Bringing Order to Data Chaos

In an era where unstructured data is growing exponentially and AI is reshaping how we work, organizations are grappling with the chaos of securing and managing this information. On a recent episode of Cyber Security America, host Joshua R Nicholson sat down with Kyle DuPont, CEO and co-founder of Ohalo, to explore how his company is addressing this challenge head-on.

From Bureaucracy to Breakthroughs

Kyle’s career began in traditional finance and technology, working at MUFG and Morgan Stanley in Tokyo and London. Frustrated by the inertia of large institutions, he caught the startup bug, initially working on anti-money laundering workflows before zeroing in on a bigger problem: unstructured data.

Founded in 2017, Ohalo started as a GDPR compliance solution. But it quickly evolved. As Kyle explains, “The compliance issue merged with security. We realized the core issue wasn’t just regulatory, it was that companies didn’t understand their data.”

Enter Data X-Ray, Ohalo’s flagship product. It empowers teams to automatically discover, classify, and protect sensitive data across vast, fragmented systems.

The New Security Perimeter: Identity and Data

The old model of cybersecurity, building a perimeter around an organization, no longer applies. With remote work, cloud storage, and AI usage, Kyle believes, “The perimeter is no longer the edge of a firewall. It’s the identity and the data that identity has access to.”

In response to this shift, Ohalo has embraced attribute-based access controls (ABAC), which move beyond simplistic role-based models to assess the actual content and context of data. “A user role is too blunt,” Kyle explains. “We need to understand whether a document is sensitive, whether it contains intellectual property, trade secrets, or regulatory material. That context should inform who gets access.”

Why Traditional DLP Fails

Kyle is direct about the limitations of legacy Data Loss Prevention (DLP) tools: “Nearly every organization we speak with has implemented DLP at some point and most found it frustrating. False positives were overwhelming, and the systems often failed to scale effectively.”

Ohalo offers a more advanced alternative. By leveraging AI to analyze not just keywords but the broader business context of each file, their solution can determine instantly and with high accuracy, what data is truly sensitive. “A CV isn’t just a name and an email,” Kyle notes. “It’s a career history. You can’t identify that meaningfully with pattern matching alone.”

Ohalo’s approach is also language-agnostic, performing well even in non-English environments. “In Japan, many organizations leave traditional DLP tools disabled because they generate too much noise. Our generative AI reads the file and interprets its meaning without complex rule-building.”

The Three-Body Problem of Data Governance

Kyle compares today’s enterprise data challenge to the “three-body problem” in physics—an unpredictable system of gravitational pull. “If you think about unstructured data as the gravity between departments, it used to be just IT and Security. Now, with the rise of AI and data initiatives, there’s a third force: the Chief Data Office,” he explains. “Everyone’s trying to use or protect the same files for different reasons. It creates organizational chaos.”

This complexity extends to the secure use of generative AI tools like ChatGPT, Copilot, and vector databases. As Kyle puts it, “We’re not just reducing risk. We’re enabling revenue by making AI use safer.”​​

Looking Ahead: MCP and AI Agents

The future, Kyle says, lies in Model Context Protocol (MCP). A proposed standard for how AI agents communicate and execute actions securely. While still emerging, MCP may play a similar role in AI that REST APIs did for web services.

“We’re experimenting with MCP internally,” Kyle shared. “Eventually, securing these model-to-model communications will become essential.”

Advice for the Next Generation

Kyle encourages young professionals to pair computer science with deep domain knowledge. “CS will soon be table stakes. You need to bring that to a specific field. Biotech is one exciting area where AI is about to transform everything.”

💡 Key Takeaway:

As AI reshapes enterprise data, we’re redefining how organizations secure and understand it. We’re not just protecting files, we’re bringing clarity, governance, and trust to the unstructured data landscape.

🔗 Schedule a meeting to know more.

🎧 Listen to the full episode on Cyber Security America on Spotify, Apple, or YouTube.