Transportation AI agent · EU posting & tachograph
EU Driver Posting & Tachograph AI Agent
Classify every leg of a journey — bilateral, transit, cabotage or cross-trade — track what each Member State then wants of it, and hold the pack for the transport manager who lodges.
- Typical delivery
- 4–6 weeks
- Deployment
- Your stack
- Lodged by name
- Leg by leg
- After launch
- Agent Care
What this agent does
Maps the journey, never lodges the declaration
In
- 01
A leg runs from the State of establishment and back, and Directive (EU) 2020/1057 exempts it.
- 02
A third stop joins that journey, and the additional-activity tolerance widened on 21 February 2024.
Reason
- 03
A cabotage job is taken, and the declaration belongs in the system before the posting begins.
- 04
A cross-trade run links two Member States, neither of them where the operator is established.
- 05
A journey only crosses a Member State, and transit with no loading or unloading is no posting.
Decide
- 06
A journey meets a roadside check, and no document beyond the closed list may be asked of a driver.
- 07
A posting ends, and a host-State request through the information system leaves eight weeks to answer.
Out
- 08
A van above the mass threshold crosses a border, and Article 2(1)(aa) has applied since 1 July 2026.
- 09
Execute write actions only inside the approval boundaries agreed during implementation.
Product statement
The agent classifies each leg and tracks what every Member State on it then wants. A named transport manager lodges the declaration; the driver operates the tachograph.
Example workflow
One journey, leg to lodgement
- 1Journey legs receivedTransport orders, consignment notes, vehicle registrations and tachograph downloads
- 2Leg context assembledThe driver, the vehicle, the Member States a leg touches and the day it ran
- 3Posting status draftedThe legs, the status each one creates, the gaps and completeness
- 4Controls appliedStatus checks, tolerance checks, country-symbol checks and completeness confidence
No human action required
Stages 1 to 4 run unaided, and nothing is lodged at any of them — the agent is classifying, and the compliance lane opens at the completeness gate.
Goes to the transport manager to lodge.
Adds a compliance-desk read first.
Compliance review
The journey is held with its legs, its gaps and the Member States they touch.
Lodge · Append evidence · Send to the compliance desk
Lodged — by the transport manager, named
- 6Journey and driver records updatedOnly where write access and records policy allow it
- 7Outcome evaluatedStatus accuracy, tolerance 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 required
8 items
- Deciding that a leg is bilateral and not a posting.
- Lodging the declaration before a posting begins.
- Answering a host-State request in the system.
- Operating the tachograph or entering a country symbol.
Automation boundary
Agent acts unaided
- ✓Mark each leg with the status it creates and the states it touched.
- ✓Track what every Member State on the journey then wants.
- ✓Count each additional loading activity already spent.
- ✓Flag the van that came into scope when Article 2(1)(aa) began to apply.
Nothing is lodged or answered except by a named person, inside agreed boundaries.
- Judging which employment law governs a driver.
- Telling an inspector that a journey was exempt.
- Setting the operation type a journey is filed as.
- Changes to journey, driver or vehicle records.
Example output
One journey leg, annotated
Our dispatch and routing agent plans against the hours already recorded; this record is what one leg of a journey created.
Journey state · single leg
Illustrative example- Leg
- Third stop, unloaded in Lyon
- Recorded as
- Posted status created, declaration due
- Operation type
- Cabotage
- Source of record
- Consignment note, 3 August 2026
- Confidence
- Held unlodged
- Held for
- The transport manager, by name
As received
Taken from the consignment note and the card download — it reaches as far as those sources do.
What the record holds
- Consignment note date
- Country symbols read
- Declaration reference
Why no status call here
Whether a leg is a posting is a Directive 2020/1057 call, not a model output.
What the score decides
Below the configured threshold the journey picks up a desk read before the manager sees it.
Action
LodgeAppend evidenceSend to the compliance deskValue
Where AI adds value
The same four claims, placed at the point in the workflow where each one applies.
Where the value lands
Value 01 – 04
Every leg
From the journey that ran it
Where the status is used
Our hours-of-service records agent works the United States document a driver signs; a fleet doing both keeps two record sets that do not translate.
The leg decides the status
Cabotage makes a driver posted and a bilateral leg does not, whatever the journey was planned to be.
What was checked, and not found
No further EU road-transport change taking effect in 2026 turned up, and that search covered only the tachograph and posting provisions — read it as not checked rather than as nothing there. Market access for these vehicles has been required since 21 May 2022, so confirm the Community licence too.
