Transcript for "Governance Beyond the Documents: Visual Collaboration, AI, and the Discovery Trail Nobody's Watching": 5432. Hello. Hello, everyone. Hi. Welcome. Welcome. Thank you guys so much for joining us today. I'm gonna... Before we get started here, I wanted to quickly introduce myself. I'm Jacqueline. I'm a marketing manager here at Reveal, and at Onna. So as... Today, I wanna give you some highlights real quick before we jump into it. One, yes. This webinar is live, which means if you guys have any questions, jump in the chat, and you guys can answer... And we'll answer those at the very end of this webinar. Or if something's relevant, we might jump in throughout, but keep that in mind. Next, this is also being recorded, so if you have any, so if you end up having jumped out early or you want to watch this back again, you can go back and watch this on demand. And then finally, yes, the slides are available for download or there's plus some additional additional resources in the docs tab right next to that messaging box. With all that said, I would love to pass this off to our speakers today. Jeffrey and Chris, take it away. Thanks so much, Jacqueline. Good afternoon, everyone. Thank you for joining us today. We're very excited to be talking with you today about governance beyond the documents. Chris and I were working on this workshop, for a little while now, and very excited about how it kinda fits in the information governance scheme. But, before we get started, let's just talk a little bit about, us. Chris, would you like to introduce yourself for the audience? Yeah. Fantastic. Thanks very much. So, yeah, I'm Chris Jones, and I come from Miro. And at Miro, I am the principal AI governance and security solutions engineer, which is a a very long title for what I do. But my involvement is is less about what people are doing on Miro boards and how people create things, the collaboration side. Instead, I focus on the the security and the governance aspect. Of this because the moment you start looking at environments like Miro, and we'll give you some examples as we go, it's a big set of unstructured data. There's a lot of ways of collaborating that goes on there. So I'm there to help organizations understand that, understand the risk involved in that, and and help mitigate that. You know, what information can they get out? Now prior to joining Miro, I spent twenty years working at one of the larger cybersecurity providers. So I've been talking about things like classification and data loss prevention since the very earliest days of that as a subject to talk about. And that helps in this subject because I've seen it from its very earliest genesis through the the simple stages into this big, sprawling environment that we have now. But, yeah, that's me. Thanks, Chris. Welcome. Glad to have you here. Miro is one of our fantastic technology partners here at Reveal and Onna. By way of introduction, I'm Jeffrey Wolff. I oversee technology partnerships here at Reveal, and I really kinda spend my day thinking about, evidence and and how work becomes evidence. And that's really what we do, at Reveal in our... With our products, Onna, Reveal Hold, Logical, and and Reveal Enterprise. Today, we're really not... This is not really a product showcase or, we're gonna talk much more about the information governance of nonstandard work as it were. But before we launch into that, I think it would be useful to hear a little bit from you. So, just by... This is completely anonymous, so no one's tagging your responses here. But if you will, could you just select one of the the options on screen. Does your information governance policy today within your organization explicitly cover tools that are collaborative, chat, AI generated content, things like Slack or Teams, things that are nonstandard work documents? Because that's really where we're gonna be focusing our efforts. And and so it'd be good to hear from the, the audience in that respect. So go ahead and put your answers in there, and, we will kinda move on. We'll come back to that. Jacqueline, we can get those poll results in a in a minute or so. Right? So we'll come back to those results in a moment. While we do that, let's talk about the central theme of of today's presentation. Kind of thought long and hard about how to summarize this, but, I know, Chris talked about his background, prior to joining Miro. I've been in the ediscovery space now for almost twelve years. Prior to that, I was in solutions engineering in enterprise content management, publishing information governance. We've both been talking IG for a very long time with our respective audiences. But information governance, as you can see by the core idea of our slide here, is built for nouns. Right? Traditional work, ESI as we call it in the eDiscovery world, electronically stored information, is based around in the past, been based around files, based around email messages, doc... Final documents, documents that have, an author, a date. And that is not necessarily how work happens today. Work now very much happens in verbs. Right? It's collaboration. It's, it's ongoing learning. It's continuous iteration, reacting to things, generating new content, involving IT. Chris, do you wanna tell us a little bit about how this kind of reflects, from the Miro side of the world, your way of work? Because that's really the basis of Miro, isn't it? Yeah. And I