Flow without context
A basic people counter does not show who was a customer, where they moved, whether they reached the checkout and how this relates to sales.
AI video analytics platform for retail and stores
A single environment for customer counting, shopper-flow analytics, queues, checkouts, self-checkout, shelves, merchandising, staff, losses and retail-network KPIs.
customersentries, exits, unique visits, repeat visits and staff exclusion
conversionlinking video streams with receipts, POS and actual sales
shelvesout-of-stock, empty spaces, planogram, facings and promo areas
queuesqueue length, waiting time and recommendation to open a checkout
checkoutsnon-scan, returns, discounts, suspicious operations and self-checkout
BI / KPIratings for stores, shifts, zones, categories and regions
Context
A store needs managed metrics, not just cameras
CCTV is already installed on sales floors, at entrances, near checkouts and in service areas, but it often works only as an archive. AI turns existing cameras into a source of business metrics: how many customers came in, where they stayed, why they did not buy, where queues grow, which shelves are empty and which operations need review.
Lost sales
A customer cannot find a product, sees an empty shelf, leaves because of a queue or does not receive help from a consultant.
Merchandising and promo
Planograms, facings, price tags, promo materials and shelf space are checked manually and irregularly.
Losses and risks
Theft, non-scan at self-checkouts, suspicious returns and internal violations require fast review.
What changes with AI
From people counting to customer-journey analytics
The platform links video events with POS, staff schedules, planograms, loyalty programs and BI. The store receives not just an entrance counter, but a managed model for traffic, conversion, queues, shelves, service quality and losses in one environment.
Customer flow
Entries, exits, unique customers, repeat visits, interest zones, routes and dwell time.
Conversion to purchase
Matching traffic with receipts, categories, checkouts, promos and lost customers.
Product availability
Empty spaces, low-stock, out-of-stock, merchandising errors, promo zones and replenishment priorities.
Network management
Ratings for stores, shifts, regions, categories, checkouts and zones on the executive dashboard.
8 moduleyflow, conversion, shelves, checkouts, queues, staff, losses, BI
1 environmentvideo, events, POS, reports, roli and notifications
10-40flowov for pilota in 1-3 stores
4-6 nedelproverka stsenariev and raschet ekonomiki tirazhirovaniya
Before
Kamery and reports zhivut otdelno
- traffic calculatessya at entrance, but not svyazan with sales and shelfmi.
- Queues, out-of-stock situations and staff absence are visible only during a manual walk-through.
- investigations kassovykh losses zanimayut hours manual viewing video.
- Tsentralnyy ofis not sees edinuyu picture by storem and regionam.
After
AI records events and calculates KPI
- Each store receives customer-journey, queue, shelf and checkout metrics.
- Events enter the log with a frame, video fragment, zone and status.
- Supervayzer sees rating toreceipt, shifts, categories and povtoryayuschikhsya problem.
- New scenarios are added without replacing existing cameras.
Platform capabilities
AI modules for stores and retail chains
Scenarios can be launched step by step: customer counting, queues and basic events first, then shelf analytics, checkout risks, staff work and BI reporting across the network.
Customers
Customer counting
Entries, exits, unique visits, repeat visits, staff exclusion and connection with conversion.
Journey
Customer journey
Routes, heat maps, interest zones, dwell time and the entrance-shelf-checkout funnel.
Shelves
Shelves and merchandising
Out-of-stock, low-stock, gaps, planogram, promo, price tags, facings and replenishment priorities.
Checkouts
Queues and checkouts
Waiting time, recommendation to open a checkout, self-checkout, non-scan and POS linkage.
Staff
Staff and service
Presence on the floor, customer response, merchandising, uniform, downtime, work zones and standards.
Losses
Loss prevention
Suspicious actions, theft, returns, service zones, conflicts and investigations.
1 284customer voshli
42%conversion in receipt
18 minsrednee time in floor
Pikovyy flow at entrance13:00-14:00 · need to usilit checkouts and zal
trafficSnizhenie konversiitraffic grows, receipts menshe prognoza
risk
not just lyudi
Counting customers how business-metrika store
System calculates not abstraktnykh people, and customer flow: otdelyaet incoming and outgoing, isklyuchaet employees, ubiraet dubli, analiziruet repeat visits and links traffic with actual sales.