Build an evidence trail
The leg, the status it created and the person who lodged the declaration 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.
Journey and order records
- Transport orders
- consignment notes
- Vehicle registrations
Tachograph data
- Card and vehicle-unit downloads
- Country symbol entries
Employment and pay records
- Contracts
- payroll
- Timesheets and proof of pay
Agent
EU posting and tachograph
- Reads the legs
- Tracks each state
- Holds for the manager
Declaration and request records
- Public interface submissions
- Information-system requests
Observability & evaluation
- OpenTelemetry
- Langfuse
- Supported monitoring/evaluation sources
Integration availability depends on the client's existing systems and API access.
Agent controls
Six screens between the model and the manager
Six screens, and only the last is narrow. Whatever comes through is laid out in the map below.
L6 · Outermost — last line of defence
Inward → L1 · closest to the model
- L1 – L2
- Test whether a journey may stand
- L3
- Puts the lodgement in a person's hands
- L4 – L5
- Keep the leg and the status behind it
- L6
- Holds the declaration unlodged when signals degrade
How Nestack evaluates it
Evaluate the whole assembly — not only the journey map that comes out.
Coverage runs the whole depth of the workflow, and every layer is cut by slice.
Surface — the pack an inspector reads
- E1Final-output evaluationDid the map give each leg the status that leg actually creates?
- E2Step-level evaluationDid the agent read the right journey, the right vehicle and the current order?
- E3Tool evaluationDid it read and write the correct journey and the correct leg?
- E4Confidence calibrationDo low-confidence journeys actually attract more desk corrections?
- E5Slice evaluationHow does performance change across specific Member States?
- E6Business outcomeHow many journeys needed a correction before the manager lodged one?
Floor — the journey the operator answers for
Failure modes
Where each failure originates in the agent
Seven failure modes, each fixed at the stage where it first appears.
Agent lifecycle
Direction of processing →
01 · Retrieval
1 mode
- JL-03
Stale journey read
The order read is not the journey the vehicle ran.
Stage gathers
The journeys, the legs, the states and the dates
02 · Reasoning
2 modes
- JL-04
Status asserted, not shown
A leg is called bilateral without its evidence.
- JL-06
Extension read as retrofit
The 2026 scope extension is worked as Article 3(4).
Stage proposes
The legs, their status and completeness
03 · Tool / write
2 modes
- JL-02
Thin journey passed forward
A journey moves on without the compliance read.
- JL-05
Leg bound to wrong vehicle
A leg is filed against the wrong vehicle.
Stage writes
Only where write access and approval policy allow it
04 · Output
1 mode
- JL-01
Lodged, leg unrecorded
The record shows a lodgement but not the leg beneath it.
Stage returns
The journey a manager lodges and an inspector reads
05 · Change / Version
1 mode
- JL-07
Silent status regression
A configuration change moves the status, not the record.
Stage tracks
Model, prompt, status rules and journey fields
- Sev-1 · a leg cleared on no evidence
- Sev-2 · wrong status reaches the record
- Sev-3 · source degrades, journey unlodged
Affected slices
Cabotage legs in France absorb the corrections
A state-level status-accuracy figure can read clean while cabotage legs in France carry most of the corrections. Nestack reports the correction rate by Member State, not only in total.
Slice performance — reported separately, not only in aggregate
Illustrative example| Slice | Failure rate | Lift | Lift vs. threshold | Status |
|---|---|---|---|---|
| Cabotage legs in France | 6.4% | 3.6× | Review | |
| Cross-trade legs in Germany | 4.6% | 2.6× | Review | |
| Bilateral legs with two stops | 2.9% | 1.6× | Watch | |
| Transit through Austria | 1.2% | 0.7× | Normal |
Bar: correction-rate lift vs. Austrian transit baseline · scale 0–4.0× · tick marks the 2.0× review threshold
2 of 4 slices over threshold
Evidence-linked improvement
What a mistaken bilateral costs
A cycle ends when the mistaken bilateral leg is a standing case. That suite is what the next journey mapped is measured against.
Improvement cycle · five stages
Switchback — the path turns at Improve and returns at Learn
- 01
Detect
Correction rate rises on cabotage legs in France.
- 02
Diagnose
The third stop that turned a bilateral run into a posting nobody declared is read back until one cause remains.
