/ 1 min read / Tech Law Talks

Getting your data AI-ready: Why records inventories matter

Effective AI governance starts with understanding what data you have, where it lives, and how it can be used. Reed Smith attorneys discuss how a well-designed records inventory can help organizations identify AI-ready data, manage risk, and accelerate AI adoption.

Transcript:

Anthony: Hello, this is Anthony Diana from Reed Smith, and welcome to Tech Law Talks. Today, we will be discussing records inventories and the importance of record inventories to AI governance. And with me today to talk about this are Therese Craparo, Kiriaki Tourikis, and Kristen Pologruto, all from Reed Smith. Thanks, guys, for joining. So why don't we just jump in? First, if you haven't heard, we did a few other ser - as part of this podcast series, we did one just generally on data governance, and then we talk a little bit about the record retention schedules and how that plays in. This one we're gonna be talking about records inventory. So why don't we just start with the basics and Kristen, if you could just explain what is a records inventory?

Kristen: Sure, Anthony. So a record inventory, it's a common tool used by organizations to compile a detailed listing of all record types, record storage locations, and other data classification data points. Inventories are generally prepared by persons designated within each business line or department, commonly referred to as record stewards, to provide a detailed overview of their record keeping practices. Record inventories meet an organization's legal and regulatory requirement for tracking and locating record information and ensuring that an organization is retaining the appropriate information in the appropriate record repositories. Record inventories enable an organization's legal and regulatory departments to comply with their requirements for tracking and locating information. They can increase an organization's overall efficiency by streamlining access to records, reducing time spent searching for information, which increases productivity. They also ensure business decisions are based on accurate and up-to-date information regarding an organization's business records. Inventories are regularly reviewed, often in intervals of once a year, and since inventories have the most up-to-date information on the records listed on the inventories based on the retention schedules, they can help some other initiatives that are within an organization. For example, record inventories can help locate non-custodial data sources for legal hold and e-discovery purposes, help organizations respond to cybersecurity incidents, and just generally help business efficiency. Inventories are vital to disposal and remediation efforts of outdated information. They facilitate defensible data minimization and they support effective disposal of records that are past the retention period and not subject to legal hold. They're also useful when rolling out new technology initiatives compliantly and securely managing records during data migrations. The nice thing about record inventories is that they're not set requirements and there's no specific format, so organizations are free to develop inventories to meet their own internal requirements and what makes sense for their businesses. But notwithstanding that, most inventories really include the following data points: You know, it's a list of business records to include references to record categories and record codes on the organization's retention schedule, a short description of the records the business is creating and managing, the location, the storage location where that record copy resides, whether it's your SharePoint location, a database, or an application. Inventories also commonly include data classification, including the level of confidentiality of the records, you know, whether this is proprietary information or not, which those types of information are really important when we start thinking about AI usage. An inventory is lastly, they commonly include information regarding personal data within business records, which can help with privacy compliance. So really a strong records inventory enables identification of the AI-eligible data sources, the exclusion of high-risk systems, and faster AI assessments.

Anthony: Yeah, thanks, Kristen. That was that was great. A lot to unpack there. So let's talk a little bit. I mean I hate to use the word data map, but in essence, what a records inventory is, is helping you map where your data is. And I hear from clients all the time, we have no idea where our data is. Like we'll say where are your contracts, where are whatever you're looking for, and they have no idea. And that's part of the concept of a records inventory is you should be able to find stuff, right? You should know where your key - sometimes we call it where your crown jewels are. And obviously we'll talk about the importance of that when we start getting to AI. One concept I want to tease out a little bit is you mentioned the word record repository. And I do think often when I'm talking to clients, they don't understand that not everything in their organization is a record. They're like, records. I have 17 versions of my contracts or I seven con - everyone has a copy of this contract. So it's a record. So I have to do something with it. I think that concept is often missed when talking about, records generally and records management, but I think particularly when we start talking about a records inventory. So I don't know if Therese or Kiriaki, you want to chime in on like the importance of defining sort of that record repository, which I know, Kristen, you mentioned is the records inventory is going to say where it is, and it's going to be your designated record repository or designated system of record. Can somebody sort of explain what is that? And how do you get there?