thought... Found this one really interesting. I hadn't considered governance from this perspective until we kind of... We're working through the presentation today, but it absolutely has shifted. There's this huge paradigm change in terms of how we work, and I do think that a lot of people from a governance perspective are still thinking of it from this noun's perspective. It's this is an email. This is a document. This is a file. This is a server. And while there may be certain information that exists out there like that, really, things have shifted. And for me, joining Miro was a a huge change. My my earliest background was in email security. And the earliest days of my career, we started talking about content security on email. It was the first time I'd seen this kind of level of data security looking for personal identifiable information, financial records, keywords, etcetera. And now to to come to a company like Miro, which is all about collaboration. It's all about having this big space where lots of different types of data and document can all live side by side. It really shows where we are. And moving to Miro, the reason I came to Miro is because I could see how we've been doing things in the background from a data discovery and from a government perspective. And I can understand how companies like Miro just don't fit that space. They need a way of helping you express that information. You need to be able to quantify this big, unstructured mess, of of of information, essentially. And that's exactly what we can see with the owner capabilities. It's taking that content and and making it re... Readable. Absolutely. Absolutely. So let's let's talk about, you know, what happens before something becomes evidence. Right? If you think about it, what we call evidence in in litigation, it used to... Like I said, we used to live... That used to live in email or finished files, and now it's scattered around places that governance programs were really never designed to look at, whiteboards in the Miro case or visual canvases and other products like that, chat channels, threads, reactions, meetings recordings, and most notably now, AI conversations. So one of the the big challenges we have is most corporations have adopted some foundational model and in many cases, multiple ones. And they have conversations with those, those models, and they, iterate. They develop content from them. And, I don't know that the information governance models have quite caught up yet to that. So that's, does this... Chris, is this matching what you you see in your space? Yeah. Very much. Yeah. Very much so because communication has evolved. Like, I mentioned the email thing. And for years, that was how I would always communicate. Whereas now I rarely use my email application because messages are happening in Slack. They're happening in Miro. They're happening in Atlassian. They're happening in all of these different environments. And they're much more collaborative. It's... Here's what we're working on now, but there's there's other content in the same... That's living in the same space. And it's... Well, here's the document that is in this environment, but it sits alongside a table. And we have our chat built into that section. So, yeah, the the way that we are communicating, the way that we are creating data is just so vastly different than where governance is up to. And, yeah, it's I think to your point, it's this living document. When you're creating evidence, where does that trail end? Probably nowhere. It's just a long line. We just have to decide where in this line are we looking. I know the, if we've got any, you know, legal departments, general counsels on the line, it's probably something that keeps them up at night, to be honest with you, because, you know, the the concept of discoverable data used to be much more finite, and now it is infinite. We, I was I was talking with an attorney. I was at a law conference last year, and I was talking with an attorney on the case that she is working on. And she was telling me that the case involved sensors in concrete that were rendering... That were capturing data and sending that telemetry to, a server. And so they had... That was part of the discoverable content for the case. This is, you know, beyond what we would imagine. This... That... That's out in the IoT, the Internet of Things, but, data data that resides in a information governance strategy or in a corporate litigation has vastly expanded from was... Where it was even five years ago. And so I wanna make sure that organizations are thinking about that and how they, how they approach that. Yeah. Before before we jump into that, for for those on the the call today, Chris Jones, that don't know what Miro is, can you talk a little bit about what Miro actually is and maybe show us a little bit about it? Yeah. Sure. So at at its core, Miro is an innovation platform. It's it's a place where you can go as an organization to to build and iterate on ideas. It's framed on on the canvas. It's it's infinite canvas elements, but as if you were using a, you know, a whiteboard in a normal environment, there's lots of different things that can be on there. It's going to be pictures. It's going to be sticky notes. It's going to be documents and tables. So the way people work with an environment like Miro is very varied, and pretty much every organization that I speak to is going to treat Miro very