Index
The agent's mission in one sentence
The funnel that can operate autonomously
Daily use cases
The business rules that the agent must respect
How to make a recommendation that does not overwhelm
The handoff: when and how to hand over the conversation to the salesperson
What information the agent must return to the CRM
How to measure the real contribution of the agent
The agent does not replace sales discipline
How to train the team to work with the agent
More than 750 dealerships already use Dealcar for their daily operations
Frequently asked questions
The agent's mission in one sentence
To transform an enquiry into a well-understood opportunity, with the vehicle identified and with a clear next step, without the salesperson having to start from scratch.
This means that the agent does not replace the human close or the advice during the visit. What it does is prevent the team from spending time on repetitive tasks: answering the same availability question twenty times a day, asking for contact details that the buyer already provided on the portal form, or confirming appointments one by one over the phone.
The funnel that can operate autonomously
The agent can accompany the buyer through five stages of the sales process, playing a different role in each.
At the entry stage, the agent identifies the source channel (portal, WhatsApp, web, advertisement), checks if the buyer already exists in the CRM with a previous history, and classifies the reason for contact. The result is a single lead linked to a file, not duplicated in three different inboxes.
See also how to design 24/7 customer service across all channels of a dealership.
At the discovery stage, the agent gathers the four key pieces of data that allow the stock to be filtered relevantly: approximate budget, intended use of the car, preferred body style, and fuel or transmission preference. It does not ask all four questions at once: it asks one, and uses the answer to decide if the next one is necessary.
At the recommendation stage, it filters the stock in real time and presents two or three options with a specific reason for each. Not a list of results: a reasoned selection.
At the conversion stage, it resolves the doubts it is allowed to resolve according to the dealership's rules (availability, published price, features) and proposes an appointment in a real slot in the diary.
At the follow-up stage, it reminds of the appointment the day before, detects cancellations, and reactivates leads that did not respond using the context of the previous conversation.
Daily use cases
The five cases where the agent adds the most value to the daily operations of a dealership are the following.
Responding to a portal lead in less than two minutes. The buyer who sends a message at 22:30 about a specific car receives an availability response and an appointment proposal before another dealership opens the following day.
Attending to WhatsApp and the web outside of business hours. Without an agent, those leads wait until the next day. With an agent, they are qualified and the visit is scheduled that very night.
Proposing an alternative when the vehicle enquired about is no longer available. Instead of replying "that car has already been sold", the agent proposes two alternatives from the current stock that match the buyer's criteria.
Preparing the context of the part-exchange car when the buyer wants to trade in. The agent collects the registration number, mileage, and general condition of the vehicle to be traded in before the buyer arrives at the dealership.
Summarising objections for the salesperson. When the buyer mentions that it seems expensive, that they need more time to decide, or that they are comparing with another dealership, the agent records that objection in the file so that the salesperson arrives at the visit knowing what they need to work on.
The business rules that the agent must respect
The agent operates within a set of rules defined by the dealership, which must be documented as configurable data, not as instructions hidden in a system that nobody can audit.
The minimum rules to define before activating the agent are: pricing and discounts (the agent can only inform of the published price, never commit to a discount without approval), vehicles excluded from automation (cars in active negotiation, reserved, or with special conditions), reservation policy (whether the agent can block a car in the system or only propose the appointment), locations and calendar (which slots it can offer and at which sites), finance (it can explain that finance is available but not calculate personalised quotes), and lead assignment order (by site, by salesperson availability, by rotation, or by specialty).
See also what AI agents for dealerships are and what they should never do alone.
When the agent faces a situation not covered by the rules, the correct response is to escalate to the salesperson, not to improvise.
How to make a recommendation that does not overwhelm
Showing fifteen cars that match the buyer's criteria is not a recommendation: it is a catalogue. The buyer has to do the work of comparing on their own, which is exactly what they were already doing on the portal before making contact.
The agent presents two or three options with a specific reason for each. A real example: a buyer asks for a diesel because they do high mileage. The agent does not assume diesel is the best option: it asks how many miles they cover per year and where they drive them. If the answer is 8,000 miles a year in town, the agent explains that there are petrol and hybrid alternatives within their budget that might be more economical for that usage profile, and presents three units highlighting the differences between them.
That conversation does not close the sale: it opens a more informed conversation that the salesperson can continue during the visit with a buyer who already has better criteria. The agent does not dictate what to buy: it explores, informs, and proposes.
Recommendations must be honest: not mentioning equipment that is not on the spec sheet, not showing cars that are not actually available, and explaining the differences between alternatives so that the buyer can make a decision based on real information.
Read also how to show cars without overwhelming the buyer from the dealership's website.
The handoff: when and how to hand over the conversation to the salesperson
The handoff is the moment the agent transfers the conversation to the salesperson. Doing it at the right time and with the right context is what determines whether the agent is an asset or an obstacle in the sales process.
The agent must transfer when it detects high buying intent (the buyer says they want to come in this week or have already decided), when there is negotiation on price or terms, when the situation has complexity that the rules do not cover (personalised finance, complaint, trade-in with paperwork issues), or when the buyer explicitly asks to speak to a person.
