Nestack Agent Care
Industries / Automotive / Floorplan agent

Automotive AI agent · Floorplan

Floorplan & Inventory-Audit AI Agent

Reconcile the lender's outstanding VIN schedule against the inventory of record, age each unit to its own curtailment date, and evidence each variance before your controller signs the certification the guaranty stands behind.

4–6 weeksTypical delivery
Your stackDeployment
Ledger-firstNamed signature
Agent CareAfter launch

What this agent does

Reconciles the position, never certifies it

In
01

A unit is floored, and its VIN, floor date and advance are read from the lender's outstanding schedule.

02

A VIN is matched to the inventory of record and resolved to one documented status, or held as a variance.

Reason
03

A status is evidenced per unit — on the lot, in service, demo, loaner, in transit, on a trade, or sold.

04

A curtailment falls due on that facility's own schedule, since timing and percentage are per-loan terms.

05

A unit is delivered, and the remittance clock runs from delivery, not from the date the contract funds.

Decide
06

A retail buyer takes free under §9-320(a) as your state enacted it, so the car goes and the money is left.

07

A payoff is tracked as identifiable proceeds, which §9-315(d) leaves unperfected after 21 days.

Out
08

A variance is written up as it arises, dated then, and kept per VIN against the inspection that follows.

09

A write happens only inside the approval boundaries agreed at implementation, and never as a signature.

Product statement

The agent reconciles the position; a named controller certifies it, and the dealer principal's personal guaranty stands behind that signature.

Example workflow

One unit, ledger to certification

AgentHuman
1Unit on the lender's ledgerOutstanding VIN schedule, DMS inventory, deal jacket or wholesale record
2Status evidencedLot sightings, service tickets, demo logs, trade paperwork and payoff records, each with its source
3Position proposedOpen variances, curtailment aging and confidence
4Controls appliedLedger-match checks, per-facility curtailment rules, remittance-window checks and confidence threshold
No human action required

Stages 1 to 4 run unaided, and nothing is certified at any of them — under §9-311(d) dealer inventory is perfected by filing, not by title, and the controller's lane opens at the confidence gate.

5DecisionBranches at the confidence threshold
High confidence

Goes to the controller to certify.

Low confidence

Adds an office-manager read first.

Controller approval

The position is held with its evidence, its open variances and the confidence.

Approve · Amend · Send to manager review
Approved — released for signature
6Lender reporting updatedOnly where write access and approval policy allow it
7Outcome evaluatedAmendments, audit exceptions, days a unit ran unremitted and shortfalls found on inspection
Amendments

Every controller amendment is counted, including exceptions raised at the audit.

What should not run autonomously

Human approval stays in control

Outside the boundary — human approval required8 items
Signing the certification the personal guaranty stands behind.
Remitting a payoff or releasing funds to the lender.
Deciding a unit is out of trust rather than unlocated.
Agreeing a curtailment, extension or waiver with the lender.
Automation boundaryAgent acts unaided
Reconcile the lender's schedule to what is on site.
Carry each unit forward with its status, its evidence and its date.
Age each unit against that facility's own curtailment schedule.
Flag variances and stale units, and hold the position.
Any write happens inside the boundaries agreed at implementation, never as a signature.
Writing off, wholesaling or trading a financed unit.
Explaining a shortfall to the auditor on the store's behalf.
Moving proceeds between the operating and payoff accounts.
Changes to facility rule sets, thresholds or approval boundaries.

Example output

One unit, annotated

Everything the agent reconciles is attached to the ledger line it was matched to.

Reconciliation output · single unitIllustrative example
Unit
Reconciled line
Days since floored
Schedule applied
Confidence
Signature required
New unit, in demo use
Status evidenced as a demonstrator from the demo log on file
Day 62 on floor
This facility's terms
92%
Named controller
As receivedTaken from the lender's schedule and the demo log on file — nothing on this side is asserted by the agent.
Evidence used Lender ledger line Demo log, dated Last lot sighting
Why this statusIt is in demo use and unsold, not a unit delivered and a person still decides.
ActionApproveAmendSend to manager review
What the score decidesBelow the configured threshold the position picks up a manager read before the controller.

Value

Where AI adds value

The same four claims, placed at the point in the workflow where each one applies.

Where the value landsValue 01 – 04
Every financed unitFrom the lender's schedule
03Reconciliation

Match to the lender's ledger

Draw on the outstanding schedule, the DMS and this facility's terms, read against the OCC booklet as at August 2026.

01Approved path

The guaranty is personal

Matched units arrive reconciled, aged and evidenced, with the open variances named.

