FeaturedOpportunity Radar20,000+ opportunities tracked worldwide, filtered for your expertise.One-time report — $29 $9.90 →

SiteAssist

Digital control-of-works platform for permits, inspections, and safety on high-risk construction sites

Cloud and mobile platform that digitizes permits, inspections, task tracking, and safety registers for high-risk construction, utilities, nuclear, and defense projects, using computer vision to verify site photos and flag missing safety equipment.

Bootstrapped - no external funding round publicly disclosed; Tracxn lists the company as unfunded (checked 2026-07-16); backers noted only as unnamed "construction industry insiders" per Wellfound raisedGrowth Stage
Founded 2020 · ~12,000 active users across ~4,000 devices as of a 2026 Ultralytics case study; named customers include Balfour Beatty, Taylor Woodrow (VINCI), Skanska, and HG Construction. Company itself is small (Tracxn: 18 employees as of Jan 2025; LinkedIn: 11-50 employees) and has raised no external funding, so this is a small, bootstrapped team serving a handful of very large UK contractors at scale rather than a broad customer base - worth noting as a contrast to the user-count figure alone.
Website →← All Tools

Why This Tool Exists

The Problem

High-risk construction, utilities, and nuclear/defense projects still run permits, safety inspections, and toolbox briefings on paper forms or spreadsheets, so site managers cannot see in real time whether a hot-work or confined-space permit has an unresolved hold point, whether a required safety photo was genuinely taken on site, or whether a subcontractor's RAMS document has actually been checked against the applicable standard. Reviewing this paperwork and the accompanying site photos manually is slow, inconsistent between reviewers, and leaves a weak audit trail if an incident later needs investigating.

The Solution

SiteAssist replaces the paper process with an offline-capable mobile app and web dashboard covering digital permits with geolocation and hold-point logic, structured inspections and checklists, task/action tracking, and registers for permits, briefings, temporary works, and equipment. A computer-vision layer built on Ultralytics YOLO models (including YOLO26) fine-tuned on SiteAssist's own construction imagery scans uploaded site photos to flag likely non-genuine images (e.g., a photo of a screen) and to count required safety items such as fire extinguishers, while a separate "AI Document Intelligence" feature reviews uploaded RAMS and other site documents against configured safety requirements. A human reviewer confirms or overrides every AI flag before it affects a permit or inspection outcome.

How You Use It

Delivery Method
SaaSWeb AppMobile App (iOS, Android)
Integrations
Not publicly specified - no third-party software integrations (e.g., Procore, Autodesk Construction Cloud) documented on the public site or app store listings
Disciplines
Project Phases
Construction Administration
Project Types
Infrastructure, Industrial, Commercial

Data Transparency

Exactly what this tool uses and how

Input
What it needs
Required:Site photos (permit/inspection evidence), Permit request details (work type, location, hazards), RAMS or other safety documents for AI Document Intelligence review
Optional:Equipment/asset register data, Toolbox talk/briefing records, Project schedule or site plan for geolocation context
Formats:JPG/PNG (photos), PDF (RAMS and safety documents)
Output
What you get
Format:Digital permit/inspection records with AI-flagged findings, surfaced through a centralized analytics dashboard
Fields:Permit status and hold points, Inspection/checklist results and assigned actions, Detected objects/hazards in site photos (e.g., fire extinguisher count, personnel-machinery proximity), Document compliance flags from AI Document Intelligence, Audit trail of approvals and sign-offs
Algorithm
How it works
Model:Ultralytics YOLO object-detection/classification models (including YOLO26) fine-tuned on SiteAssist's own construction photo dataset for hazard/PPE detection and image-authenticity checks; a separate proprietary "AI Document Intelligence" feature reviews text documents, architecture not publicly disclosed
Accuracy:Not publicly specified as a precision/recall figure. Ultralytics case study reports usage scale since January 2025 of 770,918+ images processed and 1,302,315+ objects detected (~1.7 objects/image, ~2.7 excluding background-only images), but no detection accuracy or false-positive rate is published.
Privacy
How your data is protected
Retention:Not publicly specified - privacy policy states Personal Data is retained "only for as long as necessary for the purposes for which it was collected, unless a longer retention period is required by applicable law," with no fixed duration published
Training:Not publicly specified whether customer photos or documents are used to further train SiteAssist's computer-vision or document-intelligence models; the privacy policy does not address AI/model training use of customer data
Compliance:ISO 27001:2022, Cyber Essentials (UK NCSC scheme), GDPR (referenced in privacy policy for EU/UK personal data), No SOC 2 report or certification found on public sources
API
Integration
Endpoint:Not publicly documented - no developer/API documentation page found on siteassist.com or app.siteassist.co
Method:Not publicly specified

Use Cases

  • ·Digital permit-to-work management (hot work, confined space, lifting, excavation) with hold points and geolocation
  • ·Structured safety inspections and checklists with photo evidence and audit trail
  • ·AI-assisted verification that submitted site photos are genuine and show required safety equipment (e.g., fire extinguisher counts)
  • ·Automated first-pass review of subcontractor RAMS and safety documents against site requirements
  • ·Centralized permit, briefing, temporary-works, and equipment registers across a multi-contractor project

Pricing

Free
Not offered; demo available on request via siteassist.com/contact-us
Pro
Not publicly specified - no self-serve pricing tiers published
Enterprise
Custom pricing; contact vendor (per-project/per-seat pricing model not publicly disclosed)
Website →
Research Sources & Data QualityLast verified: 2026-07-16
Verified Data (14)
problem, solution, deliveryMethod, yearFounded, vendor, website, domain, fundingTotal, customerBase, compliance-iso27001, compliance-cyberessentials, algorithm-approach, legal-entity, founders
Not Found (7)
pricing-tiers, api-documentation, integrations, data-retention-period, ai-training-data-policy, accuracy-benchmark-precision-recall, named-investors-or-backers
Our Commitment: We only include verified data from official sources. If information isn't publicly available, we mark it as "Not publicly specified" rather than guessing.