Events with context
Associate a source event with its vehicle, time and available driver or journey context. Preserve the detection source and original reference.
Connected technology / Video & Camera Systems
Connect the event, the vehicle and the evidence.
Integrate compatible cameras and mobile video recorders so available events and footage can support safety reviews, coaching and incident investigation.
Plan your integrationMAKE THE CONNECTION USEFULStart with the decision.
Start with the decision.
Design the data around it.
Video integration has several independent parts: the event metadata, camera-channel information, live-stream capability, historical retrieval and retained evidence. A successful event connection does not automatically enable every media function. Fleetly scopes these capabilities separately so teams know what is available, what depends on the device being online and what to do when evidence is missing.
PRACTICAL USE CASES
Where this integration can add value.
Live and historical media
Confirm supported channels, recording formats, connectivity and retrieval behaviour. Keep stream availability separate from the existence of an event.
Human review and AI outputs
Where an approved analysis service is connected, distinguish its classification and confidence from the final reviewer decision. Inconclusive evidence must remain visible.
Evidence handling
Define authorised export, retention and incident association. Preserve the original time and source when footage moves through a review workflow.
Every connection has a working path.
Select a stage to see what needs to happen and why it matters.
Video & Camera Systems / Detect
Receive a source event
Identify the vehicle, device, time and detection type. An alert is a signal for review, not a confirmed incident.
Read the complete workflow
- Receive a source event
Identify the vehicle, device, time and detection type. An alert is a signal for review, not a confirmed incident.
- Find the relevant media
Match the channel and recording window. Check device status and retained footage before requesting retrieval.
- Interpret the available evidence
Combine video, journey context and any approved analysis. Record confirmed, dismissed or inconclusive outcomes.
- Choose the next action
Route the reviewed event to coaching, incident management or a technical investigation.
- Retain an accountable record
Store the authorised evidence reference and review history according to the agreed retention and access process.
DATA WITH CONTEXT
Agree the meaning before moving the data.
These are scoping considerations, not field names or a published API schema. Confirm the exact contract for your implementation.
| Data area | Why it matters | What to validate |
|---|---|---|
| Event reference | Links detection and subsequent review. | Keep source identity across reprocessing. |
| Camera channel | Identifies the physical viewpoint. | Validate names, orientation and channel availability. |
| Media availability | Describes whether footage can be accessed. | Distinguish pending, unavailable and retrieved states. |
| Analysis result | Provides an automated interpretation where supported. | Keep confidence and classification separate from human judgement. |
PREPARE YOUR INTEGRATION
A useful first conversation starts here.
Mark what you have prepared and use the remaining items to shape your discussion. This checklist stays in this page session.
0 / 5Preparation items checked
Preparation only; this is not a compatibility assessment or production approval.
Opens your email app with an editable summary. Nothing is sent automatically.
COMMON QUESTIONSUnderstand the scope.
Understand the scope.
Plan with confidence.
Can an offline vehicle provide a live stream?
No. Live streaming needs the required device connectivity. Historical retrieval also depends on the integration, retained recording and a viable transfer path.
Do all camera integrations include AI?
No. Detection and analysis capabilities vary. An external analysis service requires its own supported integration and agreed processing workflow.
Does high AI confidence prove responsibility for an incident?
No. Confidence describes a model output under its own definition; it does not determine liability or replace evidence review.
What if footage cannot be retrieved?
Keep the event and retrieval status visible, record the reason where known and investigate promptly. Missing footage must not be treated as proof that nothing happened.
SUPPORTING RESOURCES
Take the next step with a clear plan.
Confirm your production scope.
Features and integration methods depend on the agreed devices, interfaces, permissions and enabled modules. Request the current technical documentation before implementation; never include passwords or access tokens in an initial enquiry.