How Rev Speeds Up Evidence Review and Case Analysis (Full Transcript)

[00:00:00] Speaker 1: Hi, everyone.

[00:00:02] Speaker 2: Thank you for jumping in and joining us today. I hope this isn't a long time spent together today as I will make it quick for us. But really what our main goal here at Rev is to continue spreading the knowledge that we are learning from the market and what we can do to truly continue to develop the platform not only for all of our clients, but also for our clients to start developing how to utilize the platform in a way in which helps them better really understand the intricacies that Rev can provide. Today's focus is on getting the contradictions and as well as the evidence or understanding what happens within the evidence. And so when jumping into our platform, a lot of people understand that that Rev not only can be based off of a transcript, but also is based off of the videos and the reports provided to it. And a lot of the ways in which a lot of the law enforcement investigators and detectives that I work with are focused on the first question being, what do I have? What can I get out of this and how do I continue this process of understanding what I have in my whole data set? And so during this conversation, my main goal is to identify the problem that we're looking at, what I'm hearing from the field, the ways teams are utilizing this, if it be within an internal affairs unit, a detective who is supporting a patrol unit or investigators that are in between the DAs, public defenders offices, and everywhere in between, as well as showing you the platform and how those teams are utilizing the platform. I'll provide some takeaways and hopefully get some questions from you all. And then I also have a document that I've kind of drafted where previously I provided a prompting guide that helped the team understand or my users understand what the best practices for prompting is. And then now I'm hopefully going to provide some better questions on the individual different types of outcomes that you're seeking and which questions I have found from the market have provided the best results. The problem that I'm seeing and what I'm hearing from the market is that there are so many cases. Everybody's dealing with so many different things at once. There's not enough time to really watch everything more than once or maybe not more than twice, more than two times. And from there, once you've listened to the 10 to 15 body cams, you still are searching for what am I missing here? I've I've heard this. I know what I'm looking at, but now I need to understand. Did they mention something that isn't even provided to me in this data set? Is there something that I need to go back and review multiple times because I was so focused on something else at that moment? So we pulled me away from the desk that now that contradiction that was said in one video could also really help me with my case for another. The other piece is that we know that digital evidence has grown by 98 percent since COVID. There's body cams in every single accident, every single report, every single situation. There's multiple people on the scene. There's multiple reports that are created. We want to make sure that you're getting the most out of this. And what I've realized is that Rev's not here to fully write every report for you. There is some document creation that we can do for you and your team. But the main goal for Rev is not to be the process where you have to be fearful to submit something and have to go sit on a stand if you made the claim because Rev wrote it. That's not our end goal. Our end goal is to allow for your team to get the most out of it and write still themselves and us be the red pen to your black pen where we're providing you information, analyzing it, and digesting it in a way that your team can then speed up that process so they can get through that X number of cases quicker and allow for them to truly do the things that they wanted to do when they chose this career. So diving into a case and what I'll do now is I'll jump into Rev's platform and we'll look at two different types of cases. One that's built for an internal affairs unit where they were looking at multiple different types of videos as well as a few different reports that were written. And we'll ask questions like give me a case chronology. Tell me the inconsistencies that were provided here. We also have a weapons possession case that we had drawn up. And from there, we'll look at some other questions that have been really helpful to dive in in more depth. And so from here, I'll jump in to Rev's platform. Can you see my screen? Hopefully. Rev's platform is very straightforward. When you jump in, it's really easy for you to bring things to us. Again, we can drag and drop. You can bring through files. You can bring multiple types of files. These are all the different types of files that we currently can accept. It keeps going. If there is any files here that you'd be looking for that your team gets a lot, feel free to reach out to me directly. I'd love to be able to support in making sure that you can get anything into the platform that you need. The first one that I'm going to look at is a motor vehicle stop. This one had the suspected possession of a weapon. And we're looking at one video, a few dispatch reports, and then there were some reports that were created alongside of that. What I've done is selected all of them together and taken them to an area called Analyze or Rev Insights. Rev Insights is our powerful, powerful area within our platform that allows for you to not only look at things individually, but together and provide any type of search that you need. Now, the first thing that I did was look at these files individually, which I can do here and I can watch the video myself and watch the transcript as well. I can click into these documents and review them here on the right side myself as well. And I can even expand it and dive in and review it so that I have a general synopsis of what I have first. But then from there, I asked a pretty specific prompt. This is the type of prompt that I fully suggest when trying to find out really intricate information. Build