Industry executives and experts share their predictions for 2026. Read them in this 18th annual VMblog.com series exclusive.
By Jennifer Crawford, CRM, CIPM, Director of Solution Engineering and SaaS Presales at Access
Most records management programs fall somewhere on the AI implementation spectrum: from curious discovery to full implementation mode. You may already be using AI at full-speed to manage your electronic records, or, conversely, you may still be exploring your options and trying to determine if AI will truly bring material value to your records management program. Wherever you fall right now, it’s useful to continually refresh your awareness of what a successful value-added AI implementation entails when it comes to electronic records management.
1. Practitioners will continue to be overwhelmed and paralyzed by the volume of potential use cases
I spoke with dozens of practitioners attending a recent records management conference, and most were simply overwhelmed by the abundance of use cases for AI, and not sure where to begin to even understand how those use cases might apply to their programs. They are feeling constrained by their daily responsibilities already, and don’t have the additional bandwidth or resources to properly investigate AI options, let alone determine which are appropriate for their budgets and program goals. This tension is exacerbated by the feeling that the train has left the station and they’re still standing on the platform. This is why it’s important to continue to offer resources to those who are still getting up to speed on how to discover and evaluate how AI could offer meaningful benefits for their programs.
2. You will still need to spend far more time and money than you think to get your content AI-ready
AI is often marketed as out-of-the-box; ready to immediately change your life for the better. However, the reality is what we keep learning from experience and funny internet memes: AI is only as good as the data and the instruction it is fed. For example, when considering the specific use case of applying AI to remediate unmanaged shared drive content, we are often surprised by how much planning and preparation is required before we can even turn it on and start to see the work begin. This will continue to be the reality, and you will need more resources, time, and budget than you expect, which leads me to the next point…
3. AI still requires more human involvement than you expect
AI is not a “set it and forget it” implementation. Most companies need intensive human involvement, especially at the outset, to define the role and rails of the AI scope, as well as define and support ongoing human-in-the-loop reviews of your AI analysis. Again, using the example of remediating shared drive content, even if it is your hypothesis that all of your shared drive data is ROT (Redundant, Outdated, Trivial) information, you will still likely need to conduct cursory reviews of the AI classification just to be sure that you aren’t disposing of something that you do in fact need to retain. Or, if the intention is for the AI to also apply retention classification to non-ROT content, you will also want to ensure that it is properly classifying records and, if applicable, moving records to the appropriate long-term repository. In short, there are many scenarios and logistics to consider even when just focusing on the remediation of unstructured data in shared drives alone.
4. AI will still not bring quick ROI, unless you are targeting very low risk data
Building on the picture we’ve built with the previous points, you can see why AI will not bring quick ROI. The only use case for quick ROI is when you already have leadership and budget support to empower AI to unilaterally make decisions to identify content, and either delete it or reclassify it without any human-in-the-loop involvement. This is a rare scenario in real life because most companies do not have this risk appetite. It is bold to take this approach because there is always a strong possibility that you could be prematurely disposing of information that legally must be retained and/or should be on legal hold. There is also the risk of improper classification of content that is indeed eligible for immediate destruction but is retained longer than legally necessary. This becomes a privacy or legal discovery hazard.
5. Vendors will still over-promise
There’s an AI bandwagon and we all think we’re supposed to be on it in some way, including those selling the virtues of AI. Be diligent when exploring vendor offerings to ensure that they fully understand your use cases, your program goals, your budget, and your time horizon. Also pursue testimonials from other customers to gather their lessons learned from vendor implementations, so that you can walk into a vendor engagement with your eyes open.
There are many strong benefits of utilizing AI to operationalize all aspects of records management; however, like any major investment, do not let a false sense of urgency distract you from the necessary due diligence to ensure a meaningful return on investment. Quality over speed is the imperative; otherwise, you risk losing far more than you gain.
##
ABOUT THE AUTHOR
Jennifer Crawford, Dir., Solution Engineering & SaaS Presales – Access
In her role at Access, Jennifer leads the Company’s pre-sales function, partnering closely with Sales, Product, Marketing, and Enablement teams to elevate Access’s technical storytelling, demo execution, and pre-sales excellence. Her focus is on driving alignment, accelerating time-to-value, and delivering differentiated client experiences across the sales lifecycle.
Jennifer brings a dynamic blend of leadership experience and technical expertise, having held senior roles at Bank of America, Iron Mountain, Gartner, and most recently, Wells Fargo. Her passion lies at the intersection of information governance, product innovation, and customer experience.
Jennifer holds an MBA from California State University and a BA in Music from the College of Charleston. She’s also a Six Sigma Green Belt, Certified Records Manager, and Certified Information Privacy Manager.
