Surveily

What it does

Surveily sells software that applies computer vision to video from a site's existing security cameras ^[1][2]. The company says it detects missing PPE, unsafe worker behavior, vehicle risks such as speeding and collisions, restricted area entries, housekeeping issues, left objects and man-down events, and sends alerts by email, app notification or SMS with an image sequence of the event ^[1][4]. Rules can be set per facility and per camera, with dashboards for reporting ^[2]. Surveily also says the system can trigger warning lights, alarms and other connected devices, and that an API passes data to EHS and business intelligence tools ^[13].

Who it is for

Surveily says it works mostly with high-risk industries, listing chemicals, oil and gas, mining, food and beverage, warehousing, logistics, retail, electricity and automotive ^[4]. Its published case studies cover a refinery, a steel mill, a chemical plant, a heat and power plant and logistics distribution centers ^[1]. The legal entity is Surveily sp. z o.o., registered in Wrocław, Poland, in February 2019, with Wojciech Tubek (CEO) and Wojciech Turowicz (CTO) on the management board ^[5][6][7].

How it is deployed

Surveily says it connects to most IP cameras that support RTSP streaming and H.264 encoding, with 720p as the minimum and 1080p recommended ^[4]. Its stated camera compatibility varies across its own pages (90%, 95% and 98%) ^[2][4][5]. It offers on-premises processing on NVIDIA hardware, cloud processing on Microsoft Azure or AWS, a portable edge unit for remote sites, and offline analysis of recorded footage ^[2]. Its security page says all video streams are processed locally, which buyers should reconcile with the cloud option ^[2][3]. Surveily says it does not use facial recognition, blurs faces by default and offers head and full-body anonymization, and that it uses AES-256 encryption at rest and TLS in transit ^[3]. The same page says ISO and SOC 2 certifications were expected in Q3 2024; we found no public confirmation that either was obtained ^[3]. The FAQ says the company aims for "+95% accuracy" in deployments but does not explain how that is measured ^[4].

Notable public customers and coverage

In March 2026 Surveily raised a €2.5 million Series A led by Momenta, with Look AI Ventures participating and a second close expected from a Polish fund ^[8][9]. It had raised a €1 million seed round in 2022 from Kogito Ventures, Satus Starter, Black Pearls VC and angel investors ^[10]. Surveily names PKN Orlen as a customer and says a refinery asphalt loading site saw a 75% reduction in workplace incidents within 16 weeks and 95% PPE compliance after six months ^[11]. An unnamed logistics provider reportedly saw 62% fewer safety incidents ^[12]. These are vendor-reported figures that have not been independently verified. Verdantix lists Surveily among the organizations in its March 2026 Smart Innovators report on video analytics for safety, and keeps a vendor profile on it ^[14][15]. We found no reviews on G2, Capterra or AWS Marketplace.

Things to ask in an evaluation

• Which ISO and SOC 2 certifications are now held, and can you see the certificate or report?
• In cloud deployments, what video or metadata leaves the site, and where is it stored?
• How is the "+95% accuracy" target measured, and what false alert rate should you expect per camera?
• How were the Orlen incident figures defined, and what was the baseline?
• Which EHS, VMS or maintenance systems have a working integration today?
• What edge hardware is needed per number of cameras, and who maintains it?

Our take

Strengths

• Several deployment options, including on premises, cloud, portable edge and offline analysis of recorded video ^[2].
• Privacy features are documented in detail: no facial recognition, default face blurring and full-body anonymization ^[3].
• A clear legal entity, named leadership and a 2026 Series A led by Momenta ^[6][7][8].

Watch-outs

• Certification status is unclear: ISO and SOC 2 were "expected in Q3 2024" with no public confirmation since ^[3].
• Outcome figures are vendor-reported, and only one case study (Orlen) names the customer ^[11][12].
• No published pricing, DPA, accuracy methodology or third-party user reviews ^[4].

Best fit: European heavy industry sites, such as refineries, chemical plants and steel mills, that want CCTV analytics processed on premises and can run a pilot to test accuracy and alert volume ^[2][4][11].