Home · Solutions · Sales & marketing
Solution · Sales & marketingOne form in Teams returns an offer band, the reconditioning cost and who must sign, in minutes
A trade-in price in twenty minutes, reconditioning included
Robots gather the car's history, the group's own sales and market listings, apply the reconditioning table and return an offer band with its approval level while the customer is still in the showroom.
Executive summary
The customer wants a price for the old car today. The used-car manager is in a meeting, and the offer arrives tomorrow.
Mientha delivers the flow on the UiPath Platform inside the group's Microsoft 365 tenant.
The customer hears a figure in the showroom, and the new-car deal that depends on the part exchange closes on the same visit.
the group CRM; the DMS platforms for purchase and resale matching; the SharePoint evidence library
Business problem
Used vehicles
A large share of retail deals carries a part exchange, and the price offered for it decides two things at once: whether the new-car deal closes today and what margin the used-car department makes months later. It has to be fast, because the customer is comparing, and right, because the group owns the car until it sells.
In most groups the appraisal is a person. The site's used-car manager prices from experience, two listing portals and a wholesaler's phone number; reconditioning is a round figure from the last similar car; the threshold is whatever he will sign. None of it is written down, so eight sites price eight ways, and the consequences arrive late: a car bought too dear sits for a hundred days and is wholesaled at a loss, one bought too cautiously leaves with the new-car deal, and finance sees the result at the stock review with no record of who priced or approved the car.
How it works today
- PersonThe adviser photographs the car, odometer and registration document and sends them to the manager in a chat with the mileage and a line on condition
- WaitingThe manager is with a customer, in the stock review or at another site; the request waits behind the day's other seven
- PersonThe manager searches two listing portals by model, year and mileage, recalls the last one the site sold, guesses the reconditioning and calls a wholesaler for a floor
- Risk of errorOffer, approval and evidence exist only in the chat, and every site applies a different threshold
- WaitingThe customer hears the figure the next day, by phone, after another dealer has named theirs
- Risk of errorWhether the price was right is discovered on the ramp, and again when the car sells, with no link back to the appraisal
Why the current process costs more than it appears
Behind every exception is an hour nobody logged.
- Forty minutes of two people per appraisal is the visible cost; across 480 cars a month it is two experienced people's working month, spent on lookups rather than judgement.
- Reconditioning found on the ramp is paid twice: as margin that was never there, and as a car waiting in the compound while the workshop fits it in between paying customers.
- Delay loses deals that appear in no report: the customer who leaves without a figure and does not come back is counted nowhere, and the new-car order goes with them.
- Informal thresholds mean nobody can say which of last quarter's offers exceeded a manager's authority, and when the one manager who can price is on leave, the site guesses or stops taking part exchanges.
Cost of inaction
Losing appraisals surface six months late, when the car finally sells below what was paid plus what the workshop booked. The rows price the gathering only; lost deals and margin given away without a written band cannot be modelled honestly without your data. What compounds is the dependence on eight individuals: every site prices as well as its manager does that day, and the group's used-car margin is the sum of habits it cannot see.
A plausible organisation with realistic proportions. The figures are there to be recalculated on your data; they are not a client result.
An illustrative Polish dealer group: three volume brands and one premium brand on eight sites, about 760 employees, a used-car centre at the largest site, one DMS per brand, one group CRM, Microsoft 365 E3 with Teams on advisers' phones.
About 480 part-exchange appraisals a month, roughly 60 a site, of which the group buys around 220; most cars are three to seven years old, a third on a full-VAT invoice.
Photographs and a mileage in a Teams chat to the manager, who prices from two portals, memory and a wholesaler's quote; reconditioning a round figure, approval whatever he signs, the answer next day.
About 40 minutes of adviser and manager time per appraisal; an overnight wait; a reconditioning guess tested on the ramp; no record linking offer, approval and resale.
An appraisal form in Teams; robots read the registration document, gather history, own sales and listings, apply the reconditioning table and return a band with its approval level; exceptions reach the manager with the evidence; appraisal, offer and resale sit on one record.
In the modelled case the adviser has a band and the name of who must sign it within about twenty minutes, 320 hours a month of gathering are released, and the tables are corrected each quarter from resale results; none of it was measured at a client.
Proposed solution
Mientha delivers the flow on the UiPath Platform inside the group's Microsoft 365 tenant. Next to the car, the adviser fills in a Microsoft Lists form in the used-car channel of Microsoft Teams: VIN, mileage, equipment and condition items ticked from checklists, how the car is owned, and photographs including the registration document. A new item starts the robot through the list-item trigger of the Microsoft OneDrive & SharePoint connector in UiPath Integration Service.
The robot gathers. A UiPath Document Understanding model trained on the group's own registration documents reads VIN, plate and first registration and compares them with what was typed; a mismatch goes back to the adviser before pricing. It then fetches specification and service record from the importer portal, mileage history and comparable listings from the vehicle-history and market-price services the group already pays for, and own sales of the model from the DMS platforms.