Therese: Sure. I mean I think that when we're talking about, you know, a designated record repository, what what's important to remember is in an organization, there is only one copy of a record. The record is the one source of that information that is the official, reliable source. There might be many, many copies all over the organization, either because individuals have convenience copies, other systems have copies of that data because they use it for another purpose. But ultimately, when we talk about designated record repositories, it is the source for that record that is the authoritative official source for that data. If you had a question and you had to give an accurate answer, what is the source you would go to for that information because you know that source is correct, accurate, complete for that record? I mean, usually what we say in terms of a process is the business needs to define what those are, right? The business knows how their business works, they know their data, it's the business's responsibility to manage their data. It's the business's responsibility to define where is my official source? What is the source that we need to focus on? And that's important for a number of reasons, right? One of which is what's the most authoritative accurate source, right? Could have lots of other, you know, copies of it that may or may not be right, could be changed, could be used for other purposes and the like. So one is what is the most authoritative source of this data? And two, it's really to easily find it. This is part of what the inventory is defining is the inventory is defining here's this most accurate and correct source for this data, for this record, if you need it.

Anthony: And I do think like just to to emphasize this and sort of like how this really works, right? So, you know, I mentioned contracts, right? If you have a contract management system, that's probably your system of record for contracts. There's probably contracts throughout your organization and everybody's mailbox, whatever. That is convenience copies, right? You can get rid of them anytime. There's no, whatever. So I think that's important. and the other thing I want to emphasize is in terms of practical, right? And and Kiriaki, we're working with a client right now where we're doing a big migration into SharePoint. So all their file shares, and they're putting in the SharePoint, and we're putting specific controls around what we're calling designated record repositories in SharePoint Online. And we're also - they have to have records inventories done before we even start this. Could you just highlight that because I think that's a classic example of why these records inventories are so important in terms of like a migration like that.

Kiriaki: I think it's important to say that record inventories have always been important or a component of information governance. It's not anything new. But I think things like AI or technology transformations add like a new dimension as to why inventories are so important. Particularly with AI, you're asking the question, not just what information do I have, but what can I do responsibly with that information? With AI, inventories are no longer tools to help us manage information or meet compliance obligations, right? They're they're helping us answer the questions about can we use this data? Can we use information for an AI model to train things in a technology migration, right? Making sure that data gets migrated and not left behind. So I think AI, technology transformations, things like this, are really forcing us to understand not just what information we have, but where it lives. What information or that data, what information does it contain? How sensitive is it? Who's responsible for it? You know, just because we have access to information doesn't mean we automatically make it available to AI. And so there's the risk of allowing information to be used inappropriately. Um, and it's not just a challenge, it's it's a real risk. But there's also operationally a challenge of how quickly can you roll out an AI initiative or a large technology migration just because the organization doesn't have enough information to make a decision, right? And so records inventory is gonna address all of these challenges. And I think Kristen touched on it a little bit, but it comes down to visibility, accountability, and I think understanding the sensitivity of the information. So like, with visibility, before you can determine if a system's appropriate or the data is appropriate for AI use, you want to know what records or what information it contains. Is it just general business information? Are there sensitive, confidential, privileged information in there? So having kind of a system level visibility gives you a starting point for assessing the risks without necessarily having to examine every individual record. Second, on accountability, it comes down to I think ownership or stewardship, right? You're proposing to connect a system to an AI tool, who do you go to? Who knows the information in the system? Who can explain the business purpose or sensitivity or any restrictions? So usually a good inventory will have as part of the data points, like, who's the business owner or the record steward, the right people who should be involved, right, in making the decisions about whether that data is ripe for AI use? The third is classification, right? How is the information classified? Is it highly confidential? Is it confidential? Does it have PII, right? That doesn't kind of tell you everything, but it gives you a position of knowledge to start from. You're not starting from scratch. A good record inventory helps you identify the AI-eligible data sources, right? It can help you identify and exclude maybe higher risk systems, and ultimately, operationally, I think it's gonna help you enable faster and more efficient AI assessments. I think it's not just about enabling responsible AI adoption, which it definitely is, but I think governance is sometimes seen as something that slows down innovation, and if done well, the opposite is true, right? Good governance is gonna help you make faster decisions because if you understand your information landscape, you're gonna make decisions more confidently, more efficiently, right? Because you're gonna be able to identify the lower risk opportunities more quickly. You're gonna identify maybe the systems that are more unsuitable. I think if every AI use case requires a new effort each time to understand what information you have, obviously, then, governance is a bottleneck. But if you have the foundations in place, you're going to be able to make decisions much more quickly.

