← All cases

AI Risk Assessment: How Cando Rail Saves 20+ Hours per Safety Report

⚡ Automation Web 22 Aug 2026 ▲ 269

Tools Used

GPT-4.1GPT-realtimeRecall.aiLangChainMicrosoft Teams integration

Results

Saves over 20 hours per risk assessment; drafts reports in seconds for less than $0.05 each.

Why AI-Powered Risk Assessments Matter for Your Business Now

Imagine being able to multiply your company’s safety checks by six, without hiring more staff or sacrificing quality. For business owners and managers in Russia and the CIS, the pressure to scale operations while maintaining strict safety standards is real. Manual risk assessments are time-consuming, expensive, and often bottlenecked by expert availability. If you want to grow and stay compliant, you need a smarter way.

That’s where AI-powered agents come in. Cando Rail & Terminals, a major player in rail logistics, faced this exact challenge. By partnering with Fractional AI, they built "Peter the Safety Agent"—an AI system that automates and accelerates risk assessments. The results? Over 20 hours saved per assessment, instant report drafts, and the ability to scale the safety program by 6x in just a fraction of the time. In this article, you’ll see exactly how they did it, what tools they used, and how you can apply these lessons in your own business.

What Was the Problem with Traditional Risk Assessments?

For Cando Rail, safety is non-negotiable. But their risk assessment process was a headache. Each assessment meant up to 16 hours of meetings with subject matter experts (SMEs), weeks of back-and-forth on revisions, and endless manual cross-referencing of PDFs. Learning from previous reports was nearly impossible, and scaling up meant even more time and cost.

  • Up to 16 hours of SME meetings per assessment
  • Weeks spent on revisions
  • Heavy manual work, little knowledge transfer
  • > "The manual process couldn’t scale: each assessment required up to 16 hours of SME meetings, weeks of revisions, and heavy PDF cross-referencing, with limited ability to learn from past assessments."

    The pain point was clear: Cando wanted to multiply the number of annual risk assessments at the heart of its safety program, but the process just wouldn’t scale. Human experts needed to remain in control—but the paperwork and logistics were holding them back.

    How Did AI Transform the Risk Assessment Workflow?

    Enter Peter the Safety Agent, a custom-built AI solution using GPT-4.1, GPT-realtime, Recall.ai, LangChain, and Microsoft Teams integration. The workflow was redesigned to keep human experts in the driver’s seat, while automating everything else.

  • Prework automation: The AI generates a first draft of the risk assessment in seconds, before the meeting even starts.
  • In-meeting voice assistant: During SME discussions, the AI transcribes and structures the conversation in real time, updating the risk assessment draft on the fly.
  • Seamless collaboration: All updates and drafts are instantly available to the team, reducing the need for endless revisions.
  • This means safety experts can focus on decision-making, not paperwork. The AI handles the heavy lifting, but humans remain fully responsible for every safety call.

    What Tools and Models Powered the Solution?

    Cando Rail’s solution wasn’t just about plugging in ChatGPT. It combined several advanced tools:

  • GPT-4.1 & GPT-realtime: For natural language understanding, summarization, and drafting reports.
  • Recall.ai: To transcribe and structure live conversations during meetings.
  • LangChain: To orchestrate the workflow and connect different AI modules.
  • Microsoft Teams integration: To fit seamlessly into the company’s existing collaboration tools.
  • Each tool played a unique role. For example, Recall.ai captured every detail of expert discussions, while GPT-4.1 turned those details into clear, structured reports. The integration with Teams meant no new apps or training were needed—AI worked right where the team already collaborated.

    > "Prework automation produces a first draft of the risk assessment in seconds; An in-meeting voice assistant transcribes and structures SME discussions in real time, generating readily updated risk assessment drafts as the conversation unfolds."

    What Business Results Did Cando Rail Achieve?

    The numbers speak for themselves. With Peter the Safety Agent, Cando Rail:

  • Saved over 20 hours per risk assessment
  • Drafted reports in seconds for less than $0.05 each
  • Laid the foundation to 6x their risk assessment program in just 17% of the original time
  • This isn’t just about saving time. Faster, more scalable risk assessments mean safer operations and the ability to grow without bottlenecks. For any business where compliance and safety are critical, this is a game-changer.

    Key outcomes:

  • 20+ hours saved per assessment
  • Reports generated instantly
  • Program scaled 6x with human oversight
  • How Can You Apply This AI Approach in Your Own Business?

    You don’t need to be a railroad company to benefit. Any business that relies on expert reviews, compliance checks, or complex documentation can use a similar AI-powered workflow.

    Here’s how you can start:

  • Identify repetitive, expert-driven processes that slow you down
  • Look for AI tools that integrate with your current systems (Teams, Zoom, etc.)
  • Pilot AI assistants for pre-meeting prep, real-time note-taking, and instant report generation
  • Keep your experts in control—use AI to automate the grunt work, not the final decisions
  • > "The system enables Cando to 6× its program in just 17% of the time, while keeping human experts fully in control of every safety decision."

    What Are the Practical Steps to Implement AI Risk Assessment?

    If you’re ready to get started, follow these steps:

    1. Map out your current assessment workflow. Where are the bottlenecks? 2. Choose AI tools that fit your needs. For voice transcription, try Recall.ai; for report drafting, look at GPT-4 or similar models. 3. Integrate with your existing meeting tools (like Teams or Zoom) to minimize training and friction. 4. Start with a pilot project—measure time saved and quality improvements. 5. Scale up, keeping human experts in the loop for all final decisions.

    Checklist for getting started:

  • Identify high-value, repetitive tasks
  • Select AI and integration tools
  • Run a pilot and measure impact
  • Scale and refine
  • Frequently Asked Questions (FAQ)

    How much technical expertise do I need to set up an AI risk assessment system? You don’t need to be a developer. Many AI solutions now offer plug-and-play integrations with tools like Teams or Zoom. Partnering with a vendor or consultant can help you get started quickly.

    Is it safe to trust AI with critical safety processes? AI handles the paperwork and drafts, but human experts remain in charge of all final decisions. This approach keeps safety standards high while reducing manual work.

    How much does it cost to implement a system like Peter the Safety Agent? The case study shows report drafts cost less than $0.05 each. Initial setup costs will vary depending on your needs, but the ROI comes from massive time savings and scalability.

    Can this approach work outside of rail or heavy industry? Absolutely. Any industry with complex documentation, compliance checks, or expert-driven reviews can benefit from similar AI-powered workflows.

    What if my team doesn’t want to change how they work? AI tools like this integrate with familiar platforms (like Teams), so your team won’t need to learn new software. Start small, show quick wins, and adoption will follow.

    Conclusion: Take Action on AI Automation Today

    Cando Rail’s story shows that AI isn’t just hype—it delivers real, measurable results for businesses that need to scale complex processes. By automating the paperwork and freeing up experts, you can boost productivity, cut costs, and keep quality high.

    If you’re ready to start, map your current workflow, identify the biggest time sinks, and explore AI tools that fit your business. The future of scalable, safe operations is here—and it’s powered by AI.

    🔗 View source