- counting customers at entrance, exit, in queues, otdele and promo-zone
- unique and repeat ponetworkteli without obyazatelnoy biometrii
- exclusion employees, okhrany, merchandisers and tekhnicheskogo staff
- conversion vkhodyaschego flowa in pokupku by POS and kolichestvu receipts
- pochasovye, dnevnye, sezonesnye and regionalnye reports by network
Pokupatelskiy put
routes, heat maps and interes to zonem sales
Video analytics shows, chto proiskhodit mezhdu vkhodom and kassoy: where customer zaderzhalsya, where proshel mimo, which zones peregruzheny, which promo not rabotayut and how menyaetsya behavior after perestanovki product.
- heat maps of aisles, showcases, islands, promo stands and checkouts
- dwell time: how long customers spend at a category, showcase or shelf
- funnel: entrance - zone - shelf - vzaimodeystvie - checkout - receipt
- and/B-testy planirovki, POSM, promo and tovarnykh razmescheniy
- comparison stores raznykh formatov and razmerov
entranceflow, repeat visits, novye customers, peak nagruzki
promopodkhody, ostanovki, interes to vykladke, funnel to receipta
shelvesdwell time, vzaimodeystvie, pustoty, errors merchandising
checkoutsqueue, ozhidanie, self-checkout, neshtatnye operations
01cameras vidyat shelf by raspisaniyu or online
02AI finds pustoty, low-stock and errors merchandising
03sobytie popadaet in zadachu employeeu
04proveryaetsya popolnenie and response time
05the category team receives a lost-sales report
06network sees rating stores and postavschikov
Pustoe mesto at shelfmolochnaya kategoriya · vysokiy oborot · 17 minut
oosplanogram narushenaSKU is out of place · promo zone · review required
plan
Shelves and merchandising
Control nalichiya, planogrammy and promo-ispolneniya
AI helps see store glazami customer: est whether product at shelf, not pustuet whether feysing, pravilno whether oformlena promo-zone, not pereputany whether products and naskolko quickly employeei zakryvayut zadachu replenishment.
- out-of-stock, low-stock, empty spaces and dlitelnost otsutstviya product
- control planogrammy, feysingov, doli shelves and promo-merchandising
- search product not at svoem meste and smeshannoy merchandising
- control tsennikov, POSM, stopperov and promo-materialov
- tasks employeeam and SLA replenishment polok
Queues, checkouts and POS
Management ozhidaniem, checkouts and kassovymi riskami
The system records queue length, waiting time, checkout load, cashier work and self-checkout events. With POS integration, returns, cancellations, discounts, unscanned goods and disputed operations can be reviewed.
- queue at kassu, self-checkout, servicenuyu stoyku and vydachu online-zakazov
- rekomendatsiya otkryt dopolnitelnuyu kassu or pereraspredelit employeea
- non-scan, product in korzine without operations, ukhod from checkouts, oshibka self-service
- returns, cancellations, discounts, manual price entry and receipt anomalies
- link events with kassoy, shiftsoy, receiptom, kadrom and videosegmentom
06:18ozhidanie in peak
5checkouts open
12POS events
Ochered vyshe porogacheckouts 1-4 · 9 customers · otkryt liniyu
SLAproduct not proskanirovanself-checkout #3 · confirmation required
scanPodozritelnyy vozvratcheckout 2 · receipt 18429 · videosegment
risk
Staff and service
presence employees and service quality
The platform helps monitor the operating standard, not a person for the sake of control: the consultant is available, the checkout is open, merchandising is replenished, the client does not wait without attention, and the store zone has an owner.
- employee presence on the floor, near the checkout, showcase, fitting room or service counter
- response time to a customer near a showcase, shelf or consultation zone
- downtime, otlucheniya, peregruzka shiftsy and neravnomernaya zagruzka
- uniform, beydzh, telefon, distsiplina and vypolnenie obkhodov
- tasks at popolnenie, uborku, telezhki, korziny and vosstanovlenie poryadka
Loss prevention
Snizhenie losses and bystrye investigations
Video analytics sokraschaet ruchnoy review archivea: sobytie already linked with camerasoy, zone, receiptom, vremenem, statusom and videosegmentom. Sluzhba security faster finds povtoryayuschiesya scheme and problemnye tochki.
