Elevate the power of your work
Get a FREE consultation today!
From unstructured legacy files to governed information: Building a defensible remediation strategy
Blogs and Articles
Let's explore what it takes to make AI-supported remediation both defensible and sustainable at enterprise scale for legacy and go forward files.
As unstructured information accumulates, organizations face a difficult question: how can they remediate content at scale without compromising accuracy, compliance, or business trust? Technology can make high-volume classification possible, but organizations still need to understand what they hold before deciding how that information should be handled.
This was the focus of our June and August Iron Mountain Education Series webinar, where I was joined by Carol Garcia, Information Governance and Management Professional, and Julian Stapleford, Lead Counsel at Wells Fargo Bank. Drawing on their experience leading a large-scale global remediation initiative at Wells Fargo Bank, we explored what it takes to make AI-supported remediation both defensible and sustainable at enterprise scale for legacy and go forward files.
When scale outpaces manual review
When organizations are dealing with billions of files, there is no practical way to review each one manually. AI-supported classification gives teams visibility into what they hold at a scale that simply isn't possible through human effort alone. For Wells Fargo, that meant remediating more than 40,000 shared drives containing nearly four billion files, making automation the only realistic way to begin understanding what the organization held.
There are too many files for humans [to process]. It has to be done by machine in some way.
The decisions that follow still need to account for the obligations and risks attached to that information. A file may be subject to retention requirements, contain sensitive information, or create unnecessary exposure if it is kept for too long. Accounting for these obligations allows organizations to use automation to support more informed decisions, not simply faster ones.
Building defensible remediation
Large volumes of legacy information can't simply be assumed to be redundant, obsolete, or trivial. Before organizations can decide whether files should be retained or deleted, they need confidence that they understand both what those files contain and the obligations attached to them. At Wells Fargo, this developed gradually, with six rounds of AI-assisted review building greater trust before remediation decisions were made.
We scanned all files six times. And then we had the human in the loop that was looking at output and making corrections as needed.
Each review cycle strengthened the file classifications. Business teams first provided representative records, or exemplars, to train the model, helping ensure it reflected real business information rather than assumptions. Legal experts helped shape a conservative, risk-based approach, while human reviewers continually refined the results as the project progressed. Rather than relying solely on automation, Wells Fargo combined governance with human expertise to support remediation decisions that could be clearly explained and defended.
Cross-functional teams build trust
Remediation affects every part of an organization, from legal and compliance to the business teams responsible for the information itself. As a result, no single function has enough context to make every remediation decision in isolation. Wells Fargo brought together subject matter experts from across the organization to contribute a different perspective on risk, governance, and the way information was actually used.
There was a lot of communication, but there was back-and-forth communication. It wasn’t a one-way street. It was bidirectional.
That same collaborative approach shaped how the project was delivered. Executive sponsorship helped establish organizational support, business champions connected project teams with employees, and regular reporting kept stakeholders informed as the work progressed. Just as importantly, feedback from the business continued to shape the project as new challenges emerged. Giving people visibility into the process—and a way to contribute to it—helped build trust in the decisions that followed.
Sustaining information governance
Large-scale remediation creates an opportunity to reset how information is managed, but the work doesn't end once legacy content has been classified. Without new governance controls, organizations risk rebuilding the same information challenges over time. Sustainable remediation depends on embedding better practices into day-to-day operations so information remains governed long after the initial remediation project is complete.
The go-forward is to not get into this mess and have to do another cleanup. So after this effort is done, business as usual is continuous scanning.
At Wells Fargo, remediation became part of “business as usual” through continuous scanning, clearer retention controls, and repositories designed to manage information throughout its lifecycle. Employee training reinforced those changes by helping people understand where information belonged and how it should be managed going forward. Regular reporting also gave stakeholders ongoing visibility into the program, ensuring governance continued after the remediation project had finished.
Automation is only the beginning
Enterprise-scale remediation doesn't end when information has been classified. Long-term success depends on embedding governance into everyday information management, so organizations can continue making defensible decisions as information grows and changes over time.
To hear how Wells Fargo approached enterprise-scale remediation, visit Iron Mountain's 2026 Education Series to watch the on-demand recording of The hidden advantage: Activating unstructured data through remediation to fuel AI.