Handover disputes
It should be possible to quickly prove what was handed to the customer, the item condition and what happened during opening.
Video analytics for pickup points
An AI supervisor for pickup points: handover and intake control, returns, queues, storage, operator work, marketplace standards and dispute cases with evidence frames.
24/7events, notifications and evidence frames without manual archive review
RTSP / ONVIFworks with existing IP cameras at pickup points
handoverhandover, opening and dispute moments are linked to video
storagestorage area, intake, shortages, access and parcel movement
queuevisitor flow, peak hours and service SLA
Cloud / On-Prempilot at one point and scaling across the network
Background
Pickup points operate under high load, tight margins and constant dispute cases
Pickup points usually already have cameras, but they remain a passive archive. When a customer says the wrong item was handed over, a return becomes disputed, a parcel disappears from storage or the operator leaves during peak load, the owner has to search video manually and often still lacks the full event context.
Returns and substitutions
Return fraud, parcel substitution, disputed defects and ambiguous actions at the desk require contextual video.
One operator handles everything
The employee accepts deliveries, serves customers, searches for orders in storage and keeps the point in order at the same time.
Standards risks
Queues, disorder in the hall, absent staff, storage zone violations and broken procedures become a risk for the point.
Where efficiency is lost
What should be controlled at a pickup point every day
Video analytics moves surveillance from archive to operations: the system captures events, attaches frames, calculates KPIs and sends a signal to the owner, manager or security team.
Handover and intake
Order handover, courier intake, opening, package damage and item condition.
Storage integrity
Unauthorized access, missing parcels, item movement, open storage zones and night events.
Queues and flow
Visitor count, queue length, peak hours, waiting time and need for a second service line.
Point standards
Staff presence, hall order, clear passages, foreign objects, signage, cleanliness and procedures.
4 zonesdesk, hall, fitting/check area and storage
1 screenevent log, frames, statuses and KPIs
2 weekstypical pilot on part of the point cameras
no replacementconnection to existing cameras
Before
Cameras simply record video
- The archive is reviewed only after a complaint, shortage or conflict.
- Handover, intake and return events are not linked to an order.
- Queues, absent operator and disorder are noticed too late.
- The network manager cannot compare points by objective KPIs.
After
AI captures events and calculates KPIs
- Every handover, return, queue or storage risk becomes an event.
- A frame, video clip, camera, zone and time are attached to each event.
- Notifications are sent to Telegram, Email, SMS, VMS or an internal system.
- New scenarios are added as the pickup point network develops.
Platform capabilities
Core modules for controlling pickup points and networks
Modules can be enabled step by step: first the clearest losses, then storage, standards, staff, integrations and a unified network control center.
HandoverHandover and intake
Capturing handover, opening, courier intake, package damage and dispute cases.
ReturnsReturns and fraud
Parcel substitution, false defect claims, item mismatch, disputed return and evidence base.
StorageStorage and back room
Unauthorized access, parcel movement, shortages, batch intake and restricted zone control.
QueueQueue and flow
Visitor counting, queue length, waiting time, peak load and service SLA.
OperatorStaff and discipline
Presence, absences, late arrivals, phone use, work zones, activity and shift KPIs.
StandardsStandards and safety
Hall order, passages, foreign objects, smoke, fire, conflict, signage and procedures.
Order handed to customercamera 2 · handover desk · 14:32
frameHandover without confirmed scanningcamera 1 · high risk · 15:08
riskBatch intake from couriercamera 4 · intake zone · 16:21
SLAHandover deskhandover, opening, disputed item, link to order time
Intakecourier, parcels, package damage, number of packages
Handover and intake
Every disputed operation gets video evidence
The system captures key service moments: order intake, customer handover, opening, item inspection, package damage and unusual actions at the desk.
- handover and opening are linked to time, zone and camera
- quick search for the needed frame by date, point, desk or event
- capturing batch intake from courier and package condition
- control of handover without scanning or confirmed action
- evidence base for the customer, marketplace and point owner
Returns and fraud
Control of returns, substitution and disputed defects
A return at a pickup point often becomes a word-against-word dispute. AI helps preserve context: who brought the item, how it was inspected, what was in the box, whether there were signs of substitution and how the operator processed the procedure.
- recording the return and item inspection procedure
- detecting package substitution or suspicious action at the desk
- capturing damaged boxes, opening, empty package or mismatch
- event with frame and video clip for review with the marketplace
- separate scenarios for expensive, fragile and often disputed categories
Return procedure videoclip 00:44 · desk camera
videoSuspected substitutionitem removed outside camera view · risk 82%
riskProcessing statussent to manager · awaiting decision
openWaiting hallqueue, order, foreign objects, conflict
Deskhandover, return, opening, operator actions
Storageparcel movement, access, shortages, intake
Check areafitting, inspection, disputed actions, abandoned items
Storage and zones
Control of storage, intake and parcel movement
For a pickup point owner, storage often remains the least transparent area: items arrive in batches, are stored in limited space, move between shelves and are handed over under high load.