- suspicious behavior in torgovom floor, at shelves and at kasse
- product removal through the entrance, service zone, warehouse or receiving area
- left objects, conflicts, falls, vandalizm, smoke and vozgoranie
- unauthorized access to the warehouse, office, ramp or service corridor
- kartochki rassledovaniy, statuses obrabotki and eksport dokazatelnoy bazy
Karta keysov
Bolshe stsenariev for retail network
for retail important sobrat not one model, and biblioteku stsenariev, kotoruyu can vklyuchat by storem, formatam, zonem, cameras and prioritetam biznesa.
customers at entrance
entrance, exit, pikovye hours, novye and repeat visits.
exclusion staff
Sotrudniki, okhrana and merchandayzery not iskazhayut traffic customers.
Konversiya in receipt
matching visitors with POS, kolichestvom receipts and vyruchkoy.
Poteryannyy customer
Chelovek voshel, provel time in floor, but not reached to checkouts.
Teplovye karty
Zony vnimaniya, slabye prokhody, peregruzhennye tochki and promo-interes.
time at shelves
How long a customer studies a category, showcase or promo stand.
Out-of-stock
Pustoe mesto, dlitelnost otsutstviya product and prioritet replenishment.
Low-stock
product zakanchivaetsya to polnogo otsutstviya at shelf.
planogram
Control merchandising, feysingov, doli shelves and tovarnykh blokov.
Tsenniki and POSM
Control tsennikov, promo-materialov, stopperov and aktsionnykh zones.
product not at meste
Pereputannye pozitsii, smeshannaya merchandising and errors merchandayzinga.
Ochered at kasse
Dlina queues, waiting time, SLA and signal otkryt kassu.
Self-checkout
Non-scan, ukhod without zaversheniya operations, zavisanie at terminala.
Vozvraty and discounts
Search receiptsykh anomaliy and link with videosegmentom.
Prodavets in zone
presence consultant, response time and help pokupatelyu.
Primerochnye
Ochered, dlitelnoe nakhozhdenie, vozvrat product and poryadok in zone.
Korziny and telezhki
Nalichie at entrance, perepolnenie zones and task employeeu.
Receiving and warehouse
Dostup, peremeschenie product, rampa, boxes, nedostachi and reglamenty.
suspicious actions
Nestandartnoe behavior at shelves, checkouts, vykhoda or sluzhebnoy zones.
Bezopasnost
Dym, fire, left objects, falls, conflicts and crowding.
Store formats
Keysy for raznykh tipov retail
Odin and tot zhe environment can adaptirovat under produktovuyu network, apteku, fashion, elektroniku, DIY, store at doma, gipermarket or punkt vydachi online-zakazov inside store.
Potok, shelves, queues and self-checkout
Maximum value comes from linking traffic, product availability, checkout speed and non-scan control.
- empty shelves and popolnenie
- queues in hours peak
- self-checkouts
Minimalnyy shtat and bystrye events
Control prisutstviya employeea, checkouts, entrance, conflicts, merchandising and zakrytiya shiftsy.
- employee at meste
- korotkie queues
- incidents at entrance
Queues, consultations and showcases
Analytics helps monitor pharmacist workload, queues, showcase availability and service quality.
- waiting time
- konsultatsionnaya zone
- showcase control
Put customer and primerochnye
Heat maps, interest in collections, fitting rooms, consultations and zone-to-receipt conversion are important.
- primerochnye and queues
- interes to kollektsiyam
- sales assistant-konsultant
Vitriny, konsultanty and dorogoy product
The system records interest in a showcase, consultant response time, access to demo zones and theft risks.
- showcase zones
- response consultant
- dorogoy product
Bolshie zaly and tyazhelye kategorii
Analytics helps allocate staff, see overloaded departments, empty spaces and warehouse movements.
- routes by otdelam
- consultations
- warehouse and ramp
Polochnoe prostranstvo and promo
Control of merchandising, price tags, promo islands, product availability and recurring network errors.
- planogram
- promo-ispolnenie
- empty spaces
Zony, checkouts and network KPI
Bolshie formaty require analitiki by otdelam, vkhodam, checkoutm, ocheredyam, shelfm and staffu.