02Human review

Put a read between preparer

In AFC v. Nunez (Bankr. N.D. Tex., March 2019) trust language and a signed guaranty were not enough on their own — dated per-unit evidence is what differs.

04Build an evidence trail

The unit, the ledger it was matched to and the controller who certified stay on the audit.

Integrations

Typical integrations

Five system groups connect to the same agent. Which of them are in scope is decided in discovery.

Floorplan lendersAlly · GM Financial
Chase · NextGear · AFC
DMS inventoryCDK · Reynolds · Tekion
Dealertrack · Auto/Mate
Dealership accountingDMS accounting module
Schedules and journals

Agent

Floorplan & inventory audit

Reads the ledger
Reconciles the units
Holds for the controller

Banking and treasuryPayoff and remittance files
Operating-account feeds
Observability & evaluationOpenTelemetry · Langfuse
Supported monitoring/evaluation sources

Integration availability depends on the client's existing systems and API access.

Agent controls

Six layers between the model and the certification

Six controls, tightening as they go inward. What clears the last one is listed in the map below.

L6 · Outermost — last line of defenceInward → L1 · closest to the model
L6Rollback / safe modePull reconciliation back to variance reporting when evaluation or production signals degrade.Roll back
L5Version monitoringTrack model, prompt, curtailment-rule and facility-configuration changes.Track
L4TraceabilityRecord the ledger line, the evidence, the variance, the amendments and who approved.Record
L3Controller approvalHold positions for a named controller; it governs release, not whether the units beneath were located.Gate
L2Policy guardrailsTest positions against this facility's own remittance window and curtailment terms, not a market schedule; a failure returns them.Restrict
L1Confidence thresholdsRoute low-confidence positions to a manager read before the controller sees them.Require review
Model corePosition proposed — open variances, curtailment aging and confidence
L1 – L2Test whether a position may stand
L3Puts the release in a controller's hands
L4 – L5Keep the unit and the ledger behind it
L6Narrows to variance reporting when signals degrade

How Nestack evaluates it

Evaluate the reconciliation workflow — not only the final count.

Coverage runs the whole depth of the workflow, and every layer is cut by slice.

Surface — the position the lender reads
Depth of coverage ▼
E1Final-output evaluationDid each VIN on the lender's schedule resolve to one documented status?
E2Step-level evaluationDid the agent use the right facility, schedule and remittance window?
E3Tool evaluationDid it read and write the correct unit and the correct ledger?
E4Confidence calibrationDo low-confidence positions actually attract more controller amendments?
E5Slice evaluationHow does performance change across specific unit types?
E6Business outcomeHow many units were found short, or corrected after the audit had closed?
Floor — the outcome the dealer answers for

Failure modes

Where each failure originates in the agent

Seven failure modes, placed at the stage each one originates.

Agent lifecycleDirection of processing →
01 · Retrieval1 mode
DH-03

DMS-side reconciling

The count is run from the DMS, not the lender's ledger.

Stage gathersLedger schedules, unit statuses and facility terms
02 · Reasoning2 modes
DH-04

Single-schedule default

One curtailment schedule is applied across three facilities.

DH-06

Funding-date clock

The remittance clock is run from funding, not delivery.

Stage proposesVariances, aging, evidence gaps and confidence
03 · Tool / write2 modes
DH-02

Low-confidence release

Agent releases a position despite insufficient certainty.

DH-05

Rationale written late

A variance is explained from current system state.

Stage writesOnly where write access and approval policy allow it
04 · Output1 mode
DH-01

Unit unresolved

A financed VIN carries no documented status.

Stage returnsThe position the controller signs and the lender reads
05 · Change / Version1 mode
DH-07

Silent threshold drift

A model or rule change widens what the agent accepts.

Stage tracksModel, prompt, curtailment rules and facility terms
Sev-1 · certified outside the boundary Sev-2 · a wrong position reaches the lender Sev-3 · evidence thin, position to review

Affected slices

A clean group figure can hide one facility

A group-level variance figure reads clean while one lender carries most of the exceptions. Nestack reports the exception rate by slice, not only in total across the group.

Slice performance — reported separately, not only in aggregateIllustrative example
SliceFailure rateLift Lift vs. thresholdStatus
Demos and service loaners7.2%3.7× Review
Units out on dealer trade4.9%2.5× Review
Sold, contract in transit3.4%1.7× Watch
Units on the lot, unsold1.7%0.9× Normal
Bar: variance-rate lift vs. on-lot unsold baseline · scale 0–4.0× · tick marks the 2.0× review threshold 2 of 4 slices over threshold

Evidence-linked improvement

A variance the guaranty rides on

A cycle shuts when the unexplained variance is a case the next release has to catch. That suite is what the next count run is measured against.