- unauthorized entry into storage or customer behind the desk
- parcel movement without an expected system action
- left, lost or long-standing parcels in passages
- checking clear passages, area overload and storage order
- separate rules for valuable, oversized and fragile orders
Queue and flow
Counting visitors, queues and peak hours
The system counts visitor flow, queue length, waiting time and operator load. The manager sees when a point needs a second service line, schedule changes or an additional employee.
- incoming and outgoing visitor flow by hour and day
- queue length at the desk and notification when threshold is exceeded
- average waiting time and service speed
- comparing network points by load, SLA and complaints
- heat zones: where customers and parcels accumulate
31visitors today
4 minaverage wait
6queue threshold
Staff
Operator work, presence and shift discipline
AI helps monitor not people for the sake of control, but the operating standard of the point: the employee is present, the service zone is covered, the queue does not grow, storage is not left unattended and violations do not repeat endlessly.
Presence on site
Capturing operator absence from the desk, long departures and an empty service line.
Working time
Late arrivals, early departures, breaks, shift activity and point comparison.
Phone and distractions
Procedure violations during customer service or batch intake.
Service speed
Time at the desk, repeated delays, peak hours and employee load.
Storage work
Routes, frequent departures, order search, access to restricted zones and performed operations.
Appearance
Uniform, badge, workplace order and customer service standards.
Conflict situations
Increased activity, crowding, unusual behavior and quick video search.
Shift training
Review of recurring violations with frames, video and clear comments.
Standards and safety
Pickup point standards, hall order and safety
The platform records the state of the point continuously, not once a week during an inspection: order in the customer area, clear passages, staff presence, foreign objects, smoke, fire, conflict and other events.
hall zonequeue, order, passages, customer experience
storage zoneaccess, intake, movement, shortages
deskhandover, return, conflict, operator action
safetysmoke, fire, left objects, vandalism
Important for 152-FZ
You can start without face recognition
The basic pickup point setup does not require biometrics: the system analyzes events, zones, actions, queues, parcels and procedures. Face recognition can be added separately only when there is a legal basis and a clear scenario.
- without mandatory collection of a customer biometric database
- on-premise for sensitive video and internal policies
- access roles, action log and archive viewing limits
- events with a frame instead of constant manual monitoring
- integration with VMS, Telegram, Email, SMS and API
Interfaces
Event log, incident cards and network dashboards
The operator sees current events, the manager sees the causes of problems, and the network owner compares points by queues, handover, returns, storage, staff and standards.
12events
3open
98%cameras online
Network dashboard
Comparison of pickup points by queues, disputed handovers, returns, storage, operator activity and standards level.
Queue above thresholdLenina 14 pickup point · camera 1
newDisputed handoverNevsky 22 pickup point · frame attached
openStorage accessArbat 7 pickup point · outside procedure
riskEvent log
Filters by point, zone, camera, event type, status, shift, employee and time.
framehandover desk · 14:32 · order found
statusunder review · owner assigned
Incident card
Evidence frame, video clip, time, camera, zone, owner and handling history.
4.1min wait
92%SLA
2storage risks
Reports and KPIs
Trends in load, dispute cases, returns, shortages, shift work and service quality.
Deployment
From one point to a unified digital control loop for the pickup network
The first stage should not try to do everything at once. It is better to run a measurable pilot on real cameras: take 3-5 scenarios, configure events, check quality and calculate the effect.
Pilot
Start with quick scenarios where the result is visible in video
For the first point, the handover desk, entrance and storage are usually enough. After the pilot, new cameras, points, roles, notifications, reports and scenarios are connected.
handover, intake and disputed desk operations
returns, substitution, package damage
queue, visitor flow and waiting time
operator absence and work zone control
storage zone, unauthorized access and shortages
events, frames, notifications and network dashboard
01
Camera and zone audit
We define the desk, entrance, storage, check area, fitting area and the point's main losses.
02
Stream connection
We take RTSP / ONVIF from existing cameras and configure zones and event thresholds.
03
Events and roles
We create the log, notifications, handling statuses, dashboards and access rules.
04
Scaling
We connect the pickup point network, new scenarios, integrations and a unified control center.
Cloud / SaaS
Fast launch, unified network analytics, notifications and reports without own server infrastructure.
On-Prem
Deployment inside the customer environment: server, roles, archive, events, reports and integrations.
Hybrid
Local processing of sensitive video plus centralized analytics, updates and model development.
AI Factory
New pickup point scenarios without a long development project
The platform supports adding new classes, events and rules: unusual packaging, new zones, specific marketplace procedures, rare conflict situations and network-specific risks.
labelingfine-tuningVLM/LLMsyntheticsrollout
Integrations
Connection to VMS, messengers and internal systems
Events and KPIs can be sent to VMS, Telegram, Email, SMS, BI, CRM, 1C, WMS, internal network management panels or marketplace process APIs.
VMSTelegramEmailSMSBIAPI
Want to test video analytics at one point?
We will select 3-5 pilot scenarios: handover, returns, queue, storage, operator or pickup point standards. After the pilot, we scale the solution to the network.