The summary the agent hands over to the salesperson during the handoff must contain six elements: who the buyer is (with their history if they exist in the CRM), what car they are interested in and what alternatives they saw, what their actual need and buying timeframe are, what main objection or question is outstanding, what actions the agent has already taken (appointment proposed, alternative sent, appraisal initiated), and what the recommended next step is.
This summary allows the salesperson to enter the conversation without having to repeat questions the buyer has already answered. If the salesperson has to ask "which car were you interested in?" again, the handoff has failed.
Read also how to organise a dealership's sales team.
What information the agent must return to the CRM
The value of the sales agent is measured by what it leaves in the CRM, not by the fluency of the conversations. If the agent maintains excellent conversations but records nothing useful in the system, the salesperson starts from scratch with every contact.
The structured data that must remain in the CRM after each interaction by the agent includes: the source channel and the campaign if there is one, the detected level of intent (browsing, actively considering, ready to buy), the primary vehicle of interest and the alternatives the buyer viewed, the questions left unanswered, the assigned owner and the next task with a date, and the reason for closing if the conversation ended without progressing.
This last piece of data, the closing reason, is the most frequently omitted and the most valuable. If 40% of leads are closed as "did not reply" without further detail, there is no way of knowing whether the problem is the price, the stock, the contact channel, or the timing of the follow-up.
How to measure the real contribution of the agent
Activity metrics (messages sent, conversations started, response time) measure that the system is technically working. Business metrics measure whether it is generating value.
The seven metrics that must be visible to evaluate the agent's contribution are: first response time (the target is less than 2 minutes for inbound leads), qualified leads over total leads received (of those who contact, how many provide enough data for a recommendation), appointments generated by the agent, show-up rate for those appointments, proposed alternatives that the buyer agreed to explore, sales where the agent was the first point of contact, and assisted sales (the agent participated at some point in the process even if it was not the first).
See also the KPIs a dealership must measure to control its business.
These metrics must be segmented by channel and time slot. An agent that generates many appointments at night with a low show-up rate might be offering appointments without enough intent or without properly preparing the buyer before the visit. An agent that has good metrics from Monday to Friday but none at the weekend may have a schedule configuration issue.
If the calendar grows but attendance falls, the agent is closing appointments without enough intent or without generating enough commitment from the buyer before confirming.
The agent does not replace sales discipline
If CRM stages are not updated, if appointments do not have a clear owner, and if reasons for loss are generic ("not interested" without further detail), the agent inherits this disorder and amplifies it.
Before activating the sales agent, it is advisable to simplify the CRM stages and agree on what each one means. The system can remind of tasks and propose closes, but management must review the pipeline and the team must update stages with real information from each conversation.
Read also how to structure a dealership's sales funnel.
Technology improves the consistency of a process that works. It does not correct a weak sales strategy or an uncompetitive product offering on its own. If the issue is stock pricing or inventory profile, the agent is going to be very efficient at generating leads that the team cannot close.
How to train the team to work with the agent
Sales team resistance to the sales agent usually stems from three causes: they do not know exactly when the agent acts and when they need to act, they feel the agent takes leads away from them without recognition in attribution, or the agent generates extra work (correcting data, reviewing summaries) without eliminating previous tasks.
All three causes have solutions, but none are resolved with training alone. The team needs to know four things: at what moments the agent is handling a conversation, how to take control of a conversation when they deem it necessary, how to correct incorrect data or recommendations from the agent, and how to record an exception when the situation does not fit the configured rules.
Adoption improves naturally when the system saves visible work (the salesperson arrives on Monday morning with the weekend's appointments already booked) and when the distribution of opportunities is transparent (each lead has a clear owner and sales attribution is fair).
More than 750 dealerships already use Dealcar for their daily operations
The Dealcar sales agent connects WhatsApp, the web, portals, and ads with stock and the CRM. Leads from all channels arrive in the same inbox with the context of the conversation linked to the customer's file and the vehicle of interest. This architecture makes it possible to evaluate the entire journey from the first enquiry to the visit and, when the agent is available, to automate the repetitive steps of that journey.
You can see how it works at dealcar.io/crm-leads-automocion or request a demo at dealcar.io.
Frequently asked questions
Does the agent compete with dealership salespeople?
Not if the division of tasks is well defined. The agent covers availability, speed, and consistency out of hours; the salesperson handles negotiation, building trust, and closing. The problem arises when sales attribution does not reflect the contribution of both: if the salesperson feels the agent is "stealing" leads without recognition, there is a process design issue in attribution, not with the agent itself.
To whom does the agent assign each lead?
According to the configured assignment rules: by site if the buyer mentions a geographic area, by salesperson availability in real time, by fair rotation, or by specialty (one salesperson managing trade-ins, another managing finance). It must be possible to make exceptions and reassignments easily from the CRM.
How does the agent learn from the dealership's conversations?
It does not automatically learn from the entire CRM history. It learns from rules, stock, and conversation examples that the system administrator explicitly decides to incorporate. You must decide what data and conversations the agent can use, rather than assuming it has access to everything.