differently. And and what might be a better way to illustrate this and to illustrate this sprawl of unstructured information that is, you know, this big issue for governance is to show you, an example Miro Board. So I'm gonna share my screen. And, you should now be seeing a big complex Miro Board, suitably named. And what we were doing is building out some information. So this is all there to kind of give you an idea of how people might use Miro. And while I'm obviously incredibly zoomed out at the moment, we can see that there are, slight information that is embedded in here. These could very easily be presented. As an organization, people use Miro. We have a a lot of AI capabilities. We've built some prototypes within this type of information and even generated images that might show how people could be using a particular tool. We have plenty of sticky notes allowing people to kind of vote on ideas, lay down their ideas and how they'll be building it. Documents and details that multiple people could contribute to. This document, by the way, is actually being generated by AI. And if we were going into that, we'd see the attribution that the links all of that together. We've got the prompt information around that here. But beyond that, there are videos that we have within there. We have table information and even things like these chat logs that are going on, because we were talking earlier in terms of building this information out. Because as a company, how would you collect this type of information? How would you pull out the chat log that we have within there? How would we have the fact that on this particular engagement that we've got, we have one message and then the reaction to that is an emoji, essentially kind of approving what's going on within here. And overall, Miro boards are very interactive. They're built all around collaboration in terms of who can I share this board with? Who do I want to invite to work on this tool with? Whether they're bringing their own ideas, whether they're reviewing my ideas or iterating on them, these boards are very much living, breathing documents. And in terms of things that I've seen, there's countless different organizations that use Miro and and tools like it, to build basically everything that you work with and that you can see. A lot of that spent starts its life in Miro, which means there's a lot of very sensitive information that people are pulling out from that. So I will stop sharing my screen now. So, Chris, tell me... And I... By the way, I I love the sticker feature in Miro. It's one of my favorites. But, obviously, when you think about that type of, product, it's often used in product design. Right? That's a big. one. But can you tell me the types of companies that use tools like Miro? Well, I'll be very coy, because I work in the the data security side. I won't name names, Sure. but, anyone who's on this webinar, if you look around your room at the moment, you will see pieces of technology in your room of things that have been built. I can almost guarantee that something within kind of arm's reach of where you are will have started its life off within Miro. If you look outside, you'll again find lots of elements. And I deal with all types of companies. Clearly, technology companies are at High Park. We also deal with, kind of educational organizations. I deal with, legal companies, architectural firms. I deal with, manufacturers, even critical infrastructure organizations, you know, very heavily invested in Miro, which means the information that they're building in there is really very, very sensitive, especially when I'm talking to, like, a critical infrastructure organization because these are the sorts of organizations that, you know, are going to be targeted by people who want access to their information. They are gonna be creating things that are really heavily regulated, that they need to keep information for long periods of time, that they need to prove ownership of detail. I was on a call just today where we were talking about something very similar, around that in terms of what should live in your Miro board. Are there things that should be restricted from that? How can they manage that? How can they prove that? And to be clear, Miro is just here. When we're talking about Miro, we're talking about an example of non... Of very nonstandard, content, right, that needs to be addressed in an information governance platform. It is not. solely about Miro. There's... Like, we we we both work in Slack every day. And Slack in and of itself is a perfect example of another application that you need to have information governance around because there is no end product. Right? A conversation on a channel or a DM isn't a living, breathing thing, that you're constantly iterating. Chris brought up the point of the, thumbs up emoji. Right? You know, the traditional way of work was you'd write... You'd create a document. That document would be finalized. You go back and forth with, a manager or supervisor. You'd you'd hone that document, and then it would... You would finalize it. And then maybe it was it was a contract, there would be signatures involved. But, ultimately, there was an end product that you would use. And if you were... If that that end product was subject to litigation or a regulatory response or a subpoena, you'd have a document to produce. It's very easy information governance. There was a record