Product
The AI operating system for logistics fleets.
Khwand replaces reactive fleet management with an autonomous employee that monitors, maintains and optimises every vehicle — from the documents and reports you already produce.
Every vehicle becomes a digital twin
Identity, operational history, maintenance history and an AI health score — continuously updated from real data, no hardware required.
The AI operates end to end
From a driver report to a completed repair: diagnose, create the work order, choose the workshop, source the parts, schedule the job.
Autonomy you control
Grant the level of autonomy you're comfortable with. Low-confidence decisions are surfaced for your review. Everything is logged.
The AI operating system
Seven specialists. One autonomous employee.
Each agent is responsible for one domain of fleet operations. Together, coordinated by an orchestrator, they run the fleet end to end — and log every decision.
Fleet Intelligence
Sees the whole fleet
Continuously understands fleet state — health, risk and activity — and surfaces what needs attention before it becomes a problem.
Maintenance
Diagnoses and recommends
Analyses driver reports and documents, retrieves similar past cases from vector memory, and creates real work orders.
Workshop
Picks the right partner
Selects and coordinates repair workshops by capability, location, cost and historical performance.
Procurement
Sources the right parts
Identifies required parts, evaluates suppliers, and estimates cost and lead time for every repair.
Scheduling
Minimises downtime
Optimises when and where repairs happen to keep trucks on the road and downtime to a minimum.
Compliance
Stays regulation-ready
Tracks inspections, certifications and regulatory gaps across the fleet so nothing lapses.
Driver Communication
Keeps drivers in the loop
Processes driver reports and coordinates notifications so issues enter the system the moment they are reported.
Orchestrator
Runs the workflows
Coordinates multi-agent workflows, creates work orders and records every AI decision in the action log.
Autonomous workflows
Multi-agent workflows that get things done.
The orchestrator sequences specialist agents to solve real fleet problems — start to finish, with a result you can verify.
Vehicle issue report
A driver reports a fault and the system autonomously analyses it, creates a work order, selects a workshop and schedules the repair.
- 1Maintenance agent diagnoses severity
- 2Similar cases recalled from vector memory
- 3Work order created
- 4Workshop selected
- 5Repair scheduled to minimise downtime
Maintenance resolution
A full coordinated response to a maintenance issue — analysis, workshop selection, parts procurement and scheduling.
- 1Fleet intelligence analyses fleet state
- 2Maintenance agent recommends actions
- 3Workshop agent selects a partner
- 4Procurement sources parts
- 5Scheduling agent minimises downtime
Workshop selection
The system picks the optimal workshop for a specific repair by comparing capability, location, cost and historical performance.
- 1Fleet intelligence analyses repair requirements
- 2Workshop agent compares candidates
- 3Best partner selected by cost and reliability
Procurement
For repairs needing parts, the system identifies what is required and sources suppliers with cost and lead time.
- 1Maintenance agent identifies required parts
- 2Procurement agent finds suppliers
- 3Cost and lead time estimated
Scheduling
After a vehicle enters maintenance or new orders arrive, assignments are rebalanced to keep trucks on the road.
- 1Fleet intelligence analyses assignments
- 2Scheduling agent re-optimises
- 3Downtime minimised
Compliance check
The system verifies vehicles against inspections and certifications so nothing lapses.
- 1Fleet intelligence reviews vehicle status
- 2Compliance agent validates requirements
- 3Gaps flagged with recommendations
Incident response
When a vehicle breaks down or is involved in an incident, the system coordinates maintenance, re-optimises the schedule and communicates with drivers.
- 1Incident severity assessed
- 2Immediate maintenance determined
- 3Fleet schedule re-optimised
- 4Affected drivers notified
Fleet optimisation
Periodic optimisation of assignments, maintenance scheduling and resource allocation across the whole fleet.
- 1Fleet performance analysed
- 2Assignments and routes optimised
- 3Preventive maintenance scheduled
Under the hood
Trustworthy engineering.
Khwand is multi-tenant by design, with PostgreSQL row-level security, a complete AI action log and a stack built for scale.
| Layer | Technology | Why |
|---|---|---|
| AI agents | LangGraph + LLM providers (NVIDIA NIM, OpenAI, Anthropic) | Multi-step reasoning across fleet domains |
| Vector memory | pgvector + sentence-transformers | The AI learns from your fleet's history |
| Database | Supabase PostgreSQL with RLS | Defence-in-depth tenant isolation |
| API | FastAPI — 17 endpoint modules | Every endpoint returns real database data |
| Background work | Redis + Celery | Event-driven maintenance checks and dispatch |
Put an AI employee on your fleet. Starting at £10 a vehicle.
Join the early access programme and let Khwand monitor, maintain and optimise your vehicles — no hardware required.