How to choose a video analytics platform (India buyer’s guide)
· Updated · Slinai
Key facts
| Non-negotiable | Works on existing ONVIF/RTSP CCTV — no camera rip-and-replace to enable AI |
|---|---|
| Slinai Core list price | ₹1,499 per camera per month after a 1-camera free month — see Pricing for Ops+ / Business |
| TCO claim | About 50% lower entry TCO vs typical US-centric alternatives with high floors or per-feature camera fees — verify in your quote |
| Security | Encryption, RBAC, audit logs, privacy-safe analytics, edge/on-prem residency options |
| Pilot | 1–2 weeks on your footage; success = clips staff act on, not a slide |
Buying video analytics in India is not the same RFP as buying a US cloud VMS. The cameras are already on the wall. The duty manager lives on WhatsApp. Legal will ask where the video goes. Finance will ask why a “smart camera” refresh is in the same deck as software.
This guide is the commercial close to the three pillars — what AI video analytics is, agentic, existing CCTV — and the FAQ. Score vendors on the same sheet, then compare named alternatives.
The scorecard (print this)
| Criterion | What “good” looks like | Fail |
|---|---|---|
| Existing CCTV | RTSP/ONVIF, NVR stays, optional gateway only | New proprietary cameras to unlock SKUs |
| Job fit | Detections you named (queue, SOP, pick SLA, LP) | Generic “smart city” reel |
| Alerts | Verified clip to WhatsApp/Slack/Teams | Email dump or dashboard-only |
| False positives | Verification step; UAT noise log | 80 events a night, nobody listens |
| Language & support | English + Hindi ops; India hours | US-only CSM, 2 a.m. Slack |
| TCO | Per-camera all-in Core; no ₹50k floor surprise | Per-feature camera fees, hidden storage |
| Residency | Edge/hybrid documented (security) | “It’s in our US cloud, trust us” |
| Join | POS/OMS when the job needs it (data) | “Conversion” from pixels alone |
| Agentic | Grounded Q&A + standing rules | Chatbot that invents counts |
Slinai’s product is scored the same way: VMS + Video AI + edge + data + alarms.
Price without lying to the CFO
Read pricing as the list, then model your camera count.
Slinai publishes:
- Free: 1 camera × 1 month.
- Core: ₹1,499 / camera / month — queue, staff, intrusion, perimeter, gate footfall, health, and the rest of Core (not a ₹2,000-per-feature meter).
- Ops+ / Business: ₹2,999 and ₹5,999 / camera / month.
- On-prem / edge-only: sales quote.
The site claim is about 50% lower entry TCO than typical US-centric alternatives with high monthly floors or per-feature camera pricing. Put your quote next to compare (Spot AI, Verkada, Rhombus, Eagle Eye, BriefCam). Do not use a blog number as a contractual discount.
Security and DPDP are RFP chapters, not PDFs in the appendix
Security & privacy should answer, in writing:
- Encryption in transit and at rest.
- RBAC and audit logs (who exported a guest-area clip).
- No facial recognition by default; behaviour and aggregate bands.
- Where recordings live (NVR vs cloud retention).
- Edge when “video never leaves this site” is a real sentence.
If the vendor cannot draw the data flow on a whiteboard, they cannot survive your counsel.
Pilot design that survives a steering committee
- One job — retail queues, restaurant queues, pick-pack SLA, or LP.
- Your streams — existing CCTV, not a hotel lobby in California.
- Named tests — alert arrives; false-positive count; one Hindi or English question if you want agentic.
- 1–2 weeks — enough traffic, not a two-day theatre.
- Decision memo — keep / expand / kill. No “let’s add 200 cameras and see.”
Book a demo is the start of that pilot, not a product tour.
Industry pages vs this guide
This post is how to buy. How to operate is:
- Retail and the footfall / conversion / queue / LP cluster.
- Restaurants & QSR
- Dark stores
- Industries index
Use FAQ for the short answers your steering pack will copy.
Next reads
Common questions
Should we buy cameras and analytics from the same vendor?
Only if you were replacing cameras anyway. Forcing a new camera family to ‘unlock AI’ is usually a TCO trap. Prefer ONVIF/RTSP on the fleet you have.
Is cloud mandatory?
No. Many Indian sites need edge or hybrid for WAN or residency. A platform that is cloud-only will fail a plant or a hill-station store even if the demo looked fine in Bengaluru.
What accuracy number should be in the RFP?
Require UAT on your footage for named detections (footfall, queue, SOP, or LP). Brochure percentages are starting points. Slinai cites 99.25% on priority detections on customer footage — still validate.
How do we compare US-centric vendors?
Score WhatsApp delivery, Hindi/English ops, India support hours, GST invoicing, and ~entry TCO. Use the compare pages for Spot AI, Verkada, Rhombus, Eagle Eye, and BriefCam.
What is a fair pilot?
One or two sites, 1–2 weeks, your streams, named success tests (alert path, false positives, one POS join if conversion or cash matters).
Related pages
Guides in this cluster point to live product, industry and use-case pages.
Related guides
Restaurant queue management: host stand to takeaway
Restaurant queue management on existing CCTV: time host-stand parties, billing and takeaway counter waits, and catch walkaways before guests leave.
Cold storage ingredient compliance in cloud kitchens
Cloud kitchen cold-storage analytics flags door-open dwell and ingredient pulls that sit in the warm aisle — process evidence beside a logger, not a replacement for it.
Cloud kitchen complaint forensics from KOT timestamps
Cloud kitchen complaint forensics ties a refunded order ID to the right prep-camera window using KOT and POS timestamps, not bag logos.