of authority. That is not the case in in the new way of work. Right? Because if I... If Chris sends a message saying, here's here's my version of the project. Is this good to go? And I am... As his manager give a thumbs up to that. I... That's my consent. My consent was an emoji. And Chris now. take that... Takes corporate action on it with my approval. And if we wind up in litigation, that reaction is important to the litigation, which is amazing. when you think about it. And we were talking about this earlier today, but, you know, there's a reason that this is all called being signed off on because it was a physical signature at one point. We've evolved from that. clearly, but the thumbs up might seem quite clear. Right? We we we may understand that, but people bring personality into emojis as well. So it isn't just a thumbs up. Someone might go okay or 100%. I've seen so many varieties of that, and I suspect everyone on this call knows exactly what we're talking about. In any of these tools, be it Slack or Miro or any of the thousands of other collaborative and unstructured data tools, People do work in emojis and Okay. people communicate in different ways. How do you track that? How do you give attribution to that? The metadata, the timing on that, link it with everything else that was happening at that moment, even versioning. Yeah. Yeah. And so that's a great segue into, we'd be remiss to not mention what Onna is. Right? Onna is one of Reveal's, premier solutions. It is one of the few products on... In the space that collects Miro boards and 30 types of other content from other, cloud solution manufacturers. But the intent behind it is to create a defensible product that is searchable, reviewable by by primarily the legal teams, but, to allow you to collect that data in a way that it has context. Right? So for Miro, that's... That could be challenging. It could be, you know, it could be sharing, taking a board like the complex one that Chris showed and making sense of it. Right? Providing a metadata record for every element in that board. So if somebody asks, well, when Jeffrey put his thumbs up sticker on that section of the presentation on that Miro board, he was providing approval to move forward with that and when did that happen and can we go back and capture that information? So again, this is not meant to be a product showcase, but it is meant to show that you are able to capture though... That type of metadata even on a nonstandard document type. So, you know, what we're talking about here is all of is all of the assumptions that we have today around how we manage content. I think this this slide encapsulates that pretty well. It's it's it's a hard thing to do today because modern content really breaks all of those governance programs that we've built up over the years for other traditional electronically stored information. Right? An email is a final version of something. When you send it, that is the final version of the document. Now you can conflagrate that, of course, because we have moved in the last few years into the world of modern attachments. You used to have an... If you had an email message and you had an attachment, that was the final sent version. However, we we switched away from that into sending a link to documents in the... Both the Microsoft ecosystem and the Google ecosystem. Those are collaborative documents. Right? So what is the final version of that Excel spreadsheet or that Google Sheet that you attach to an email message? Is it the version that was, is it is it the version of that document that existed when the message was sent, when the message was received? Is it the final version after, you know, at the time you do a collection for litigation? There's a lot of... Still a lot of questions around that. But you can see how applications like Miro or Slack, really just break that model of there is not one final version. Chris Jones, do you wanna talk about, authority record... You know, authority in records and where they live. and So... Yeah. And there's some interesting things for this. I've seen this come up in a load of different forms over the years. And we talk about, yeah, documents and links. You know, what was that that current version you you mentioned? But then it it also comes down to, does this version even exist anymore? I've dealt with cases where people have needed to try and retrieve data, not just in in Miro or elsewhere, where someone has come in and actively deleted it. And this might not be malicious, although it absolutely could be. It could be accidental. Very likely someone doesn't even realize that what they're doing is the wrong thing to do. But when you come back and try to govern this information, how do you track what was going through there? Which author's type of information are you going to be... To be looked at? And where is that data? Are there multiple versions of it? I think it's a it's a very key point. And coming to to Miro is... I saw a really interesting shift. Beforehand, I was talking about, data loss and data awareness and data control in the broadest sense. And it'd be this constant thing of, okay, well, can we cover this piece of technology? Can we see what's happening here? And certainly with Miro, it was... The question was flipped in terms of, are you tracking the data from from Miro? And you should be asking yourself the same question. Are we tracking the content that