Load Procurement
ActiveScans connected load boards continuously and scores each load against your real cost per mile.
- Boundary
- Never surfaces freight below the configured margin floor.
- Recent
- 184 loads evaluated · 6 above threshold · 1 booked
The trucking company
that runs itself.
Agents find freight, price it against your real cost per mile, dispatch only HOS-legal drivers, catch faults before they strand a truck and invoice on delivery — inside policies you write.
Operating·3 agents, 5 engines
Representative operating picture- LOAD PROCUREMENTevaluating 184 loadsDAT · Truckstop
- RATE INTELLIGENCEmargin validated$1.84/mi breakeven
- COMPLIANCEHOS verified8h14m drive remaining
- DISPATCHmatching driver + tractorUnit 284 · Dallas
- MAINTENANCEfault detectedSPN 3364 · derate risk
- FINANCEinvoice preparedPOD attached
Decisions
DISPATCH Load #48291 → Unit 284 · projected contribution +$684
COMPLIANCE PHX → DEN refused · driver 1h42m short of HOS buffer
RATE Load #48310 declined · $1.61/mi below breakeven
MAINTENANCE Unit 112 booked · aftertreatment fault, Thursday 07:00
FINANCE Invoice #2291 prepared · $3,140 · awaiting approval
FUEL Transaction flagged · 61gal on a 50gal tank, Amarillo
Agent · reasonsEngine · deterministic
Ecosystem
Works with the systems that already run your fleet.
Your telematics, load board, fuel card and accounting subscriptions stay in your own name. We act on your behalf, never resell the feed, and never mark up a provider you already have a contract with.
TelematicsAPI adapter
TelematicsAPI adapter
Load boardAPI adapter
Load boardAPI adapter
AccountingAPI adapter
Fuel cardFile import
Fuel cardFile import
Fuel cardFile importProvider-agnostic by design. Adapters sit behind one internal carrier model, so a new ELD or load board is a connector — not a rebuild.
Fuel cards are an import, not an API. WEX, Comdata and EFS publish no transaction API, so those land as your own export. We would rather say so than call it an integration.
Built to connect across the logistics stack
Roadmap · not yet availableCategories the connector layer is designed to take next. Named as categories rather than vendors — we have no agreement with any provider in these areas, and a logo here would imply one.
- Mapping & routing
- EDI
- Insurance
- Maintenance
- Payments
- ERP
- Warehouse
- Broker systems
- Factoring
- Drug & alcohol consortium
Architecture
One intelligence layer across your entire operation.
Fragmented software is the reason a carrier cannot answer a simple question like “should this truck take this load?” The ELD knows the hours, the load board knows the rate, accounting knows the cost, and none of them know each other. One state model is what makes the decision possible.
Systems
What comes in
- Load boards
- ELD / telematics
- Fuel cards
- Accounting
- FMCSA
- Documents
- Maintenance
- Driver records
Intelligence layer
Unified operational state
Every source normalised into one model of the carrier, per tenant.
PoliciesCost per mileHours of servicePositionsLoadsDriversEquipmentFinancialsRisk
Agent layer
- Procurement
- Dispatch
- Predictive maintenance
- Pricing · HOS · fraud · compliance · billing
Actions
What goes out
- Book freight
- Assign equipment
- Refuse unsafe dispatch
- Schedule repair
- Flag fuel fraud
- Create invoice
- Escalate exception
Every action leaves through the same gate: the policy engine decides whether it executes, waits for a human, or is refused — and the decision is recorded either way.
The distinction
Your TMS records what happened.
logistick.ai decides what happens next.
A system of record stores information. A dashboard visualises it. Both still require a person to read the screen, hold six constraints in their head and act before the load is gone. logistick.ai operates on the information directly.
- 01Signal
Truck 284 goes empty in Dallas
- Telematics reports the delivery
- Driver Dre M. · reefer · 53ft
- 02Context
What the system already knows
- 8h14m drive time left on the 11-hour clock
- Operating cost $1.84/mi
- 62h of the 70-hour cycle used
- 03Decision
Load #48291 clears every policy
- DAL → MEM · 452 mi · $1,240 · $2.74/mi
- Deadhead 18 mi · under the 80 mi ceiling
- Projected contribution +$684
- 04Action
Dispatched under policy
- Driver and tractor assigned
- Broker confirmation sent
- Above the auto-book threshold — no human needed
- 05Audit
Written down before anyone asks
- Inputs, matched policy, reasoning and result recorded
- Reversible, and attributable to the agent that acted
Worked example using the demo carrier. The arithmetic is real; the load is illustrative.
The workforce
Specialists that share one operating state.
Three of these reason. The rest are deterministic by design — nobody should want a language model deciding whether a driver has hours left. Each one works from the same state, so a decision made by one is visible to all.
Dispatch
ActiveMatches freight to equipment and to drivers who are legally able to run it.
- Boundary
- Never dispatch below the required HOS buffer.
- Recent
- Refused PHX → DEN · driver 1h42m short of the buffer
Predictive Maintenance
ActiveReads fault codes as they arrive, grades severity and turns the serious ones into shop bookings.
- Boundary
- Derate-risk faults escalate the same day.
- Recent
- Unit 112 · SPN 3364 · booked Thursday 07:00
Rate Engine
DeterministicPrices every load against fixed and variable cost — fuel, driver pay, maintenance reserve, insurance, tolls, overhead and equipment.
- Boundary
- Anything under breakeven is flagged before booking, not at settlement.
- Recent
- Breakeven $1.84/mi · load declined at $1.61/mi
Hours of Service
DeterministicEnforces the 11-hour drive limit, 14-hour window, 30-minute break and 70-hour cycle on every candidate assignment.