Improvement cycle · five stagesSwitchback — the path turns at Improve and returns at Learn
01Detect

Variance rate rises on one facility.

02Diagnose

The car that sold on Friday and stayed on the ledger until the audit is read back until the cause is one thing.

03Improve

The change is numbered on the way out, with the units that revealed it attached.

04Verify

A variance case that has not cleared keeps the release parked until it does.

05Learn

It joins the suite for good, and the reconciliation rules are edited to match.

Learn → DetectThe return edge. Detection next time runs against a suite that is one case longer.

Typical build scope

Twelve workstreams across six weeks

The build scope read against the delivery timeline. Week structure follows the six-week plan — discovery, sources, reconciliation workflow, evaluation, integration, then production validation and handover.

Workstream Week 1Week 2Week 3Week 4Week 5Week 6
01Floorplan reconciliation discovery and boundary work.
02Lender-ledger and DMS assessment.
03Per-facility curtailment and remittance rule mapping.
04Unit and ledger ingestion, normalisation.
05Reconciliation logic and evidence binding.
06Confidence scoring and variance routing.
07Controller approval workflow.
08Lender-ledger and DMS integration.
09Variance and curtailment cases.
10Guardrails and reconciliation controls.
11Unit-trail instrumentation.
12Deployment, documentation and Agent Care handover.
12 workstreams · 6 weeks · bar shows the weeks a workstream is active — several run in parallel Final scope and sequence confirmed in discovery

Engagement tiers

What each tier includes

Rows are the capabilities named in each tier's scope. Higher tiers include everything below them.

Capability✓ in scope · — not at this tier PilotOne lender, one store ProductionProduction ledger feeds AdvancedMultiple lenders / rooftops
Introduced at Pilot
Reconciliation to your ledger
Controller approval
Reconciliation-accuracy baseline
Introduced at Production
Reporting by lender
Approval workflow in your systems
Approved write-back
Ledger and DMS integration
Introduced at Advanced
Complex multi-facility terms
Multi-stage store approvals
High unit volume
Multi-lender inventory controls
Build price From $5,000 From $8,000 Custom quote
Final build priceConfirmed after discovery based on integrations, workflow complexity, transaction volume, approval controls and deployment requirements.
Separate from buildBuild pricing is separate from recurring Agent Care, which covers managed monitoring, evaluations, incidents and verified improvements after launch.

What we need from you

What you bring, and what we build with it

Each input maps to a piece of build scope and a week in the delivery timeline.

You bringWe build with it
01Your facility terms and ledger structure Unit and ledger ingestion and evidence mappingWeek 1
02Representative completed certifications Reconciliation baseline, rule extraction and evidence bindingWeek 2
03Your curtailment and remittance decisions Curtailment, remittance and automation-boundary mappingWeek 1
04Access to relevant APIs, feeds or exports Lender-ledger and DMS assessment, then integration setupWeek 2
05Positions you would not want certified Out-of-trust cases and the evaluation suiteWeek 4
06What no certification may cover Confidence scoring, variance routing, guardrails and approval controlsWeek 3
07Named controllers to approve positions Controller approval workflow, then pilot and production validationWeeks 5–6
Nothing else is required Deployment, documentation and Agent Care handover are ours.

Delivery timeline

Four phases across six weeks

Bands sit where the audit calendar puts them, which is why week 5 holds two phases.

Phase W1W2W3W4W5W6
Discovery W1
Build W2 – W3
Evaluate W4 – W5
Pilot & Launch W5 – W6
Week focus W1Reconciliation discovery, facility mapping and the automation boundary W2Source integration and the reconciliation baseline W3Reconciliation workflow, confidence logic and approval controls W4Evaluation suite, out-of-trust checks and failure-mode testing W5Ledger integration, pilot facilities and targeted corrections W6One audit cycle run under the controller, then Agent Care handover
Reading the bandA bar spans the weeks its work is named in, and no others. The week 5 overlap is real.
At the end of W6Once the cycle validates, Agent Care assumes the monitoring.
DurationSix-week plan shown · typical delivery 4–6 weeks depending on scope confirmed in discovery.

Next step · Automotive AI agent

Build a floorplan agent around the ledger your lender actually audits.

Show us your facility terms, your lender's schedule and who signs — today that is the person who prepared it. You get a per-unit reconciliation, a dated rationale on each variance, and a position a controller can certify.

Nestack Agents · Floorplan & inventory auditAGT-AUT-16 · Agent Care available after launch