gets put into these additional applications? Because they're gonna have different records of different versions. They're gonna have new authors that will have worked on pieces of information. They will be using AI tools that will maybe rewrite it, maybe add to it, maybe have even created the the document that the... You're then seeing embedded into emails elsewhere that you assume came from the user. And, absolutely, Miro was what cemented that concept in my mind because I saw how people were treating Miro compared to a SharePoint information... A SharePoint repository because everyone understands SharePoint. It might be terabytes worth of data, but we know the idea of what we're talking about there. It's a familiar standard. You look at the other applications you've got there. Some of them will be brought in. Some of them could be even shadow IT. But they're going to behave in a very different way, and they need tools that can pull that information into into the right way and and give you the right data. Chris, you mentioned shadow IT, and it's... I think it's probably worth exploring that one for the audience. For for those that aren't familiar, shadow IT is when applications are brought into an enterprise that are not directly sanctioned by IT. IT hasn't deployed the application. Miro, Slack, those are both actually great examples of products that are available in some free fashion, freemium as it were, where people can start working and even collaborating, inviting their colleagues and start working in it before the the the organization has an enterprise version of it or has established a governance policy around it. Do you wanna touch a little bit on that? Yeah. Interestingly, that was how I originally came across Miro. In my personal life, I started using the tool, but it would have been considered as a piece of shadow IT. I've also worked with other organizations. And again, this is before, I joined Miro, where we would be using technologies that would discover shadow IT applications that they had to see what people are using on their endpoints. And again, Miro came up a lot. Slack came up a lot. There's dozens and dozens of different examples. And everyone here, again, you're gonna have an idea of what your teams are using. You may not know everything out there. But yes, channel IT is clearly going to be a way where lots of data can suddenly be used in different environments. And I think we're going to come and touch on this later, but this also then starts brushing up on the AI Yes. argument. Yeah. I don't wanna get was gonna ahead of. myself. No. No. No. That's that's exactly where I was gonna go next because the one of the... I think what I see the most lately in my own organization is note takers. Note takers are the latest version of shadow IT. If you look at the fourth bullet on the slide, which is my favorite, assumption of modern content was that the human wrote it. So we... What we used to deal with, we used to deal with documents that were written by humans or email messages that were written by humans. That is not necessarily the case anymore. Right? So we are in a in a situation now with all of the the foundational models that are all available, whether they're paid or or free or freemium, is people are using them within a professional environment, within a corporate environment and generating content. They're having conversations. Hopefully, they're not exposing any information that they shouldn't read to those models, And, hopefully, corporations are controlling that. But, there... Those conversations. and the documents that are generated as a result of those conversations that they iterated on, those are also, that's also content that needs to follow some sort of information governance policy. So I think it'd be interesting if we come back to this. If we were running this same webinar in a few years' time, would that assumption be not that a human wrote it, that an AI wrote it? Are we then gonna need to be proving... We're already in this situation. We need... People need to be proving their involvement with it. And I think it's only going to get bigger and bigger a problem as AI becomes more prevalent in terms of how people work. Was this generated by you? Was this generated by an AI? Yeah. Absolutely. Absolutely. I was, I was being pinged by my colleague to let you know that the poll results from earlier that I ever actually put up, showed that the strong majority of the audience actually said that some. of their, I'll put this slide back up so you can see it. So some of them was the strong majority from the audience. So, you know, that implies that we are... That organizations are thinking about, this content, but not able to address all of it today. And so it will... IG policies, like everything else, will iterate over time. But, in the new governance world, these are the the questions that we have to ask ourselves. Who changed what and when, really? And, Chris, do you wanna touch on any of these? Yeah. So this for me is the the backbone of what I talk about on a day to day basis because, as I said, my my focus is on data security. It's, you know, what what is on the board, for people to work on. And, again, you you can... I'll use Miro as my example, but you can ascribe this to any application that people are working with that's, you know, not not structured. I work with organizations that need to go back and prove what something looked