KPI TECHAI agents for car dealerships: what each tool does
The tools of the KPI Tech platform are AI agents for car dealerships, in sales and in service: they answer the phone, score calls, check leads and point out customers who have a reason to come back. All of them work on one customer knowledge graph, and decisions on price and terms stay with a person. One of the tools is Sufler from KPI Tech, the consultant agent for salespeople, service advisors and the BDC: during a conversation with a customer it prompts what to say and what not to promise.
The agents take over retyping data, keeping track of deadlines and ringing round the customer base. A person gets a task with a due date from the agent.
The dealership phone and customer calls
Receptionist
Answers the phone during peak hours and after the dealership closes. It qualifies the enquiry, books service and hands the call to a person when a decision is needed.
Call Analyzer
Transcribes calls and scores customer intent, adherence to the script, lost opportunities and complaint signals. The result lands on the customer record, next to their cars and visits.
Leads, CRM deadlines and conversations about a new car
Lead Validator
Checks leads from web forms and from Meta before they reach a salesperson: whether the contact is real, how hot the lead is, which model it concerns and when the customer plans to buy.
CRM Controller
Watches the quality and timeliness of work on leads: whether there is a note, a status and a next step, and how funnel conversion looks. It works on read-only access, so it changes nothing in the CRM.
Sales Prospector
Looks for reasons to talk about a new car: readiness to change the car, the move from service to sales, the end of a finance agreement and part-exchange potential. The end-of-term playbook only runs where we have confirmed access to finance data.
Reasons for a service visit, and the quality of customer care
Service Prospector
Points out service leads: dormant customers, recommended work left unbooked, service intervals, warranty expiry and service campaigns. Every reason is its own playbook, so you can see why a given customer is on the list.
Quality Loop
Tracks the post-visit SLA, catches dissatisfaction and complaints and makes sure they are put right. It invites customers to leave a review whatever they thought of the visit.
Prompts during calls, messages to customers and a messenger for the group
Sufler
Knows the customer, the vehicle and the history, and during a conversation prompts what to say and what not to promise. It works in text and in voice, in a management role and in an employee role.
Personalizer
Runs email and SMS communication, either automatically or with human approval, and writes conversation scripts. It builds the content from what is known about the customer and the car, so every customer gets a different message.
Sufler Connect
A secure messenger for a dealer group, designed for NIS2 requirements, in which AI agents work alongside employees. It has company-account sign-in, roles, multi-factor authentication, channels and alerts from the other tools.
One customer record and tool calibration
Knowledge Graph
Joins the customer, their vehicles, work orders, visits, contacts and consents into one record. We pull data from the DMS, the showroom CRM, telephony, mailboxes, manufacturer feeds and marketing, and merge duplicates on phone, email and chassis number.
Learning Loop
Checks what happened to scores and recommendations after they reached the CRM, and calibrates the parameters of the other tools on that basis.
How the tools hand work to each other
- The Service Prospector hands its list of service leads to the Receptionist or to a person, who makes the calls.
- A few weeks of Call Analyzer recordings calibrate the Receptionist before it goes live.
- Sufler works on the same graph where the Call Analyzer stores its scores and the Receptionist stores recordings and notes from calls.
The usual rollout order in a dealer group
We often start with Sufler or with the Call Analyzer. The Service Prospector, the Sales Prospector and the CRM Controller come after the Call Analyzer. The Receptionist goes live after calibration, followed by the Lead Validator. Every month after the start you get a report: what the tool did, what came of it and where it needs correcting.
The full rollout order and stagesQuestions about AI agents for car dealerships
Where do the agents get their data on the customer and the car?
The knowledge graph joins the customer, their vehicles, work orders, visits, contacts and consents into one record. We pull data from the DMS, the showroom CRM, telephony, mailboxes, manufacturer feeds and marketing, and merge duplicates on phone, email and chassis number. Every other tool works on that layer.
What permissions does an AI agent get, and what data does it see?
Every agent has an identity, permissions and an action log, and the data of each group is isolated. An agent sees only what its role needs, and never more than its owner sees in the CRM and DMS. GDPR and the DPA are part of the contract. We send the architecture description on request: hi@kpitech.pl.
What contract covers each new tool a dealer group adds?
We work under a master agreement for the group and a separate SOW for each tool and agent. The contract covers response times, the data processing agreement, a billing unit we do not change mid-term, and the right to exit with help migrating the data.
The rollout starts with one tool
The demo is an hour of conversation about one dealership and one pipeline. Once the tool goes live, we look at the numbers together with your team.