By Ocient Staff
From October 6–8, 2026, our team was on the ground at ISS World Middle East and Africa in Dubai, UAE, the largest gathering of law enforcement, intelligence and financial crime investigators across both regions. Over three days, around 100 sessions ran across seven tracks, from lawful interception and 5G to OSINT, cyber defense and crypto investigation, alongside 20 hours of hands-on training and a busy exhibition floor. Trying to summarize it all would miss the point. The more useful story is the thread that ran through almost every session.
Every threat on the program, from encrypted messaging to deepfakes to crypto laundering, is now at heart a data problem. The agencies that pull ahead will be those that treat their data foundations as a strategic capability rather than an IT project. Four takeaways stood out: encryption is pushing investigators toward metadata, AI is outrunning the data layer beneath it, long-term history is becoming evidence, and sovereignty is now a requirement.
Going dark moves the fight to metadata
Encryption is now the default. Criminals and hostile actors rely on encrypted messaging apps, VPNs, Tor and satellite terminals, and 5G standalone networks encrypt subscriber identities by design. A striking number of sessions focused on what can still be learned when content cannot be read.
The answer is to change what gets analyzed. When investigators cannot read the message, they study the pattern around it: who connected to whom, when, from where and how often. Call detail records (CDRs), internet connection records (ICRs), network metadata, location data and financial transactions become the primary evidence. The consequence is easy to miss. Metadata is high-volume by nature, and its value only emerges when billions of records are correlated across sources. Going dark does not reduce the data burden. It multiplies it.
AI has arrived. The data layer has not.
AI was the headline of almost every product session: agentic investigation assistants, private language models running on-premises, and automated fusion of OSINT, location and intercept data. Much of it was impressive. Yet one session made a telling admission: more data is now slowing investigations down rather than speeding them up.
That is the paradox government faces. AI is only as good as the data it can reach, and agentic AI raises the stakes. An AI agent asks far more questions, far faster, than any analyst. If the platform beneath it cannot keep pace, the agent either stalls or reasons from an incomplete picture, and an incomplete answer delivered with confidence is more dangerous than no answer at all. The bottleneck is no longer the model. It is the data foundation.
History is evidence
The threats on the program rarely reveal themselves in a single moment. Trafficking routes, laundering networks and influence campaigns emerge over months, sometimes years. Yet too many agencies still sample their data or age it off after a few months, because keeping it is too costly or too slow to query.
That is a quiet but significant capability gap. No one can find a pattern in data that has already been deleted. The US intelligence community said the same thing at DODIIS 2026: you can’t find what you don’t keep. Retaining full-fidelity history, and querying it as quickly as today’s data, turns the archive from a cost into an investigative asset.
Sovereignty and trust are now requirements
A clear trend across the Middle East and Africa is the demand for sovereign capability. Governments want AI and analytics they control: hosted on their own infrastructure, free from vendor lock-in, and able to run sensitive workloads offline. Several sessions were built around exactly this.
This is about more than data residency. Investigative tools must operate within lawful frameworks, with clear governance and auditable decisions. The more AI is woven into investigations, the more important it becomes that agencies can show how a conclusion was reached and what data it rested on.
What this means for government
Taken together, these takeaways point to one conclusion. The next intelligence advantage will not come from a single clever tool. It will come from a data foundation that can ingest everything, retain it for as long as it matters, answer questions at the speed AI demands, and stay under national control.
That is the problem the OcientAIQ Unified Data Platform was built to solve. OcientAIQ ingests, retains and queries full-fidelity data from 100TB to exabyte scale, with sub-second queries on a petabyte, deployed on-premises, in the cloud or hybrid. It does not replace the tools on the exhibition floor. It gives them the foundation to work at full scale.
Agencies working through any of these challenges can contact Ocient to start a conversation.