like at a certain time. It will be a case of we've developed this and now another organization has developed something similar. Can we show where we were at this staying? It's it's showing you workings out on the tests, essentially. They will need to have every version of it because a board is a living, breathing thing. So how how can we see where we were and how we got to this point? It then also comes into interactions because a board isn't just a place where I go. It's a place where I am on, where my whole team can come and access that board because I've shared it with them, where this contractor who's outside of my organization has been pulled in, so somebody from Reveal as we're building our demo board, earlier today. I think there's lots of different ways. You want to capture every single one of those interactions because they're important. Just seeing a thumbs up on on a conversation history means nothing unless you can say that was given by Chris, who would be able to do that. Same with changes. If I delete something, who deleted that? Because, you know, it's not all about malicious actors, but they are a part. It's also about when people do things accidentally or because they don't realize that that's problematic. You need that trail, which links into the objects themselves. The metadata on a piece of any creative content is critical. And I think there's only gonna get more so now that we have AI in the mix. Was this created by Chris? Was this created by an AI? Essentially, which model created a particular piece of content? Or was this created by Chris with an AI? Prompt information, how did this even come to be? If I'm building a document on there, what went into building that document from the AI's source? And then, again, all the things that link to that, all of the sources of the information. If you can't get a full picture of the information, you're likely going to find yourself in a problematic area. And that's the world we're in at the moment. It's not just we have this data. It's what of this can I prove? What of this do I have an evidence trail for? I was at, I was at a conference a couple weeks ago, and they were touching on, AI interactions within collaborative chat. That came up as a topic of conversation. And what is... Where it's going. Right? So every every application, Miro included, has an AI component to it at this point. You have some sort of way to interact with a model, and affect the work that you're doing in that application. Collaborative chat is no different. Whatever your. product choice is, they all have they all have some sort of chatbot that interacts. What's what's changing though is that the level of interaction of the chatbot. Right? So the chatbot effectively becomes a, a teammate in that channel now as opposed to answering questions for you on the side. The the chatbots can be members of a channel and can just answer without being prompted. So if I ask a question in a in a group of 15 people and my chatbot is in that channel and it knows the answer, it's gonna just chime up. I didn't actually ask it to. I didn't prompt it to. So treating AI as a, as a colleague from the... Just from the standpoint of information governance becomes important because no different than Chris responding or Jacqueline responding, AI in that case responded with information and other people took action based on that response. Yeah. So that... You know, right there, that's that's an IG policy to consider is that if you're using a collaborative chat application, maybe making sure that you have the ability to capture not just the human interactions, but also the the the AI agents that are that are acting in those channels as well. The... I I I bring us back to the law for a second. You know, 37 e, going going back to the Federal Rules of Civil Procedure, 37 e, is a reasonable steps approach. Right? It's the... What reasonable steps are are needed to make sure that you secure the right information for litigation or for discovery? And that standard still applies when we're talking about collaborative content like this, nonstandard document content. But the real question is what is reasonable actually mean? You know, how hard is it to capture this information when the evidence changes to Chris's point ten, fifteen times a day? Right? If you're in a in a board or you're on a Slack channel constantly, reacting, that information is not stand... Not just once or twice, maybe... It may be ten, twenty, 30 times a day that there there are changes to that evidence. So, keep that in mind as you're building your information governance policies. Let's, let's kinda switch gears a little bit into a couple of scenarios. We thought it would be useful to take what we were talking about here and put this into real world scenarios. So, Chris, you wanna set this one up? Yeah. So this is a scenario that we came up with, but I have actually come seeing this in kind of real world situations dozens and dozens of times. The organization is building something up. They have a particular piece of road map. It's all set in stone, and somebody senior in the organization leaves, goes elsewhere. And, you know, despite what they've said and signed as they've moved on, lo and behold, the new organizer... The company they've gone to comes up with the same idea. How can we prove if we have the new organization or if we're the old organization, how do you prove when that idea came