Nestack Agent Care
Industries / Transportation / Hours-of-service agent

Transportation AI agent · Hours-of-service records

Hours-of-Service Records AI Agent

Corroborate every duty day the driver signed for against the documents that arrive afterwards, run each submission clock from whichever trigger falls later, and hold the record for the compliance manager.

4–6 weeksTypical delivery
Your stackDeployment
Driver-signedClock per doc
Agent CareAfter launch

What this agent does

Corroborates the day, never certifies it

In
01

A duty day closes, and § 395.8(f)(7) leaves the certifying signature to the driver who was there.

02

A signed record leaves the cab, and § 395.8(a)(2)(ii) sets the window for reaching the carrier.

Reason
03

A document reaches the driver late, and § 395.11 runs its clock from that day, not from the trip.

04

An expense receipt covers on-duty not driving time, and § 395.11 counts it a supporting document.

05

A day draws more paper than the ceiling, and § 395.11 caps what the carrier is obliged to retain.

Decide
06

A record is received by the carrier, and § 395.8(k)(1) starts the retention period at receipt.

07

A copy stays in the cab, and § 395.8(k)(2) wants the run of recent days available on duty.

Out
08

An ELD operator manual is no longer carried: § 395.22(h)(1) was rescinded effective 22 July 2026.

09

Execute write actions only inside the approval boundaries agreed during implementation.

Product statement

The agent assembles and corroborates. The driver certifies the record of duty status; a named compliance person reviews it and retains it.

Example workflow

One duty day, document to retention

AgentHuman
1Duty record or document receivedElectronic logs, bills of lading, dispatch records, expense receipts and payroll sheets
2Day context assembledThe driver, the duty day, the category each document falls in and the date it carries
3Day state draftedThe documents present, what they corroborate, the gaps and completeness
4Controls appliedClock checks, category checks, corroboration checks and completeness confidence
No human action required

Stages 1 to 4 run unaided, and nothing is certified at any of them — the agent is corroborating, and the compliance lane opens at the completeness gate.

5DecisionSplits at the completeness gate
Day corroborated

Goes to the compliance manager to review and retain.

Anything unsupported

Adds a compliance-desk read first.

Compliance review

The day is held with its gaps, its clocks and the driver who signed it.

Retain · Append document · Send to compliance
Retained — by the named compliance manager
6Driver and log records updatedOnly where write access and records policy allow it
7Outcome evaluatedCorroboration depth, clock coverage, desk corrections and what review found
Corrections

Each compliance-desk correction is counted in the evaluation.

What should not run autonomously

Human approval stays in control

Outside the boundary — human approval required8 items
Certifying that a duty day is true and correct.
Editing or annotating a record of duty status.
Deciding that a document is genuinely missing.
Destroying a record before retention has run.
Automation boundaryAgent acts unaided
Bind each document to the duty day it corroborates and the date on it.
Time each submission clock from the later of two dates.
Pursue a document that has not yet reached the driver.
Mark the on-duty hour that no receipt or dispatch record can support.
Nothing is certified or retained except by a named person, inside agreed boundaries.
Judging whether a driving window was exceeded.
Telling an inspector the day is corroborated.
Setting which documents a duty day requires.
Changes to logs, payroll or dispatch records.

Example output

One duty day, annotated

Our driver support agent stops the moment a driver cites hours, fatigue or a defect; this record is what one duty day left behind.

Day state · single driverIllustrative example
Duty day
Recorded as
Driver
Source of record
Confidence
Held for
Certified record, § 395.8(f)(7)
Certified by the driver, one on-duty hour unsupported
Long-haul driver
Electronic log and fuel receipt, 3 August 2026
Held unretained
The compliance manager, by name
As receivedTaken from the electronic log and the documents filed against it — it reaches as far as those sources do.
What the record holds Electronic log entry Fuel receipt, dated Dispatch record
Why no compliance call hereWhether a day was driven lawfully is a Part 395 judgement, not a model call.
ActionRetainAppend documentSend to compliance
What the score decidesBelow the configured threshold the day picks up a desk read before the manager sees it.