- many entrances and checkouts
- heat maps
- rating otdelov
Vydacha zakazov inside store
Control ocheredey, employeea, zones khraneniya, vydachi, vozvrata and spornykh situatsiy with customers.
- queue vydachi
- employee in zone
- dokazatelnye frames
Interfaces
Executive dashboard, event log and reports by network
operational komanda sees events and tasks, security team rassleduet incidents, categoriesnyy menedzher monitors shelves, and tsentralnyy ofis sravnivaet storey and regiony by edinym KPI.
42%conversion
7.8%OOS risk
04:12ozhidanie
Executive dashboard
traffic, conversion, queues, out-of-stock, losses, service, tasks and rating stores at odnom ekrane.
Pustaya shelfstore #24 · molochnaya kategoriya · 17 minut
oosOchered vyshe porogacheckout 3 · 8 customers · SLA risk
SLANon-scanself-checkout #2 · proverka employeeom
scanZhurnal events
Filtry by storeu, zone, camerase, tipu events, kasse, employeeu, kategorii, statusu and time.
kadrpodtverzhdenie events and zone at izobrazhenii
SLAowner, status, kommentariy and response time
Kartochka events
Podtverzhdayuschiy kadr, videofragment, zone, receipt or shelf, status, owner and istoriya obrabotki.
#12luchshaya conversion
#7rost queues
#3OOS risk
Reyting stores
Sravnenie toreceipt, regions, shifts, otdelov and categories by business-metrikam and povtoryayuschimsya problemam.
Implementation
Pilot in 1-3 stores and scaling at network
A rational start is to take scenarios that quickly show economics: customer counting, queues, shelves, checkouts and one or two loss-prevention events. After stabilization, shelf analytics, staff, POS anomalies and network BI are added.
Pilot
Nachinaem with izmerimykh stsenariev and realnykh video streams
For the first stage, camera quality, visibility of entrances, checkouts, shelves and zones, POS access, clear event thresholds and success criteria are important.
counting customers and exclusion staff
queues, ozhidanie and signal otkryt kassu
out-of-stock and tasks at popolnenie
self-checkout and kassovye events
event log with kadrom and videofragmentom
ekonomika tirazhirovaniya at network stores
01
Obsledovanie
Kamery, rakursy, zones, POS, planogrammy, format store and requirements khraneniya.
02
MVP
connection flowov, basic events, zones, zhurnal, reports and pervye pravila reaktsii.
03
Pilot
Doobuchenie, nastroyka porogov, otsenka tochnosti, lozhnykh srabatyvaniy and business-effekta.
04
Tirazhirovanie
Integratsii, roli, BI, dostavka modeley at storey and unified control center networkyu.
On-Prem
Video, events, archive and integrations ostayutsya inside infrastructure customer.
Hybrid
Lokalnaya obrabotka video plyus tsentralizovannaya analytics, obnovleniya modeley and reports by network.
Cloud
Bystryy start pilota, esli politika kompanii pozvolyaet obrabatyvat flowi in oblake.
AI Factory
Novye scenarios without proekta with nulya
Platform podderzhivaet biblioteku modeley, fine-tuning under konkretnuyu network, VLM/GenAI for novykh events, but-code pravila, versionirovanie and dostavku modeley at storey.
labelingfine-tuningVLM/LLMrulesrollout
Integratsii
link with systemmi retail network
Sobytiya and KPI can peredavat in VMS/NVR/CCTV, POS, ERP, WMS, CRM, loyalty, task manager, service desk, BI/DWH, LDAP/SSO and vnutrennie API.
VMSPOSERPWMSCRMBISSO
Legal framework
Bazovaya analytics works without biometrii
Customer counting, queues, shelves, checkouts, staff and security events do not require identifying the customer. Face recognition can be connected separately only when there is a legal basis and a clear scenario.
Russian software
Razvertyvanie in zakrytom environment
STATANLY can be deployed on-premise or hybrid, connected to existing cameras and integrated with internal retail systems without transferring sensitive video to an external perimeter.
Khotite proverit videoanalitiku at 1-3 stores?
We will agree on a fast pilot: customer counting, queues, shelves, checkouts, self-checkout, staff or loss prevention. We will test value on real video streams and POS data.