It’s hard to overstate how quickly artificial intelligence has gone from an abstract concept to a frontline tool in operational decision-making.
In 2025, security leaders are navigating not only the perennial risks of theft, sabotage, and natural disasters, but also an increasingly volatile threat landscape shaped by geopolitical instability, supply chain disruption, and cyber-physical convergence. And at the same time: expectations for responsiveness and efficiency against a backdrop of limited capacity have never been higher.
For those tasked with protecting global operations, the mandate is becoming increasingly clear: evolve, or risk falling behind.
To that end, AI offers a compelling pathway forward for many teams. One that promises to streamline the security risk management lifecycle, surface insights faster, and bring clarity to chaos. But embedding AI tools in your security stack responsibly also requires a balance of innovation and grounded expertise, especially for those managing complex asset footprints.
At Human Risks, our focus is on reimagining what technology can do for security teams on the ground. The AI workflows we develop are designed with practicality in mind. Helping scale security teams’ impact across large physical footprints by putting the right data – and the right decisions – into the hands of those who need them most. But before we dive into our approach, let’s break down the current state of play across the wider industry in April 2025.
The Strengths and Limitations of AI in April 2025
In September 2024 Human Risks partnered with Decis Intelligence to publish a comprehensive joint whitepaper on AI Applications in Enterprise Security Risk Management.
As we shared at the time – and have discussed with our customers since – the AI landscape for security leaders is continuing to mature significantly, particularly in the realm of large language models (LLMs), machine learning, and predictive analytics. Fast forwarding to today, increasingly accessible AI tools can process and interpret vast volumes of unstructured data – from incident reports and open-source threat feeds to internal communications – and distil them into structured insights in seconds.
In the context of Enterprise Security Risk, this means AI tools are highly effective at:
- Summarising incident reports across geographies in real-time.
- Detecting patterns in historical data to predict likely threats.
- Suggesting mitigations based on similar scenarios across an asset footprint.
- Automating classification of risk scenarios based on potential impact.
However, the reality is it’s not a silver bullet. While AI models excel at handling large datasets and identifying patterns, they also struggles= with context and nuance – areas where human experience remains vital. Risk ratings, for example, are deeply contextual and often informed by on-the-ground intelligence or organisational risk appetite, which an AI system simply can’t interpret correctly.
Likewise the quality of AI outputs will always be heavily dependent on the quality of the data that feeds them. In environments where data is siloed, outdated, or incomplete, AI analysis will often be misleading. That’s why Human Risks champions a human-in-the-loop approach – augmenting judgement, not replacing it.
So what are the pain points that AI can help solve?
By automating the mundane and enhancing situational awareness, AI allows security teams to focus on what they do best: making critical, timely decisions.
1. Information Overload
Modern security teams are inundated with data – incident logs, travel advisories, threat intelligence, operational updates – but often lack the capacity to turn these insights into informed analysis. AI can ingest this information from multiple sources, highlight what’s relevant, and push timely alerts to the right stakeholders.
2. Delays in Reporting and Decision-Making
Traditional reporting processes can take days. Or even weeks. Automated analysis using AI tools can significantly speed up this process – summarising findings, suggesting mitigations, and even generating draft reports for review at the beginning of each day. This not only improves agility, but frees up time for strategic work.
3. Lack of Standardisation in Risk Assessments
Most organisations operate across multiple jurisdictions and business units, each with their own approach to risk. Leveraging AI tools to standardise terminology, criteria and formatting across jurisdictions and language barriers can play a significant role in creating a unified view of risk across an entire organisation. Format never feels important, until you have 15 different descriptions of the same causal event in a dashboard.
4. Subjective and Inconsistent Risk Scoring
Risk matrices – which in spite of their limitations are the backbone of most organisations’ approach to security risk – are often filled out manually and can vary dramatically across regions and teams. By leveraging AI to do the heavy lifting on the analysis that needs to inform scoring exercises, and recommend baselines based on real historical patterns and asset context, AI can drive dependable outcomes from risk scoring exercises without having to have a security expert in the room.
5. Inadequate Foresight and Scenario Planning
With predictive models, AI can simulate potential threat scenarios and offer probability-weighted outcomes, helping teams prepare better response plans. For example, if supply chain risks rise in a certain region, AI can alert teams and recommend scenario-based responses based on past disruptions.
So what’s our focus? How Human Risks is deploying AI across our customers’ workflows
At Human Risks, we believe the best use of AI is one that keeps humans in control, while letting technology do the heavy lifting. By embedding AI into everyday security workflows, our goal is to enable security leaders to act faster, standardise processes across regions, and allocate resources more strategically.
- AI-powered threat suggestions that dynamically surface based on asset data, geography, and recent alert activity.
- Incident pattern recognition allowing for immediate identification of emerging trends and hotspots as part of the incident reporting process.
- AI contextual intelligence summaries embedding what asset owners and regional leads need to know, such as via our recent integration with decis.ai, directly into risk assessment processes.
- Auto-generated recommendations and assessment summaries that help teams produce high-quality, consistent outputs across language barriers and levels of expertise.
Interested in learning more? Connect with the team and book a demo.
Stay ahead of risk. Stay human.
