Authored by Tim Rollins, Director of Content Marketing, Exterro

Preservation law was written for a world of paper, where "keeping records" simply meant not throwing banker's boxes away. Modern corporate data doesn't work like that, and the gap between legacy legal assumptions and digital reality is where millions of dollars in cost, risk, and friction now live.

On a recent episode of the Data Xposure Podcast—hosted by Mike Hamilton and brought to you by Exterro—guest Andy Hansell provided a clear-eyed reality check on this disconnect. Hansell, currently Director Counsel at Target, brings a rare perspective: he has practiced at a major law firm, worked as an eDiscovery consultant, and now sits in-house on the enterprise client side.

Having seen discovery battles from every seat at the table, Hansell has developed a knack for challenging conventional legal industry wisdom. Here are four contrarian takes from his conversation on Data Xposure that every General Counsel, legal ops leader, and litigator needs to hear. 

You may not agree with all of them–heck, I don’t necessarily agree with all of them. But they’re interesting takes, based on years of experience and expertise, and it’s important to think about them if you’re in-house counsel involved in eDiscovery.

Hot Take #1: AI Isn't a Defense Cost-Saver—It’s an Arms Race

There is a widespread narrative in corporate legal departments that artificial intelligence will act as a silver bullet for slashing outside counsel bills and document review budgets. The logic seems sound: if AI can summarize files and draft motions faster, legal spend should plummet.

Hansell sees a very different future: AI in litigation is fundamentally an arms race.

"Plaintiffs may adopt this stuff faster than defendants," Hansell noted during the podcast. Because plaintiff firms typically operate on contingency fees, every hour spent investigating or drafting claims dilutes their profit margins. They have an immediate financial incentive to leverage AI to automate repetitive tasks.

Instead of lower legal bills, corporate defense teams should prepare for:

• Scraped & Targeted Claims: Plaintiffs using AI tools to scour corporate websites and public data to spot regulatory violations and launch serial litigation.
• Higher Claim Volumes: Commoditized, well-drafted demand letters and complaints hitting corporate doors faster than ever.
• More Resilient Complaints: Claims crafted with AI to specifically withstand early motions to dismiss.

For corporate defense teams, adopting AI isn't about pocketing easy savings; it's about building the internal capacity required to keep pace with an AI-powered opposition. 

Some Additional Thoughts

Overall, this take makes solid sense. But on any given matter, getting to the facts faster and saving time and limiting data volumes can translate directly into savings. Time will bear out if there’s an explosion in plaintiff’s attorneys filing claims in the hopes of seeing what spaghetti sticks to which walls.

Hot Take #2: Preservation Law Belongs in the Paper Era (and It's Bleeding Budgets)

Judicial preservation expectations remain stubbornly caught in a paper world. Judges often view ordering a company to "preserve everything" as a zero-cost ask because they still imagine physical documents sitting harmlessly in storage. In reality, litigation routinely reaches back years into past transactions stored on legacy software that the business no longer actively uses. Preserving that data is anything but free.

To satisfy broad preservation obligations, organizations are forced to keep unsupported software online and maintain aging, physically failing hardware solely for potential discovery. "Trying to convince the business that we need to keep this stuff online, up to date on new hardware, for the sole purpose of meeting a preservation obligation is not only expensive, it just doesn't make a lot of sense," Hansell explained.

Furthermore, attempting to migrate legacy data to modern cloud platforms to keep it safe carries severe operational risks: system updates during migration can inadvertently cause permanent data loss, exposing the company to sanctions. Hansell argues it's time for procedural rules to grant safe-harbor protections or fee-splitting mechanisms for legacy data preservation.

Some Additional Thoughts

I’ve seen enough savvy judicial opinions around AI, data volumes, and novel data sources to counter-argue that there are definitely some savvy judges out there who understand technology and want to see counsel adopt it effectively in the name of producing quick and efficient legal outcomes. That said, the cure Hansell proposes is sound. If procedural rules aren’t sufficient to provide the safe harbor and fee-sharing arrangements desired… then that’s something a good ESI protocol can hopefully help with!

Hot Take #3: In-House eDiscovery Is 10% Law, 90% Detective Work & Persuasion

When law firm attorneys picture in-house practice, they often envision high-level legal analysis and trial strategy. But inside a major enterprise, the day-to-day reality of eDiscovery is much grounded in internal diplomacy.

Hansell breaks the role down into two key components:

• Detective Work: Outside counsel usually receives a refined, finished product. In-house counsel must act as a detective to map out where data actually lives, who owns the systems, and where "shadow IT"—unapproved apps used by workers to get their jobs done—might be hiding.
• Persuasion: System administrators and IT leads reside in the technology department, not legal. In-house lawyers rarely have direct authority over them. "Other than vague threats... I don't have a whole lot to make them do what I want them to do," Hansell shared. Getting results requires building relationships, explaining the why, and practicing persistent, polite persuasion.

Hot Take #4: Million-Document Reviews Are Compliance Theater—Only 50 Documents Matter

The eDiscovery industry is built around processing massive data volumes. Discovery disputes often devolve into fights over broad collections and terabytes of information.

Yet Hansell offers a sobering perspective on what actually determines the outcome of a dispute. Whether a case involves 1,000 documents or 10 million, the entire matter will ultimately turn on a tiny handful, usually no more than 40 or 50 key records.

"A jury, a judge, or whatever is never going to look at broader than that," Hansell emphasized. Those core documents must dictate the real strategy: do we settle, or do we litigate?

Once you accept this truth, modern eDiscovery splits into two distinct tasks:

1. Strategic Discovery: Finding those ~50 game-changing documents as early as humanly possible to guide business decisions.
2. Compliance Discovery: Processing, reviewing, and producing the remaining mountain of data as cheaply and efficiently as possible simply to satisfy court obligations.

Some Additional Thoughts

While I don’t have the hands on experience of in-house eDiscovery, this take rings 100% true to me. That’s why Exterro takes pains using our ARMOUR AI-powered workflows to surface the relevant facts faster, whether you’re in an eDiscovery matter or conducting a forensic investigation. Getting to the fact faster gives you a leg up on strategic decisions, like “Settle or litigate?” and gets counsel providing true insight and value faster–hopefully faster than the opposition or regulators.

Eating the Elephant One Bite at a Time

For legal teams trying to modernize their data risk programs, Hansell’s primary advice is simple: don't panic. You cannot fix every legacy system overnight. Focus first on your largest data repositories, talk directly to custodians to uncover how systems are actually being used, and ensure that auto-delete policies are immediately paused when litigation looms.

Want to Hear the Full Conversation?

Tune into this episode of Data Xposure to hear Andy Hansell and host Mike Hamilton discuss alternative fee agreements, legal ops benchmarking, and why eDiscovery programs are "written in sanctions motions".

Listen to "When Data Preservation Isn’t Free" with Andy Hansell on Data Xposure

‍