AI SaaS platform

case study

SaaS Platform SoterAI

An AI platform for automatic risk detection, incident management, and ensuring compliance with safety regulations
and internal procedures.

The development of SoterAI followed a structured process: we began with in-depth research, identified key problems, formulated a product hypothesis, and then designed solutions that directly addressed those challenges.

Tup here to explore - UX/UI solution

Problem

Every successful product starts with identifying the problem, as it defines the true value of the solution.

Understanding user pain points helps set priorities, validate hypotheses, and reduce risks.

Managing industrial and corporate risks requires enormous resources:

⚠️ 27% of specialists’ working time is spent on routine tasks (document processing, compliance checks, manual reports).
⚠️ Existing AI tools produce inaccurate answers in 38% of cases if they don’t take into account corporate context.
⚠️ For large corporations with tens of thousands of employees, a scalable administration system is critical — but most solutions fail to provide it.

This leads to: higher costs, process errors, low trust in digital tools, and slow adoption of innovations.

Research

To shape the product hypothesis, I carried out several types of research.

First, I conducted interviews and surveys with OSHA experts and risk managers. OSHA experts are occupational safety specialists who deeply understand the requirements of the U.S. regulatory authority — OSHA (Occupational Safety and Health Administration). Their insights helped validate the problems faced by real professionals.

Second, I ran a competitor analysis. Most existing solutions turned out to be either too narrow in focus (standalone AI assistants) or overly complex to use. For example:

Solutions like Invigilo, Rapid Monitor, and Protex AI concentrate mainly on camera-based hazard detection, while SoterAI also works with documents and automates reporting.
Arvist AI focuses on warehouses and logistics, whereas SoterAI is universal and suitable for any type of production.
DeepEyes is limited to GMP environments, while SoterAI covers all industries.
Kognition AI works only with cameras, but SoterAI goes further with regulatory analysis, predictive insights, and process automation.
SafetyIQ emphasizes training and predictions, while SoterAI is more practical — it automates inspections, reports, and investigations.

Finally, I mapped the user journey using existing Soter products (like the safety manager dashboard). This revealed critical bottlenecks: manual data entry,

slow approval processes, and low report accuracy.

Product Hypothesis

Build an AI platform that:

Automates routine risk management processes,
Uses company context for accurate responses,
Provides multi-level administration for global corporations,
Reduces implementation costs and time.

Design Solutions & Impact Matrix

Task	Solution	Result	UX Value
Corporate management (45k+ employees)	Multi-level administration system (hierarchy, roles, tags)	Global admin manages 10+ regions without errors	Scalability, transparency, cost reduction
Increase AI accuracy	Contextual AI trained on corporate data + proprietary AI assistant	Accuracy ↑ to 97% vs. <35% when testing documentation in standard chats	User trust, reduced risk of wrong decisions
Risk analysis automation	Risk Register — interactive visual table	Automatic risk detection	Transparency, control
Lower entry barrier	Onboarding screen with chat assistant	Users started via chat (CTA) without issues	Low entry barrier, higher engagement
Reduce routine workload	Workflows (batch processing, autogeneration)	Document creation time 5–7 min (vs. 20–25 min before SoterAI)	Time savings, reduced cognitive load
Task centralization	Action Dashboard (similar to Jira inside the product)	Increased team engagement → scalable adoption	Convenience, productivity boost
AI personalization	AI Preferences (tone, format, priorities)	User satisfaction +87%, less manual corrections	Personalized experience, faster workflows

Results

We work hard every day to make life of our clients better and happier

Time-to-Task

Routine tasks that used to take 20–25 minutes are now completed in just 5–7 minutes. That’s a 3–4x acceleration, freeing specialists’ time
for more valuable work.
AI response accuracy

High-performance AI delivering 97% accuracy, reducing errors in decision-making and boosting user trust across the organization.
ROI

The system pays for itself within 9–12 months, while the industry average is around 18 months. Faster ROI makes adoption more attractive for large enterprises.

Early Achievements

First enterprise client — Marsh (45,000 employees worldwide). Key factor: multi-level administration system. Read more here

Adoption by existing Soter client base through SoterAI integration

Despite focus on enterprise subscriptions, individual packages also show steady demand with 2–3% monthly growth.

SoterAI has evolved beyond a simple AI assistant into a comprehensive

safety and risk management platform. It automates routine tasks, significantly reducing specialists’ time, delivers decision accuracy of 97%, and scales efficiently

to corporations with tens of thousands of employees.

The platform has already demonstrated fast ROI,

with payback expected in less than one year.

Contact me:

kolbenkom@gmail.com

Poland
Warsaw