Pricing is a table, not a model. The used-car director owns the rules: expected retail price from own sales and listings, target margin per brand and age band, a cost per stock day, a reconditioning table priced per condition item by the aftersales director at the workshop's own rates, and the threshold table of who may sign what; margin-scheme and full-VAT resales are taxed differently, hence the ownership field. The result is a band (floor, target, ceiling) with the reconditioning sum by item, the comparables and the approval level. An offer outside the band, fewer than three comparables or a history flag raises a UiPath Action Center task with the evidence for the site manager, or above a second threshold for the used-car director. Every appraisal is written to the group CRM and later matched by VIN to the purchase and resale in the DMS, so Power BI shows the chain per car.
UiPath Orchestrator queues, triggers and audit; UiPath Integration Service connectors for Microsoft OneDrive & SharePoint and Microsoft Teams, and Connector Builder; UiPath Document Understanding (custom document model, pre-trained Invoices model, Validation Station); UiPath Action Center actionable notifications in Microsoft Teams; Microsoft Lists; Power BI
Form and checklists, document model, gathering robots, the three tables, band card, approval tasks, CRM record, VIN matching, report and runbook
The DMS platforms, the group CRM and the importer portals through their interfaces or screens; the vehicle-history and market-price services through an interface where the subscription includes one, otherwise the screen
How the automated process works
- PersonNext to the car, the adviser fills in the form in Teams: VIN, mileage, equipment, condition items, how the car is owned, and photographs
- AutomationThe new item starts the robot; Document Understanding reads VIN, plate and first registration from the photo, and the adviser corrects any mismatch
- SystemThe robot collects specification and service record from the importer portal, mileage history and comparable listings from the two subscriptions, and own sales from the DMS platforms
- AutomationThe pricing table produces the retail estimate, deductions, reconditioning sum, band and approval level; a history flag, a mileage inconsistency or too few comparables raises it
- PersonThe adviser receives the band, the reconditioning breakdown and the approval level in the channel, and offers inside it or asks for an exception
- PersonAn offer outside the band reaches the site manager, or above the second threshold the used-car director, as an Action Center task in Teams with the evidence: approve, adjust or decline with a reason
- AutomationThe written offer goes to the customer on the channel they asked for; a reminder before it expires goes out only where that channel's consent record allows, and an objection stops it for every brand
- AutomationThe appraisal is written to the CRM; when the DMS records purchase and resale, the robot matches them by VIN and the report shows the chain on one line
Human-in-the-loop model
Automation handles
- Reading the registration document and service invoices, and checking the typed VIN and mileage against them
- Gathering history, specification, own sales and comparable listings, the same way for every car
- Applying the tables, routing exceptions with the evidence, and matching each appraisal to purchase and resale
People decide
- The offer itself: the adviser chooses the figure inside the band and negotiates with the customer
- Every exception: managers approve, adjust or decline on the evidence
- The condition of the car, assessed by the adviser and later by the technician; the robot prices only what they tick
Before and after
Systems and integrations
Every entry can be checked in vendor documentation. The evidence class is stated next to each one.
Inputs
- the appraisal form in Microsoft Lists with photographs
- the importer portal per brand
- the vehicle-history service
- the market-price service
- used-car sales in the DMS platforms
Automation layer
- UiPath Orchestrator
- UiPath Robots
- UiPath Integration Service
- UiPath Document Understanding
- UiPath Action Center
Target systems
- the group CRM
- the DMS platforms for purchase and resale matching
- the SharePoint evidence library
- Power BI
Human touchpoints: the band card in the used-car channel in Microsoft Teams; Action Center tasks in Teams for exceptions; Validation Station for unreadable documents
Technologies used
gathering and pricing on the list-item trigger; queue, retry, log and audit per appraisal
Astarts on the new form item, posts the band, wraps the services that offer an interface
Aa model trained on your registration documents reads VIN, plate and first registration; the pre-trained Invoices model reads service invoices; Validation Station for doubtful fields
Aexception tasks for managers, with evidence, due dates and audit
Athe form on the adviser's phone, the channel where the band arrives, the approval tasks
AIllustrative economic model
The arithmetic is open, so it can be argued with.
A trade-in costs the group forty minutes before anyone names a price, and that is all the model counts: the adviser's photographs and message, the manager's portal searches, the wholesaler call and the reply. €25 an hour blends the fully loaded cost of a sales adviser and a used-car manager. Nothing here was measured at a client; deals won and losses avoided are outside the arithmetic.
Run the numbers on your data
An illustrative estimate from your own inputs. It models released capacity; it is not a promise of savings.
