International Security Journal hears exclusively from Botan Osman, CEO of Restrata about using AI to break down siloes and to tackle new threats.
The once very separate streams used to ensure people safety, manage travel risk and anticipate and react to critical events are rapidly converging.
The industry is waking up to the fact that, while operations centres may work in siloes, threats don’t.
In today’s threat environment, the interconnected nature of the world means that a cyber-incident can trigger a physical response, or a seemingly localised incident can quickly become international.
Shared frameworks, common language and quick access to information in real time are now essentials, not aims.
People first
One thing that hasn’t changed, however, is that people are still front and centre when it comes to evaluating risk and limiting exposure.
The top priority for any GSOC head remains to keep the organisation’s employees safe.
The difference is that people no longer behave in the way they did before.
Work forces are increasingly dispersed, so keeping track of employees and their threat exposure, not only while travelling but also while working from different locations (including their homes) as well as online, calls for a more co-ordinated approach.
Technology plays a role here for certain, but the bridging of siloes and a move towards joint operations will be the true game changer.
Enter AI
No conversation about technology in 2026 would be complete without mentioning AI.
There is much pressure to adopt it across the board in many sectors, not just corporate security.
Many corporate leaders are seeing it as a quick way to solve staff shortages, inefficiencies and operational bottlenecks.
But AI, however sophisticated, is not a panacea.
An all-or-nothing approach to implementing it, often by simply bolting it on as an additional layer to existing systems, won’t fix the situation.
In fact it could easily make things worse.
AI needs to be viewed in context.
My view is that in Operational Resilience and Corporate Security it is part of an important ‘triangle of trust’, involving software, AI and crucially, people.
In this dynamic, software provides precision, such as exact calculations, audit trails etc, AI is the interpreter, there to recognise patterns, detect anomalies and people make the all-important judgement calls.
When dealing with numerous fragmented and siloed systems, AI can be an extremely valuable conduit between people and increasingly complex data, helping them see the forest not the trees, so to speak.
However, anchoring it in strong data sets with robust and comprehensive guard rails is vital to avoid decision-making based on bogus assumptions.
Guiding principles It is these guiding principles that sit behind Restrata’s operational AI, rosa, which has been built into our resilienceOS platform.
The platform is purpose-built for security and resilience professionals managing people safety, travel risk and critical events.
rosa combines natural language interaction, UI embedded actions and agentic workflows with deep integration into resilience OS and operational resilience data, enabling security teams to rapidly assess threats, monitor incidents, coordinate emergency response and maintain safety.
By operating in real-time across regions, time zones and devices, rosa provides immediate support through natural conversation, removing barriers between security professionals and the intelligence they need.
In rapidly evolving situations, it helps close the gap between recognising a threat and acting on it.
The need for mature data systems
Like many similar systems in other industries, our AI draws from customer data – in this case including personnel locations, asset information, travel itineraries and incident feeds – to deliver insights specific to the situation at hand.
From coordinating multi-channel crisis communications to verifying employee safety across global operations, rosa uses the information to automate the workflows that matter most.
In order for this to be effective though, that data must be robust and accurate.
A simple mapping exercise to establish the maturity of an organisation’s data environment is a great starting point for any would-be implementer of AI systems.
By understanding where it sits on the ‘maturity ladder’, an organisation can begin to introduce AI in a way that complements its operational systems.
This type of assessment is a vital part of our customer onboarding process, laying the foundation for future success.
By understanding exactly where the key data is held, and exactly which data the AI is using to inform its decisions, security leaders can have confidence in its interpretive abilities.
From there, tasks that previously required manual effort across multiple systems can now be handled via an intuitive UI and agentic support, freeing teams to focus on strategy and decision-making.
Inevitable evolution
As the geopolitical landscape quickly evolves, pressure on security teams will only intensify and the need to speed up decision-making around threats and possible exposure will become mandatory.
The recent events in the Middle East have highlighted that environments previously considered safe, such as Dubai, can be exposed to risk overnight and teams must be ready to react at a moment’s notice.
Existing systems, such as those designed for Travel Risk management, have been severely tested during the crisis and have in many cases been found wanting due to a lack of up-to-date location information on people and assets.
There is no doubt that AI will be central to delivering this ‘new normal’, just as earlier technology like GPS played a game-changing role in years gone by.
The question is no longer ‘if’ but when.