- Boundary
- A dispatch that would break a clock is refused outright.
- Recent
- Every proposed assignment checked before it is offered
Fuel & Fraud
DeterministicReconciles card transactions against tank capacity, position and time to surface theft and card misuse.
- Boundary
- Volume above tank capacity is always flagged.
- Recent
- 61 gal on a 50 gal tank · Amarillo · flagged
Eight capabilities are registered as callable tools. Every invocation is logged against the run that made it.
Control
Autonomy without surrendering control.
Autonomy you cannot bound is not a feature, it is a liability. Every agent runs inside a policy you write, and you decide how much rope it gets — starting at none.
Agents may book loads automatically when projected margin is above 18%, deadhead is under 80 miles, and every hours-of-service check passes.
min_margin_pct: 18 · max_deadhead_mi: 80 · hos_ok: true → allow
Turn autonomy up gradually
- 1Observe
Agents run and record what they would have done. Nothing moves.
- 2Recommend
Proposals reach the queue with their reasoning and projected outcome.
- 3Approve
A human with the right role releases each action.
- 4Automate
Actions inside the policy execute on their own. The rest still wait.
Policy engine
Deterministic conditions, evaluated in priority order. First match decides: allow, require approval, or deny.
Approval thresholds
Anything above a value, or outside a boundary, waits for a person.
Role-based access
Six permission sets separate reading, writing, approving, fleet, money and PII.
Tenant isolation
Every row carries its carrier, enforced in the database rather than in application code.
Credential isolation
Provider credentials and OAuth tokens are encrypted with AES-256-GCM before storage.
Decision history
Inputs, matched policy, reasoning and result recorded for every proposed action.
Human override
Any proposal can be rejected, and any automation switched back to approval.
Incident escalation
Failures in unattended jobs open an incident and notify a channel a person watches.
What we do not claim. logistick.ai holds no SOC 2, ISO 27001 or HIPAA certification, and publishes no uptime guarantee. The controls above are implemented in the product; the certifications are not, and we would rather write that here than let you find out during procurement.
Operational state
Your operation, alive.
In most stacks a truck is nine rows in nine systems. The ELD knows where it is, the TMS knows what it is hauling, accounting knows what it cost, the maintenance vendor knows what is wrong with it, and nothing knows all four at once.
Here a truck is one object with a position, a driver, a legal clock, a mechanical condition, an operating cost and a revenue contribution. That is the reason an agent can make a decision a dashboard cannot even frame.
Unit 284 · resolvedAvailable
- Unit
- 284 · 2022 Freightliner Cascadia
- Position
- Dallas, TX · empty 22 min ago
- Driver
- Dre M. · 6 yrs · clean MVR
- Hours
- 8h14m drive · 62h of 70h cycle
- Equipment
- Reefer 53ft · reefer hours 4,120
- Health
- SPN 3364 active · derate risk
- Economics
- $1.84/mi operating · $2.74/mi last load
- Documents
- Insurance current · IFTA Q3 filed
- Contribution
- $18,430 this quarter
Platform
Sixteen modules. One operating system.
Not a suite of products that happen to share a login. Every module reads and writes the same operational state, which is why dispatch knows what compliance knows and finance knows what both of them did.
Operations
Moving freight, every day.
- Command CenterThe carrier at a glance
- DispatchAssignment and live progress
- LoadsBoard search, scoring, booking
- Fleet & ELDPositions, odometer, faults
Intelligence
Deciding and accounting.
- Agent ConsoleRuns, tools, policies, audit
- ApprovalsThe human gate
- FinancialsPer-load profit and settlement
- Fuel CenterSpend, IFTA, fraud detection
Risk
Staying legal and insurable.
- ComplianceDocuments and expiries
- InsurancePolicies and claims
- Carrier DiscoveryFMCSA authority and safety
- MaintenancePredictions and work orders
Network
Capacity, people and equipment.
- Driver MarketSourcing and DOT screening
- Shop NetworkRepair booking
- Truck MarketEquipment listings
- OutreachBroker and shipper pipeline
Outcomes
Optimise the metrics that actually move a carrier.
We are not going to quote you a percentage. logistick.ai has no production carrier to measure yet, and every number on a page like this one is invented until it is not. These are the measures the system is built to move, and the ones we will report against once there is a fleet to report on.
Revenue per truck
Loaded-mile ratio
Deadhead miles
Operating margin
Fuel cost variance
Truck utilisation
Days to invoice
Maintenance downtime
Compliance exposure
Dispatcher workload
900+
Trucks in one workspace, load-tested
16
Modules on one operational state
3 / 5
Reasoning agents / deterministic engines
Built for growing fleets and load-tested at 900 trucks in a single workspace — evidence of where it has been proven, not a ceiling.
Why now
The stack fragmented. Nothing sat above it.
A carrier running 200 trucks now operates seven systems that each know one piece of the business and none of which know each other. The integration point has been a dispatcher with eight tabs open, retyping between them.
Two things changed. Nearly every one of those systems now exposes an API, so the state of a carrier can finally be assembled in one place. And models became good enough to reason over that state — well enough to propose a decision, and cheaply enough to do it for every truck, continuously.
That makes an execution layer above the fragmented stack possible for the first time. Not another system of record competing for the same seat — a layer that reads them all and acts.
ELDTMSLoad boardFuel cardAccountingMaintenanceComplianceOne operating state
Where this goes
The next great logistics company may run with far fewer people, operating far more freight.
Today that means a carrier whose freight is found, priced, dispatched and billed by software that never sleeps, under policies its owner wrote. What it means at network scale is a question worth building toward.
Or read the integration detail first — most engineers do.