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 duty dayFrom the driver who signed it
03Record

Where the record is used

Our dispatch and routing agent plans against the driver's recorded hours; this page is the record those hours came out of, read after the fact.

01Approved path

The person who was there signs

§ 395.8(f)(7) leaves the certifying signature to the driver, while § 395.8(e)(1) binds the carrier too.

02Human review

What was checked, and not found

The carrier retention ceiling is verified in the text of § 395.11; the widely repeated rule that the first and last document of a duty period must be among those kept was looked for in that section and not found there.

04Build an evidence trail

The day, the document that corroborates it and the driver who certified stay on the record.

Integrations

Typical integrations

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

Electronic logging devicesELD records · driver logs
Certified records of duty status
Trip and dispatch systemsTMS · dispatch records
Bills of lading and itineraries
Expense and payroll recordsFuel cards · expense systems
Settlement sheets and receipts

Agent

Hours-of-service records

Reads the day
Corroborates the hours
Holds for the manager

Fleet messagingFleet management system
Mobile communication records
Observability & evaluationOpenTelemetry · Langfuse
Supported monitoring/evaluation sources

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

Agent controls

Six gates between the model and the manager

Six gates, and no two of them shut on one thing. What passes all six is named in the map below.

L6 · Outermost — last line of defenceInward → L1 · closest to the model
L6Rollback / safe modeNarrow the agent to record assembly when evaluation or production signals degrade.Roll back
L5Version monitoringTrack model, prompt and clock rules; the § 395.22(h)(1) removal took effect on 22 July 2026.Track
L4TraceabilityLog each document, its category, the duty day it corroborates and every read of the record.Record
L3Manager releaseHold the day for the compliance manager; it governs release, not whether the day was lawful.Gate
L2Clock guardrailsTest each clock against § 395.8(k)(1) and § 395.11 as configured; retention starts at receipt.Restrict
L1Confidence thresholdsRoute a thinly corroborated day to a compliance read before the manager sees it.Require review
Model coreDay assembled — the record, the documents, the clocks and completeness
L1 – L2Test whether a day may stand
L3Puts the retention in a person's hands
L4 – L5Keep the day and the document behind it
L6Narrows to record assembly when signals degrade

How Nestack evaluates it

Evaluate the whole assembly — not only the duty record that comes out.

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

Surface — the record an inspector reads
Depth of coverage ▼
E1Final-output evaluationDid the record hold what the duty day actually generated?
E2Step-level evaluationDid the agent read the right driver, the right day and the current log?
E3Tool evaluationDid it read and write the correct duty day and the correct document?
E4Confidence calibrationDo thinly corroborated days actually attract more desk corrections?
E5Slice evaluationHow does performance change across specific driver groups?
E6Business outcomeHow many days needed a correction before the manager retained one?
Floor — the record the carrier answers for

Failure modes

Where each failure originates in the agent

Seven failure modes, each pinned to the stage where it first shows.

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

Stale log read

The log read is not the one the driver later certified.

Stage gathersThe drivers, the days, the documents and the dates
02 · Reasoning2 modes
JG-04

Day asserted, not corroborated

A duty day is called supported with no paper behind it.

JG-06

Proposed rule read as live

The English proficiency NPRM is treated as in force.

Stage proposesThe days, their documents and corroboration
03 · Tool / write2 modes
JG-02

Thin day passed forward

A day moves on without the compliance read.

JG-05

Document bound to wrong day

A receipt is filed against the wrong duty period.

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

Retained, corroboration unrecorded

The record shows retention but not what supported it.

Stage returnsThe record a manager retains and an inspector reads
05 · Change / Version1 mode
JG-07

Silent retention drift

A settings change shifts the clock while the record stands.

Stage tracksModel, prompt, retention rules and corroboration fields
Sev-1 · a day retained on no documents Sev-2 · wrong document reaches the day Sev-3 · source degrades, day holds open

Affected slices

Drivers on late receipts absorb the corrections

A driver-level corroboration figure can read clean while drivers on late receipts carry most of the corrections. Nestack reports the correction rate by driver group, not only in total.

