, and at the end of the day, it’s personalized to their store or to their dealer group, whatever that looks like. And it’s just that next level tool that will make your lives a lot simpler,” she said.
“As long as we tell it to keep it within the voice of Mississauga Toyota, I just think that’s massive. It’s not recreating something, it’s just, it’s enhancing who you are as a business and how you represent yourself,” said Gubasta.
She said people are concerned AI will be used to replace people, but she said dealerships are in the people business, and this just enhances their skills and abilities.
Smart Application Manager
Another provider is DecisioningIT, which uses a smart application manager widget called SAM to pre-qualify customers for auto financing. The AI-powered tool uses a predictive matching system to connect consumers with potential lenders, and doesn’t discriminate based on credit history.
“SAM allows dealers to pre-qualify their customers in real time, clearly identifying their purchasing power without affecting their credit score,” said André-Martin Hobbs, Chief Product Officer.
Dealers can integrate SAM directly into their website or in-dealership via QR codes. Once this is done, Hobbs said, they can immediately identify qualified prospects, optimize their financing offers (such as rebates), and maximize F&I product sales. “The results are compelling: increased sales and profitability, plus a higher CSI index.”
Hobbs is convinced AI will continue to profoundly transform the automotive sector, suggesting its impact will be felt in the short term via the accumulative and proactive management of customer data, more efficient operations, and an improved user experience.
The company will be launching a version of the widget called SAM Leisure in mid-April, along with a new tool specialized in vehicle trade-in valuation.
“Automated monthly reports generated by AI will be available to easily identify inefficiencies and help optimize the financing (F&I) process,” said Hobbs.
Overall, he said AI and tools like SAM automate repetitive processes and allow sales and finance teams to focus on high-value tasks, like personalized customer interactions and revenue optimization.
Hobbs is convinced AI will continue to profoundly transform the automotive sector, suggesting its impact will be felt in the short term via the accumulative and proactive management of customer data, more efficient operations, and an improved user experience.
“Ultimately, different AI tools will interact with each other to increase their effectiveness tenfold,” said Hobbs. “In the longer term, we will have a fully connected automotive ecosystem, where vehicles communicate directly with dealerships and anticipate maintenance, upgrade, or even replacement needs, creating a fully predictive and frictionless customer experience.”
AI Glossary Primer
- Artificial Intelligence (AI): Software that mimics human intelligence — like learning, problem-solving, or making decisions.
- Machine Learning (ML): A type of AI that learns from data and improves over time without being manually programmed.
- Neural Network: A computer system modeled after the human brain — it’s how AI can recognize patterns, like matching a customer with the right vehicle.
- Model: The brain of the AI — it’s trained on data and used to make predictions or decisions.
- Training Data: The examples or information the AI learns from. More (and better) data = smarter AI.
- Algorithm: A set of rules or steps the AI follows to make decisions or solve problems.
- Natural Language Processing (NLP): How AI understands and responds to human language — like when a chatbot answers customer questions.
- Prompt: A message or instruction given to the AI — like a question, command, or request.
- Agent: A smart AI “worker” that can complete tasks or automate processes — like Auto Agentic’s AI agents that follow up with leads or analyze sales data.
- Automation: Using AI to handle repetitive tasks — freeing up human time for more important work.
- Accuracy: How often the AI gets things right. Important when evaluating AI tools.
- Hallucination (in AI): When AI generates information that sounds correct but isn’t true or accurate. It’s not lying — it’s making its best guess based on patterns, but sometimes gets it wrong.
- Bias: When AI makes unfair or unbalanced decisions — often due to flawed or limited training data. Something ethical AI providers actively guard against.
- Explainability: How easy it is to understand why the AI made a certain decision — critical for trust and transparency.
- Real-Time Processing: AI that works instantly — like chatbots or real-time pricing tools.
- Predictive Analytics: AI that forecasts future outcomes — like which leads are likely to convert into buyers.
- Generative AI: AI that creates things — like writing emails, generating images, or composing responses.
- Agentic AI: AI that acts like an intelligent assistant — not just answering questions, but taking initiative, making decisions, and completing tasks on your behalf.
- Large Language Model (LLM): A powerful AI trained on vast amounts of text — used to power tools like ChatGPT.
- API (Application Programming Interface): A way for different software systems to talk to each other — often how AI gets integrated into dealership software.
- Use Case: A practical example of how AI can be applied — like lead follow-up, inventory analysis, or automated service reminders.
Source: Barry Hillier, Auto Agentic