Anthony: Yeah, and I think as you as you pointed out, like again, if you have a records inventory, it's both one, it could be higher risk, right? Because the records inventory will tell you this is a record that maybe you don't want an AI, or maybe you want to have special controls around because it's highly confidential, but it's a record, and it's sort of a crown jewel, and you want to have the theory of when can I use it? When can I not? So it's from a sort of risk perspective, but it's also - the flip side is also this is probably the best data you have, right? It's not the drafts, it's not the - the noise. This should be really good data, if you want to use. One, as I mentioned, it's good to find things, right? If you want to find the contracts, it's always better to go to a contract management system and say, okay, let me find the contract, and is it the right version and all that kind of stuff. But it's also like if you want to use AI, you now know where the golden source is for your organization. Again, something that people don't, I don't think, think enough about is making sure that you have a really good records inventory because it's incredibly valuable. Let's talk about, stepping back a little bit in terms of like how do you do this, right? It's great, I've got my record retention schedule. You guys are telling me to do a records inventory. Where do I start, Therese?

Therese: That is always the key question, Anthony. Where do I start in putting together what seems like it can be a very daunting task, I think, and can feel very overwhelming. Ultimately, as I said earlier, knowing where records are and what designated record repositories are is really a business activity. The businesses know their data. They're expected to know their data. And I think generally where we start is, you know, you could have your template for your inventory, but what we're asking the businesses to do is claim your records. What are the records from our retention schedule, which we discussed in previous podcasts? What are the records that I as a business create, own, receive? What are the records? And then where are they? And it is somewhat of a mapping exercise, as you said earlier, but it's really saying, if I have - we'll just use our contracts - if I am a business line that creates or receives or owns contracts, mapping that out, where - what is the source of it? Is it a group share? Is it a contract management system? Who owns that? So who in the business is responsible for owning it? That does not mean that that poor person is responsible for every single contract for the entire business. It means that person is responsible for knowing where the data is and for knowing that data source, right? They don't need to know the ins and outs of every contract or every record that they're talking about, but you're saying who's the owner? Who's responsible for making sure yes, we have a data source for this? This is a data source. I generally know what's in it, and I know that this is what it is for our business. Right, so when you're talking about it, we usually start with having the businesses go out, we send them the template inventory and say, hey, claim your records, map where records to where they are stored. And we're asking you to tell us what you do in the ordinary course of business, right? That's really what we're asking the businesses to do. Usually it's, you know, coordinated by someone. So typically a records management function is helping to coordinate this, to be the compliance sort of function that makes sure the businesses are doing that, but it's sending the inventories, having the businesses walk through to claim their records, map out where those records are stored, and I think that a lot of times there's a - it's a learning process. This isn't always something that until now has really been integrated as part of a business process. And I think we're seeing more and more with our clients, because of the criticality of this, because of the importance of it, we are seeing it become part of a - the regular sort of BAU business function. And I think one of the examples of this that is my favorite to tell is that had an organization and the RIM team, the records team, was working to implement this inventory process, right? We're gonna ask the businesses to assign record donors and to claim their records and to do this, which is an additional ask, right? It's an additional ask at that time on top of their day to day. And so it went up to the senior management. And it went up to the chief operating officer and to the CEO to say, look, if we're gonna be asking the business to do this, we want to justify that we're asking it. We want to make sure that this is a business process that, you know, the company values. And, you know, everyone was very nervous going into that meeting. Are they gonna understand? And the COO looked at the presentation for about three minutes and said, "So you're telling me we expect our businesses to understand and manage their data?" Of course, that is your responsibility. It's not always that easy, and it's not always that easy to convince people that it's important. But I think the key is, right, you have a process for we're gonna give the business the template that we want them to complete. We're going to send that to the businesses. You give them training and all of that to understand what's expected of them to fill it out. And it's really slowly integrating this into this is what we do as a business day-to-day. We manage our data properly. We know where our records are because that's the most important data. And we need to know where that is for so many things, including for AI, but also, right, just for regulatory compliance and, you know, financial reporting and all of those kind of things. So I think it's we start with, you know, the outreach to the business and having them inventory it and pull that data together. And then gradually over time that becomes, right, part of the business process where they certify it each year. Yes, this is still correct, or well no, we've changed where we store our records. But I think that's it's really engaging the businesses to do this at their level, to identify where, you know, how they organize and manage their data.