into... When it was generated? What's on the board and when? Who configured it? Has anybody exported or tried to steal the information? And again, we're in the AI age. So was this person's idea? Was this concocted by an AI? Now, if you are only looking at a situation... Again, I'll use Miro as an example, but it could be anything. If we were only looking at what we have on the board now, well, that doesn't tell us anything. Clearly, the idea is on the board now. The board creation date could be any time. That doesn't prove when the information got on to the board or who built it, which is exactly where this kind of evidence led approach, this ability to pull the information and get all the metadata is going to save you. Because then an organization can 100% say this was created by this person on this date. We have it in our logs. We can show you it. We can prove that it holds up to legal scrutiny. Exactly. So you wanna make sure that whatever whatever governance policy you create or whatever solution structure you create allows you exactly what Chris is describing. Right? The ability to take that data, export it out. Miro as an example, just like any other company has discovery compliance APIs that allow programs to export that information along with intact metadata, and make sure that you have that. Make sure you're collecting that, especially if there's a potential for litigation. You wanna make sure that you have a way of of exporting, not just what the static version of the board looks like today because that's, to Chris's point, not necessarily gonna be very helpful, But how that board, iterated over time and all of the scamps of who authored the changes and how people reacted to those changes, whether they were AI or human generated. So it's important to make sure you have that available to you in a in a format that is contextual and and usable. Let's look at another one real quick. I know we're we're almost at the end here. The prompt nobody preserved. So as we touched on already, more and more, work is being conducted in, No. Claude or ChatGPT or Gemini or Copilot. You you take your pick of the foundational model that you'd like to work with in your organization. But a lot of work is being conducted there. And so, one thing that we've have heard a lot from our customers from a eDiscovery standpoint, a litigation standpoint is how do we get that information out? Because what I will tell you is that the foundational models themselves, they're great at providing the neural network. They're not really designed or entirely around compliance, though. Compliance is somewhat of an afterthought. Most of them have the ability to extract some or all of the data, but you need external tools to help you do that. So, that's a classic opportunity for a product like Onna. We've recently built connectors for, Gemini, for ChatGPT, for Claude, more are coming. And these are designed to capture entire conversations. Right? So the prompt from the human, the response from the AI, any documents that are generated, assuming the, assuming the model allows that type of export can also be captured and and, extrapolated. It's really important that your IG policy takes into account the AI tools, but, it's more than just that. Right? Chris, do you wanna touch on, like, Miro AI and how that is different? Yeah. Yeah. Exactly. Because this becomes a really interesting use case. So I'll I'll explain how we have it set up. But, again, this could be plenty of other technologies. Within a Miro board, you have the Miro AI agent. You can converse with it, you can ask it questions, it can create things, it can research, it can even act as a connector to other external tools if you need to. There's a lot of MCP capabilities within that. But importantly, it's happening on the board. So as far as any of your tools would be considered, this isn't me working on the Claud application or in Copilot or Gemini or whatever. It's a Miro board, and this is what it's being created for. And this always comes back to that conversation we were having earlier on, the shadow IT. Yes, Miro and tools like it are built for collaboration. But when they have AI embedded, which and let's be fair, pretty much every tool out there is touting AI elements now. What can people build within there? So there's plenty of regulations that would kind of talk around this, EU AI being the kind of the biggest one from recently there. But, you know, the point here is, can that prompt information be preserved? If you have tools out there that people are using that they might then go and use as a workaround from your AI policies, you want to make sure that you control and pull that prompt data back because if they are putting sensitive information into those tools, if they're building something, if they're creating documents that they export elsewhere, certainly depending on the type of organization you are, you're going to need a record and a control of that. So, yeah, for me, this is such a big part. AI governance is part of my job description. I speak to organizations about this sort of thing all the time. And it's a big concern for people. What are people doing? Are we even getting value? Yeah. It's a great point. We've touched a lot on, you know, visual whiteboard tools and tools like, you know, collaborative, chat like Slack. But there are plenty of other nontraditional data sources that you you as an organization need to