Index
Sales: responding and moving the opportunity forward
Reception and telephony: ensuring no call goes unanswered
Vehicle purchasing: deciding what deserves to enter stock
Stock and pricing: acting before the car ages
Marketing and digital storefront
How to prioritise the 25 cases
Three pathways combining multiple cases
How to estimate value before the pilot
Most common prioritisation mistakes
Dealcar and AI agents for dealerships
Frequently asked questions
Sales: responding and moving the opportunity forward
Case 1. Answering availability questions with current stock. When a lead asks "do you have a diesel Golf under 80,000 km?" outside of business hours, the agent checks the inventory in real time and responds with the available matching vehicles. Not with a "we will call you tomorrow" that loses the lead.
Case 2. Identifying budget, fuel type, usage and purchase timeframe. Before the sales representative resumes the conversation, the agent has already gathered the four or five details that determine which car to show. The representative enters with context, not starting from scratch.
Case 3. Recommending alternative vehicles when the requested one does not fit. If the car the lead asked about is already reserved or has more kilometres than they are looking for, the agent proposes the three closest alternatives from stock with the arguments for why they fit. This reduces abandonment due to "they don't have what I'm looking for".
Case 4. Scheduling a visit or test drive. The agent proposes available times, collects the lead's confirmation and adds it to the assigned representative's calendar. Without anyone having to exchange additional messages to arrange the appointment.
Case 5. Summarising the conversation for the sales representative. When the representative resumes contact, they receive a structured summary: what the lead asked, which car interests them, what budget they have and what objections or doubts they raised. They do not have to read through the entire WhatsApp conversation to get up to speed.
Case 6. Proposing the next follow-up based on lead status. If the lead visited the car but did not close the deal, the agent suggests when and how to make the next contact based on behavioural patterns: "they have viewed the car twice on the web, it is worth calling in 48 hours with a finance proposal".
These six cases share one condition: the AI has to know which car and which person are being discussed. An agent that responds without connecting to the CRM and inventory can be fast and at the same time completely useless.
Read also why lead response speed is the factor that impacts conversion the most.
Reception and telephony: ensuring no call goes unanswered
Case 7. Answering out-of-hours calls. The agent receives the call, identifies the reason, gathers the caller's details and offers an appointment or basic information. The dealership does not lose the lead because it is closed.
Case 8. Classifying calls by department. Sales, vehicle purchasing, administration and after-sales have different processes. The agent identifies the reason for the call and directs it to the correct channel from the very first second, without the caller having to explain themselves twice.
Case 9. Recovering missed calls. If a call could not be answered, the agent calls back within minutes, notes the reason and logs it in the CRM so the team can follow up with context. No missed call goes unfollowed.
Case 10. Confirming appointments and detecting cancellations. The day before a visit, the agent sends an automatic confirmation. If the lead cancels or does not confirm, they notify the sales representative in time to reorganise the schedule.
Case 11. Transferring an urgent call with a prior summary. When a call requires immediate human attention, the agent transfers it to the available sales representative with a 10-second briefing: who is calling, about which car and how urgent it is.
Vehicle purchasing: deciding what deserves to enter stock
Case 12. Gathering details from an individual wishing to sell their car. The agent collects make, model, year, kilometres, condition and asking price via WhatsApp or the web form. The buying team receives the complete vehicle profile without making any initial phone call.
Case 13. Detecting listings that match the buying policy. The agent monitors portals and professional platforms looking for cars that match the stock profile the dealership is seeking: make, year, kilometres and price within the target range. Every morning there is a filtered list of opportunities ready to review.
See also the best car auction platforms for dealerships.
Case 14. Comparing the asking price with market benchmarks. When a buying request comes in, the agent cross-references the asking price with the market price of similar cars on portals. If the price is out of range, they flag it before the buying team invests time in a physical inspection.
Case 15. Estimating margin after all costs. From the proposed purchase price, the agent calculates the estimated margin, including transport, typical reconditioning for that type of vehicle and the expected market selling price. It is not a definitive valuation, but a first filter that discards unprofitable deals before the inspection.
Case 16. Prioritising models by demand and historical rotation. Using the dealership's sales data, the agent identifies which vehicle profiles have rotated fastest and with the best margin in recent months. This information guides which buying opportunities to prioritise when multiple are available at the same time.
Stock and pricing: acting before the car ages
Case 17. Alerting of vehicles with incomplete photos or equipment. When a car is published on portals without the minimum number of photos or with empty required fields, the agent detects it and notifies the manager before the listing sits active for days without converting.
Case 18. Detecting prices outside the competitive range. The agent periodically compares the price of each published vehicle with similar cars on portals. If a car is significantly above the competitive range without generating leads for more than 10 days, they flag it for review.
Case 19. Suggesting reviews by tranches of days in stock. At 30, 60 and 90 days, the agent generates a list of cars that have exceeded that threshold with a proposed action for each: price drop, platform changes, new photoshoot or liquidation decision.
Read also how to calculate and control stock days in a dealership.
Case 20. Identifying duplicate, reserved or incorrectly published stock. A car appearing twice on the same portal, a reserved vehicle still listed as available or an ad with an outdated price are errors that the agent automatically detects without anyone having to check portal by portal.
Marketing and digital storefront
Case 21. Writing specific descriptions for each vehicle. Instead of generic or identical descriptions across cars of the same model, the agent generates personalised texts based on the vehicle specification data: extras, history, kilometres and sales conditions. The result is a unique listing for each car, not a filled-in template.
Case 22. Removing backgrounds and unifying the visual style of photos. Photos taken on the dealership lot with varying backgrounds are automatically processed: background removed, replaced with a neutral or branded one, and exposure corrected. The entire catalogue has the same visual presentation without manual photo-by-photo editing.