Business benefits
- The customer hears a figure in the showroom, and the new-car deal that depends on the part exchange closes on the same visit
- Reconditioning is priced per ticked item before the purchase, not discovered on the ramp after it
- Every offer carries its evidence and approval, and a small site without an experienced manager makes the offer the used-car centre would make
- 320 hours a month of gathering move to robots, and losing appraisals become visible per site and adviser within the quarter
The management view
- Pricing discipline becomes a table with an owner and a version, applied identically at eight sites instead of eight managers' habits
- An offer above a site's authority needs an approved, time-stamped task; the flow enforces the threshold, not habit
- Appraisal, purchase and resale are linked for every car, so the tables are corrected from evidence each quarter and growth no longer needs an experienced manager at every site
Board-level KPIs
Security and governance
The automation holds exactly the rights it needs, and not one more.
- The registration document carries the owner's personal data: the robot extracts vehicle fields only, the image stays in a restricted SharePoint library under a retention label, and the band card never shows the owner
- Robots run in the EU region of UiPath Automation Cloud with a dedicated login per portal, subscription and DMS, read rights plus the one write the process needs, and secrets in an Orchestrator credential store backed by Azure Key Vault
- Segregation of duties is the threshold table itself: an adviser cannot approve their own exception, a manager cannot approve above the site limit, and every approval carries a name and a time; the tables change only with the used-car director's sign-off, and Document Understanding runs under AI Trust Layer policies
Why now
Used cars are where a group's margin is made and the part exchange is where it is decided; every quarter of chat-and-phone pricing books more cars nobody can defend, beyond the modelled €8,000 a month of gathering
More cars are coming back: 597.4 thousand new passenger cars were registered in Poland in 2025, 8.3% more than in 2024, 350,500 with an alternative drivetrain (PZPM and KPMG, February 2026); those are the part exchanges of the next three years, and the electrified ones will test any pricing table
Every component is documented product: a Document Understanding model trained on your own documents, a list-item trigger from Microsoft Lists, Action Center tasks completed inside Teams
Relevant executive roles
Pricing becomes a table he owns and corrects from resale results, and exceptions reach him with the evidence before the purchase
Reconditioning is priced before the purchase from the workshop's own rates and compared with what was booked
The used-car margin is traceable per car from appraisal to resale, and losing purchases show within the quarter
Common questions and objections
They keep doing so: the table produces a band and the manager decides everything outside it, on evidence already gathered. What changes is that the band exists at every site and learns from what the cars made.
It gives a market figure for a model, and knows nothing of your reconditioning rates, stock-day cost, own sales or approval limits. The robot reads it as one input and adds the rest.
The form asks for what they already photograph and type into a chat, and returns a band in minutes instead of a figure tomorrow; a site that skips it shows in the report as the one with the most exceptions.
When this is not the right solution
- Single-site dealers or groups appraising a few dozen cars a month, where one experienced manager with a written threshold and a shared list suffices
- Groups whose DMS keeps no resale result per car; a table cannot be corrected from results nobody kept, so that record comes first
A question for the next management meeting
Last quarter we bought cars from our own customers: how many of them sold for less than we paid plus reconditioning, and which of those offers was approved by nobody?
Implementation approach
Delivery runs in stages, so it can be stopped at any point.
We deliver
- One quarter of your part exchanges matched by VIN to purchase, reconditioning booked and resale, to show where a band would have sat
- The form and checklists in Microsoft Lists, agreed with advisers at two sites and the aftersales director
- The document model trained on your own scans and the gathering robots, each with its own login and scope
- The three tables as versioned, owned objects, the band card and approval tasks in Teams, CRM record, VIN matching, Power BI report and runbook
We need from you
- Twelve months of part exchanges from the DMS platforms: appraisal, purchase price, reconditioning booked, resale price and days in stock
- The used-car director to own the pricing tables and the aftersales director the reconditioning prices
- Logins for the portals, subscriptions, DMS platforms and CRM, with confirmation that automated access is permitted under each contract
Stages
Discovery
One quarter of appraisals matched to resales; sources, subscriptions, each site's thresholds and habits
Design
Form and checklists, tables, thresholds per site and brand, the band card, the security model
Build and validation
Document model, robots, rules, Teams touchpoints, CRM record and report, then three months of appraisals replayed against what the cars made, checked by the used-car director
Go-live and tuning
Two sites first, every band reviewed by the manager for a month, then site by site; quarterly correction of the tables
Departmental. Effort follows the number of DMS platforms and importer portals, whether the subscriptions offer an interface, and how far the sites' pricing habits differ.
The reconditioning cost is guessed at the appraisal and discovered on the ramp.
Send us thirty part exchanges from last year with the offer, the reconditioning actually booked and the resale price. You get back a written read-out of where a band would have sat on each and which offers needed a signature.
Compare thirty appraisals with their resaleThe neighbouring process usually has the same problem
Nine channels, four brands, one receptionist deciding who calls the customer back. The Friday lead waits until Monday.
View solution Operations & qualityThe right car at the right site, and demonstrators in viewStop moving cars between sites on a phone call and finding out where the demonstrators are by asking.
View solution Sales & marketingThe next car offered while the current one is on the rampYour owner base drives through your workshops every month while sales buys leads from portals.
View solutionIndustries we deliver this in most oftenAutomotive retail