be aware of in your data map. Right? And I just highlighted a few of them on the screen right now. Hyperlinked documents is a... Is... Was kind of the starting point here. If you go back to the Noom case, Nichols versus Noom in in, 2021, that was five years ago. The ruling was that hyperlinked documents weren't considered attachments for purposes of production. Like, if your request for production came in for emails and their attachments, hyperlinked documents weren't considered attachments. But the courts have gone different... In different ways since then. The world has changed a million times by... In the last five years. So keep in mind that hyperlinked documents do matter, and... But what version of those documents is what might be relevant? So make sure your IG policy has a way to address the versioning of documents if your organization links rather than attaches to emails. But beyond that, you you deal with tools like, Confluence, which are Wiki based, also constantly updated and collaborative, ticketing systems, you know, ServiceNow or Jira. We we we work with Zoom as a partner as well. So video messaging and and the transcription of of of chat conversations and the side conversations we're going on on-site of that, and the AI components to those as well. Plenty of ephemeral messaging apps. We... We've all encountered those. If you're in the Discover world, you know the the the mess that ephemeral messaging causes us, from a collection and forensic, processing standpoint. And then, of course, like coding tools, even coding tools, today. You know, Chris mentioned code. A lot of work happens now in code code. And so that iteration of of content, it used to be easy to capture, code samples as part of a litigation case. It's not necessarily the case anymore because they're now being iterated on by... Primarily by AO. These are things that... Yeah. So I was gonna say add one... are aware of. Yeah. Go ahead. Sorry. And these these non traditional sources, you take a tool like Miro and, again, lots of other ones, all of these actually could exist in that one area. So suddenly. you've got this loads of nontraditional ways all in another nontraditional bucket. And I think we touched on this one already. Make sure you treat AI as a contributor. You know, any attribution of of AI generated content, should be tagged as such and preserved and such. I want to leave you with some takeaways as we're coming up on the, the end of the presentation here. But these are some tips, I think, that can be useful in your work, as you try to build your information governance policies to deal with nonstandard documents. Right? Make sure your, your... All of your collaboration chat AI tools are in your data map and you have a retention schedule for them. Set your policies around AI generated content and conversations. Make teams understand that compliance, you know, is an important aspect of their work product. How they deal with it and how they work together is affected by how you deal with the the collaborative aspect of that. Make sure legal holds are extended to these types of content too. Like, don't just focus your legal holds on email and traditional documents anymore. Expand them beyond that. And and make sure that all the stakeholders are in the room. I think that's the most important one. Make sure all stakeholders are available. So, you know, last slide here, three questions for you to take away and come back with tomorrow. Where does our most consequential thinking actually happen in our organization today? If we had to construct it for a court or regular, could we? And do we know which of our content was actually, shaped by AI? So keep keep those in mind. I think there was one question I wanted to jump to. Jacqueline, do we have a minute for one question? Chris, this is for you. A question from the audience. And, Yes. like, So I just you see that. one? Yeah. Yep. Do. you wanna... So very quickly, it's... Basically, am I understanding correctly the data being collected is primarily content and activity created in Miro rather than data pulled in from Microsoft 365 other platforms? So we're using Miro as an example today because it's a great showcase for what unstructured data could be. The reality is, absolutely, Microsoft 365 could be is gonna be used for these sorts of things. But it's it's a little bit more standard. You will have other applications, other tools that don't fit that mold. And we wanted to use Miro to showcase what that difference is and to say kind of what experience we're having from people and organizations that do use and rely heavily on that unstructured data format. So hopefully that's helpful. Thank you for that, Chris. Jacqueline, you wanna take us out? Pleasure, guys. I wanna appreciate you. I wanna thank you all for joining us today. As you... As I said earlier at the very beginning of this webinar, it was recorded. So if you miss any part of it, feel free to jump back to the same link tomorrow and you should be able to watch it back. Additionally, the slides that were mentioned today were... Are... You can find in that docs tab right next to the chat feature, as well as some additional resources we have for you guys. If there's any questions you guys have, Jeffrey and Chris both have their emails here. Feel free to reach out to them. Thank you guys so much for joining us today, and we hope to see you again soon. Thanks. Thanks, everyone. Bye,