Slice performance — reported separately, not only in aggregateIllustrative example
SliceFailure rateLift Lift vs. thresholdStatus
Drivers on late receipts9.0%3.6× Review
Team-driven duty days6.4%2.6× Review
Drivers on short-haul exception4.0%1.6× Watch
Single drivers on one lane1.5%0.6× Normal
Bar: correction-rate lift vs. single-lane driver baseline · scale 0–4.0× · tick marks the 2.0× review threshold 2 of 4 slices over threshold

Evidence-linked improvement

What an uncorroborated hour costs

A cycle is done when the uncorroborated on-duty hour has become a case the next release must pass. That suite is what the next day filed is measured against.

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

Correction rate rises on drivers with late receipts.

02Diagnose

The fuel slip that surfaced three weeks late and started its own clock is read back until a single cause is left standing.

03Improve

Number the change; the days that drove it are filed underneath it.

04Verify

Each touched day case is run once more, and one red holds it back.

05Learn

It stays on as a standing test, and the corroboration rules move with it.

Learn → DetectThe return edge. Every later day meets a suite that has grown by one case.

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, record assembly, evaluation, integration, then production validation and handover.

Workstream Week 1Week 2Week 3Week 4Week 5Week 6
01Duty-day discovery and automation-boundary work.
02Log, document and payroll sources.
03Document-to-category and clock-coverage mapping.
04Duty record ingestion.
05Driver, day and source binding.
06Corroboration scoring and review routing.
07Compliance-manager retention workflow.
08Log and payroll-system integration.
09Corroboration and clock cases.
10Guardrails and certification controls.
11Day-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 group, one retention cycle ProductionProduction duty-record workflow AdvancedMultiple groups / terminals
Introduced at Pilot
Record assembly to your categories
Compliance-manager release
Document-category baseline
Introduced at Production
Reporting by driver
Review workflow in your systems
Approved write-back
Electronic-log integration
Introduced at Advanced
Multi-terminal fleets
Cross-category document packs
Large driver fleets
Multi-clock retention controls
Build price From $5,000 From $8,000 Custom quote
Final build priceConfirmed after discovery based on integrations, workflow complexity, fleet size, 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 drivers and the days they signed for Fleet mapping and duty-record captureWeek 1
02Representative logs, receipts and dispatch records Source binding, clock logic and the corroboration baselineWeek 2
03Your document categories under § 395.11 Day mapping, driver binding and the automation boundaryWeek 1
04Access to relevant APIs, feeds or exports Log, receipt and dispatch-source assessment, then integration setupWeek 2
05Days you would not want certified Corroboration cases and failure testingWeek 4
06What no duty record may prove Corroboration scoring, review routing, guardrails and release controlsWeek 3
07A named compliance manager to retain Retention 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

Each band is the honest working time one phase takes, and the fifth is shared because that work is.

Phase W1W2W3W4W5W6
Discovery W1
Build W2 – W3
Evaluate W4 – W5
Pilot & Launch W5 – W6
Week focus W1Duty-record discovery, day mapping and the automation boundary W2Source integration and the corroboration baseline W3Record assembly, clock logic and release controls W4Evaluation suite, corroboration cases and failure-mode testing W5Log integration, pilot days and targeted corrections W6One retention cycle run under the compliance manager, then Agent Care handover
Reading the bandEach bar spans only the weeks the work it names actually occupies. The fifth holds two, because the work does.
At the end of W6When the duty record validates, Agent Care takes the agent on.
DurationSix-week plan shown · typical delivery 4–6 weeks depending on scope confirmed in discovery.

Next step · Transportation AI agent

Build a duty-record agent around the day your driver already signs for.

Show us one duty day and the documents behind it. That signature has to survive a roadside inspection, a compliance review and the whole retention period on your own shelf. Planning against recorded hours is a different agent.

Nestack Agents · Hours-of-service recordsAGT-TR-20 · Agent Care available after launch