Anthony: And then ultimately, just tying it back to the theme of this on AI governance, is you really can't have effective AI governance without a good records inventory, full stop. Obviously you could try and do it, but we all know that so much of the focus of using AI is well, you can only use it for certain types of data, and you're excluding certain types of data. And sometimes they do it based on data classification, as Kristen mentioned, uh, but even the records inventory is helpful in sort of classifying your data, like, okay, if it's HR data and it's HR records, that's likely going to be highly confidential with sensitive personal information. so it is really helpful and I think critical for AI governance. And I think it is, as we keep mentioning, often overlooked when people are talking about AI governance. They sort of skip that. So as we close this out, yeah, I will just open it up for everyone to sort of give any final thoughts on the importance of records inventories to AI governance. Kristen, any final thoughts?

Kristen: My only final thought was that kind of going back to our first in the series of these talks, to be able to have an effective record inventory, you have to first start with that retention schedule in really classifying your business records.

Anthony: Yep, fair. Kiriaki, any final thoughts?

Kiriaki: I mean I think again, it goes down to record inventory, both help you in terms of mitigating risks of using AI, but also operationally, just really help you roll out AI initiatives that much faster.

Anthony: Excellent. And Therese, final thought.

Therese: Look, I think this is the new world of data management, right? This has been coming for a long time. We've gradually been realizing that data management, records management is the foundation, for, you know, running our companies, being able to maximize business profit, business efficiency, and the like. And we're just seeing it exponentially greater with AI, right? If you want to be able to use AI to its maximum value and getting the most value for your company, you need to understand your data first. And the foundation of a records inventory is what gives you that starting point to be able to get the most value you can out of the technology that you have.

Anthony: Yeah. And I think that value is important when you think about how much money being is invested by organizations you're listening to, right? If you're an organization, you're investing so much money in AI, and it's I don't want to say wasted, but you should invest some money in getting your records inventory done. Alright, thanks all. This was great. We're gonna have obviously additional podcasts in this series, so we hope that everyone joins us for those as well. Thanks all.

Outro: Tech Law Talks is a Reed Smith production. Our producers are Shannon Ryan, Amanda Saunders, and Mason Kautz. For more information about Reed Smith's Emerging Technologies Practice, please email [email protected]. You can find our podcast on all streaming platforms, reedsmith.com and our social media accounts at Reed Smith LLP.

Disclaimer: This podcast is provided for educational purposes. It does not constitute legal advice and is not intended to establish an attorney-client relationship, nor is it intended to suggest or establish standards of care applicable to particular lawyers in any given situation. Prior results do not guarantee a similar outcome. Any views, opinions, or comments made by any external guest speaker are not to be attributed to Reed Smith LLP or its individual lawyers. 

All rights reserved.

Transcript is auto-generated.

Related Insights