eDiscovery Daily Blog

• November 10, 2025

Masters Conference: Seattle, 2025 |  Article by Sheila Sadaghiani, Regional Director of Sales, CloudNine

Speakers:

• Doug Kaminski, Infinnium
• Mike Russell, Expedia Group

Session Abstract:

Let’s face it, most organizations are drowning in data and struggle with gaining control. At the same time, threat actors are targeting those who hold a lot of data and especially sensitive data. Storage used to be cheap, but now we’re seeing vendors in the data ecosystem monetizing that volume to a greater degree. Add the element of GenAI and the data footprint grows exponentially as does the need to control any exposure.  Join us for this important session to learn what you can do to address this growing issue!

Why Information Governance Should Come Before AI in eDiscovery

At the recent Master’s Conference Legal in Seattle, I sat in on a powerful panel discussion about information governance and it really resonated with what I see daily in eDiscovery sales. As an account representative for CloudNine, where we offer CloudNine Review for eDiscovery document review, I often work with corporate legal teams and large law firms who are struggling under the weight of their own data.

One of the panelists, Doug Kaminski, Chief Revenue Officer at Infinnium, shared an important point: organizations are creating and duplicating data faster than they can manage it, and without a strong governance framework, they end up losing control of their information. His insight underscored what I experience firsthand, many clients simply don’t know what data they have, where it’s stored, or how much of it is redundant.

Another panelist, Mike Russell from Expedia Group, expanded on that idea by emphasizing the practical side of governance. He noted that governance isn’t just about meeting compliance obligations, it’s about enabling business agility. As Mike explained, when organizations take the time to understand their data landscape, they’re not just preparing for litigation or audits; they’re improving collaboration, reducing security risk, and setting the stage for smarter decision-making across departments.

Much of this data sits in Microsoft 365, local drives, and shared servers, sometimes duplicated multiple times across custodians. When litigation arises, the result can be overwhelming. I’ve seen organizations collect 10 terabytes of data when the actual relevant set might only be 500 gigabytes to 2 terabytes. That overcollection drives up processing, hosting, and review costs dramatically, all because no one had visibility into the data landscape beforehand.

To put the scope of the problem in perspective, several statistics were mentioned during the panel discussion:

• Over 50% of enterprise data is considered “dark data”, information that organizations store but don’t actively use or even know exists.
• ️ 30% to 50% of stored data in most organizations is duplicated or redundant, increasing costs and risk exposure.
• The average enterprise holds over 10 petabytes (10,000 terabytes) of data, yet less than 10% is typically relevant for eDiscovery or compliance purposes.
• Data storage and management costs rise an estimated 35% year over year when governance is not in place.
• Unstructured data makes up roughly 80–90% of total corporate data, making it the hardest, and most expensive, to manage in litigation.

Doug Kaminski also emphasized that this lack of visibility doesn’t just inflate discovery costs, it creates security vulnerabilities. Disorganized, unstructured data is a prime target for threat actors. When sensitive information is stored haphazardly across systems, it’s not only harder to find for litigation but easier to exploit in a breach.

There’s also a growing misconception that AI can fix these issues. It’s true that AI has tremendous value in accelerating review and surfacing insights, but as Doug noted during the panel, “AI is only as good as the data you feed it.” I couldn’t agree more. If an organization’s data is chaotic and duplicative, AI simply amplifies that noise, leading to higher costs and less reliable results.

That’s why, in my opinion, information governance must come first. Governance creates the structure that makes AI effective. It’s proactive, not reactive. It prevents unnecessary spending, improves security, and lays the foundation for more accurate and efficient discovery when litigation occurs.

AI is an incredible tool, but it’s not the solution to poor governance. Clean, well-managed data allows AI to reach its potential. Without that foundation, even the most advanced technology becomes an expensive workaround.

When you roll these into a financial model, companies that implement governance-first strategies often realize 2× to 3× higher ROI within 12–18 months compared to those with no formal governance.

Factor	Governance-First	No Governance
Data Reduction	25–40% average decrease in total data volume via defensible deletion and deduplication	Data sprawl grows unchecked (duplicate data often 29%+)
Operational Efficiency	Faster response to DSARs, discovery, and remediation (2–5× improvement)	Delays in retrieval, indexing, and review cause higher costs
Risk Mitigation	Reduced breach exposure and fewer sanctions due to proactive classification	Increased incident response and legal costs due to unmanaged data

Putting governance first doubles the organizational ROI by cutting waste, improving compliance, and unlocking automation potential, while ignoring governance leaves money and risk on the table.

The takeaway from the Seattle Master’s Conference was clear: before we can rely on AI to revolutionize eDiscovery, we must first reimagine how we govern data. Information governance isn’t just a compliance initiative; it’s the cornerstone of every successful discovery strategy.