Check the comparison of AI tools for car photos.
Case 23. Creating social media posts from the car's file. From the vehicle details, the agent generates the text and format suitable for Instagram, TikTok or Facebook: including the price, the most relevant extras and a call to action. Without the sales representative having to write anything extra.
Case 24. Generating campaigns for vehicles with the most days in stock. Cars that have been in stock for more than 45 days without generating leads enter an automatic activation campaign: posting on social media with a highlighted price, potential price reduction and mention in emails or WhatsApp messages to leads who previously searched for that profile.
Case 25. Creating themed pages by make, body style, city or need. To improve local SEO and organic visibility, the agent generates specific pages such as "used SUVs in Bristol" or "second-hand hybrid cars under £20,000". These pages capture traffic from specific searches that the general catalogue page does not intercept.
How to prioritise the 25 cases
It makes no sense to activate all 25 at once. Four questions help to order them by real impact before committing resources:
Question | If the answer is high |
|---|---|
How many times does it occur per month? | Gains priority by volume |
How much does the error cost when it fails? | Needs more human control |
Are the data available already? | Can be piloted sooner |
Is the result measurable in 30 days? | Is a good candidate for a pilot |
A small dealership with few leads but many cars to photograph will probably get more immediate value from the photo and description cases (21 and 22) than from sales ones. A dealership with 200 or more leads a month, where many arrive out of hours, should start with cases 1, 7 and 9. A business with excess stagnant stock should look first at cases 18 and 19.
The quadrant matrix helps to avoid dispersion:
Quadrant | What to do | Examples |
|---|---|---|
High impact / high ease | Pilot first | Out-of-hours response, photos, descriptions |
High impact / low ease | Prepare data first | Predictive pricing, automatic purchasing |
Low impact / high ease | Automate if it frees up actual time | Conversation summaries, appointment confirmation |
Low impact / low ease | Discard for now | Projects with no clear metric or urgent problem |
Read also how to use AI to respond to leads in a dealership.
Three pathways combining multiple cases
Cases are not isolated actions: in practice, they link together into complete flows. Measuring each case separately can hide whether the entire pathway is working.
The sales pathway combines cases 1, 2, 3, 4, 5 and 6 into a single flow: the lead arrives via a portal, the agent responds with available stock, gathers the buyer's details, recommends alternatives if the car is not a fit, schedules the visit and summarises the conversation for the representative. The indicator of success is not how many messages the agent answered, but how many visits took place and how many of those visits ended in a sale.
The acquisition pathway links cases 12, 13, 14, 15 and 16: an individual requests a valuation, the agent collects the car data, compares it with the market, estimates the net margin and prioritises it based on the historical rotation of that profile. The final approval always rests with the buying team, but preparation time is significantly reduced.
The stock pathway connects cases 17, 18, 19, 21 and 22: the car enters the system, the agent detects if photos or equipment are missing, generates the description, checks the price against the market and warns at 30, 60 and 90 days if it has not rotated. The objective is to reduce the time between when the car arrives and when it generates its first lead.
How to estimate value before the pilot
Before committing time and budget to a specific case, the calculation of expected value should include three variables. The monthly volume of the task to be automated (how many missed calls, how many out-of-hours leads, how many cars to photograph). The savable time per occurrence multiplied by the cost per minute of the team's work. And the expected impact on sales: if case 7 (answering out-of-hours calls) generates 5 additional appointments a month and the conversion rate from visit to sale is 30%, that is 1.5 incremental sales a month that previously did not happen.
This estimation does not seek to justify the project but rather to identify which hypotheses need to be tested in the pilot: how many of those out-of-hours calls generate valid leads? How many of those appointments actually take place? If the pilot data confirm the hypothesis, it makes sense to scale up. If they do not confirm it, the cost of having tried it is limited.
Most common prioritisation mistakes
Choosing the case that had the most spectacular demo without verifying if the problem it solves is frequent in your own business. Automating a task that occurs two or three times a month, where the time savings do not justify the implementation effort. Ignoring the human review and correction time that any AI system requires: it is not zero. Mixing sales, after-sales and purchasing cases under a single success metric, making it impossible to know what is working. And failing to close projects that do not meet the results threshold agreed upon in the pilot: leaving them "on standby" consumes attention without generating value.
More than 750 dealerships already use Dealcar for their daily operations
Several of the 25 cases described in this article are already part of the daily life of dealerships working with Dealcar: AI photo enhancement directly from the vehicle file, automatic description generation from specification details, detection of stagnant stock with alerts for days in inventory, and centralisation of leads from all channels into a single CRM so that no contact is left unfollowed.
You can see how they work at dealcar.io/crm-leads-automocion or request a demo at dealcar.io.
The platform is constantly evolving in the area of AI. If you want to see which features are active today and what is on the roadmap, request a demo at dealcar.io.
Frequently asked questions
Where should a dealership that has never used AI start?
With the one that solves the most frequent and most measurable problem in the business. For most independent dealerships, the cases with the highest immediate return are the out-of-hours response (case 7 or 9) and description generation and photo enhancement (cases 21 and 22). Low risk, high volume of occurrences and a measurable result in under 30 days.
What data does the dealership need to have for the AI agents to work well?
The inventory must be updated in real time in the management system. The CRM must have the history of each lead with its status. And WhatsApp conversations must be centralised in the system, not on the personal mobile of each sales representative. Without these three inputs, the agent can respond quickly but cannot respond well.
Do AI agents replace sales representatives?
No. Agents cover out-of-hours availability, accelerate qualification and automate repetitive tasks (photos, descriptions, scheduled follow-ups). Business decisions, closing complex deals and managing unforeseen situations still require the human team. What changes is how much time that team spends on tasks that the AI now handles.