- 03
Improve
Every change leaves numbered, with the journeys that caused it filed beneath.
- 04
Verify
A single red journey case is enough to keep the release back.
- 05
Learn
The case is permanent, and the status rules are rewritten with it.
Learn → Detect
The return edge. The next journey is measured against a suite 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, status mapping, evaluation, integration, then production validation and handover.
Workstream
Week 1Week 2Week 3Week 4Week 5Week 6
- 01Journey-status discovery and automation-boundary work.Week 1
- 02Order, tachograph and payroll sources.Week 1, Week 2
- 03Leg-to-operation and Member-State duty mapping.Week 1, Week 2
- 04Journey and leg ingestion.Week 2
- 05Driver, vehicle and state binding.Week 2, Week 3
- 06Completeness scoring and review routing.Week 3
- 07Transport-manager lodgement workflow.Week 3, Week 4
- 08Journey and tachograph-system integration.Week 5
- 09Status and tolerance cases.Week 4, Week 5
- 10Guardrails and lodgement controls.Week 4
- 11Journey-trail instrumentation.Week 5, Week 6
- 12Deployment, documentation and Agent Care handover.Week 6
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 Member State, one year | ProductionProduction posting workflow | AdvancedMultiple states / operators |
|---|---|---|---|
| Introduced at Pilot | |||
| Status mapping to your journeys | ✓ | ✓ | ✓ |
| Transport-manager release | ✓ | ✓ | ✓ |
| Operation-type baseline | ✓ | ✓ | ✓ |
| Introduced at Production | |||
| Reporting by member state | — | ✓ | ✓ |
| Lodgement workflow in your systems | — | ✓ | ✓ |
| Approved write-back | — | ✓ | ✓ |
| Tachograph-record integration | — | ✓ | ✓ |
| Introduced at Advanced | |||
| Multi-state operations | — | — | ✓ |
| Cross-state evidence packs | — | — | ✓ |
| Large journey volumes | — | — | ✓ |
| Multi-state posting controls | — | — | ✓ |
| Build price | From $5,000 | From $8,000 | Custom quote |
- Final build price
- Confirmed after discovery based on integrations, workflow complexity, journey volume, approval controls and deployment requirements.
- Separate from build
- Build 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 bring
We build with it
01Your journeys and the states they touch
→Journey mapping and leg capture
Week 102Representative orders, notes and card downloads
→Source binding, status logic and the operation-type baseline
Week 203Your State of establishment and the licences held
→Status mapping, journey binding and the automation boundary
Week 104Access to relevant APIs, feeds or exports
→Order, tachograph and payroll-source assessment, then integration setup
Week 205Journeys you would not want declared
→Status cases and failure-mode testing
Week 406What no posting declaration may cover
→Completeness scoring, review routing, guardrails and release controls
Week 307A named transport manager to lodge
→Transport-manager workflow, then pilot and production validation
Weeks 5–6
Nothing else is required
Deployment, documentation and Agent Care handover are ours.
Delivery timeline
Four phases across six weeks
Every bar is the true working span of a phase, so the fifth week has to hold two of them in here.
Phase
W1
W2
W3
W4
W5
W6
- DiscoveryW1
- BuildW2 – W3
- EvaluateW4 – W5
- Pilot & LaunchW5 – W6
Week focus
W1Journey discovery, status mapping and the automation boundary
W2Source integration and the operation-type baseline
W3Leg classification, tolerance logic and release controls
W4Evaluation suite, status cases and failure-mode testing
W5Record integration, pilot journeys and targeted corrections
W6One posting year run under the transport manager, then Agent Care handover
- Reading the band
- Each bar spans only the weeks its own work is named for. The fifth week holds a pair because the work does.
- At the end of W6
- Once the journey record validates, Agent Care takes over the watch.
- Duration
- Six-week plan shown · typical delivery 4–6 weeks depending on scope confirmed in discovery.
Next step · Transportation AI agent
Build a posting agent around the leg that decides the status, not the plan that preceded it.
Show us one journey and the legs it is made of. Send the week that worried you and we will map each leg to the status it creates, before the next declaration falls due. A United States hours record signed by the driver is a different agent.
- Typical delivery
- 4–6 weeks
- Deployment
- Your journey records
- Lodged by name
- Leg by leg
- After launch
- Agent Care available
Nestack Agents · EU posting
AGT-TR-24 · Agent Care available after launch