Nsi Approved Security Companies List Data: Dataset Structure and Field Coverage

It is the public Company Finder run by the National Security Inspectorate (NSI), the UK certification body that trades as Insight Certification Ltd from Sentinel House in Maidenhead. NSI describes the register as holding over 2,000 approved companies, covering the United Kingdom and - through its CerticCS schemes - the Republic of Ireland. Every entry is a firm NSI itself certifies, tagged with an approval tier (NSI Gold, NSI Silver, NSI CerticCS or SIA ACS) plus the specific scopes it is certified to deliver.

Scope families in the finder span Security Systems (intruder and hold-up alarm systems, access control systems, CCTV surveillance systems, detector activated CCTV to BS 8418, scaffolding alarms), Fire Safety Services (BAFE scopes including fire detection and alarm systems), Guarding Services (keyholding and alarm response, door supervision, close protection, security guarding, cash and valuables in transit), Alarm Monitoring (alarm receiving centres, lone worker monitoring, remote video response centres) and Specialist Services.

The finder is a GET form on nsi.org.uk/company-finder/, and query parameters drive server-rendered results. A call such as ?at=NSI+Gold&search=company&company-name=alarm&perpage=5 returns matching firms directly in HTML - verified during research. The at parameter selects the approval type, company-name takes free text, and location searches post Lo and La longitude-latitude values plus a range in miles.

Region filtering is unusually deep: the finder covers 49 UK police forces plus separate guarding and fire regions, so you can pull every firm whose certified scopes claim response coverage in your force area. Results paginate at 10, 50, 100 or all records, sort randomly or A to Z, render in a map view, support favourites cookies and printing, and can go through a server-side PDF export of the result set.

Two practical limits: Google reCAPTCHA protects parts of the flow, so polite rate limiting is essential, and there is no official bulk export. One sampled record shows the granularity on offer - BHP Alarms Ltd in Londonderry (BT48 0LY), approval type NSI CerticCS, holding both CerticCS ARC and CerticCS Electronic Security schemes.

Each record is a structured object rather than a flat card: CompanyID (a GUID from NSI's register), CompanyName, full postal components (AddressStreetOne and AddressStreetTwo, AddressTown, AddressCounty, AddressRegion, AddressPostCode, AddressCountryCode), email, phone and website, geocoded Latitude and Longitude feeding the map view and distance search, the ApprovalType enum, a Schemes array - NACOSS Gold, Systems Silver, CerticCS ARC, CerticCS Electronic Security are examples - and ApprovedServicesOffered.

ApprovedServicesOffered lists the certified scopes themselves, such as design, installation and maintenance of intruder and hold-up alarm systems, each carrying a ResponseRegionGroup and the police regions covered. That field is what makes the register usable for territory work: you can see not just whether a firm installs alarms, but which force areas its certified scopes claim.

One structural difference matters before you automate anything against them. NSI serves server-rendered HTML with a PDF export path; SSAIB renders results client-side with mustache-style placeholders such as Installer_Name, and its deeper portal requires selecting a certification scheme and solving a captcha before any records return; FDACS serves a Next.js shell whose body text only appears after JavaScript rendering. Same procurement question, three different collection realities - which is why the quality scores differ even though all three are free to query.

Filter by scope before name. Match the scope family to the job first - an intruder-alarm install maps to Security Systems scopes, a monitored account to Alarm Monitoring scopes such as alarm receiving centres - then constrain by postcode distance and check each candidate's ApprovedServicesOffered for response coverage in your police-force region. Only then compare trading names, because two local firms can hold very different scheme strings.

Size the risk against real exposure. data.police.uk publishes monthly street-level crime, outcomes and stop-and-search for 44 forces; the National Crime Victimization Survey interviews roughly 240,000 people in about 150,000 households each year, capturing burglaries never reported to police; and for US metros the LA Crime Data set carries 1,004,894 block-level geocoded LAPD incidents covering 2020 through 2024. Matching incident density to certified-scope coverage turns a logo check into an evidence-based shortlist.