a changing referral summary for this case. So thinking of the pass off between an officer to a DA or to an investigator. Cite the specific evidence and the timestamp supported and flag any inconsistencies. And then look at the information and so on. And one thing that I like to add every once in a while is to understand that I want to make sure, even though Rev's best practice is that we don't go and cite anything for you, in this specific case, I wanted to ensure that it didn't. And so being integral of what you provide and asking, similar to a Chad TPT or Gemini or a Claude, is making sure that you add some type of question or some kind of statement ensuring the outcome that you're looking for. So do not add any case law or make up any information provided. Within our platform, we not only can call out the charges filed here based off of the information provided, we also call out where we found information. And so we will highlight in those reports, allowing for you to go and look at the original source, as well as the text and identify where something was found. What I hear from a lot of my clients is that this is a huge piece to allowing for the team to really just double check all of their work. The inconsistencies flagged is the big part of our conversation today. And a lot of it's focused around each one of the files provided and what individual change there has been. And we even highlighted, you know, how big of an inconsistency it is. In 17, it calls out the glove compartment. Someone said the glove box. In some cases, that may not be a huge deal, but I've heard from my clients that even a change of words, especially when reading Miranda rights or getting statements from people can be a big difference. So making sure that those align is a huge piece. The purpose for travel was another reason. Something like I'm going home. Or we can review the video. And then understanding that I'm just driving around. About 30 seconds later. Can be a big difference. These inconsistencies keep popping up and allow for you to review them from each step with these citations. Now, why is this one of the first things that people do? I think one of the things that I talk about with my attorneys that I work with is you never really know how much information that you have here. And it may be different on the law enforcement space. But often another question similar to inconsistencies that will be asked is what information here is referenced. But is not provided to me in the data set that's here. And so if it's referring to a separate report, it's referring to a dispatch log or a booking report, something that is in this data set that needs to be passed off to them. That is another amazing question that can be provided to just thinking about the data set as a whole, not even within the transcript. And then the last piece being now with REV, we can draft up documents. And so the document, this whole output, this whole answer here is a huge way for you to then export and even export with the citations as a table of contents at the very end. Once exported, it can be taken to a supervisor. It can be taken to another officer to review and double check. This is an amazing way for you to jump in and maybe even change up some of the words and put into your own work, into your own tone. So changing some of these em dashes, knowing that it's an AI drafted document, or even in some ways, I have people adding disclaimers such as this document. Can't see. Was drafted by AI. Knowing that that is what is required for you to pass off to your team. The other portion that allows for you to do is you can highlight sections now. So maybe this inconsistency here is the most important one. I can highlight here and I can ask a separate question like, tell me more about this. Specific. Section. And now because I've dragged and highlighted this, the whole platform will now focus on its next response being about this inconsistency. So now as a separate document, this document can be provided in the same sense as a separate piece. The creation of documents are speeding up the process for a lot of our teams. And I think this is a huge time saver when it comes to creation of any in any time that that you have a data set in front of you. To the second example that I wanted to provide was focused on a family disturbance or or an internal affairs use case. A lot of body cams, a lot of videos, a lot of interviews associated to it, as well as some reports that were created by the team. To help with a more clean data set. I can take that to analyze like I did before. And it'll separate there or in the same sense. I jumped in and built one out of speed or in focus on time. Instead, build a case chronology. This was a prompt that we already had created here. We have dozens, if not hundreds of different ones here, allowing for you to categorize them. Focus on what you're looking to get what your outcome is. In chronology is one of our main use cases by a lot of my clients. I said build a case chronology. It has an internal prompt and it provided us this answer. It gave us each date. Alongside quotes. Hyperlinks. Take me directly to the videos and allowed for me to understand every time something was mentioned. All the way down. From May to June. The other piece being the exact same ideal that we were thinking of before where point out the inconsistencies of these files and anything that's missing or that's referenced in this data set.

[00:14:09] Speaker 1: We have multiple inconsistencies in this data set.

[00:14:11] Speaker 2: Again, a huge value for you to continue to jump in and find value out of our outputs. We have multiple inconsistencies here again highlighting the value that our system believes could be provided and and how important it could be. And again, a huge value for you to continue to jump in and find value out of our outputs. Is there any questions that the team here has about how we're finding inconsistencies or trying to find the constants associated to data. As I know we have a few participants here.

[00:14:50] Speaker 1: Yeah, thanks, Dylan. We'll go ahead and open it up to Q&A now. The first question that we have is, what are some other questions you have seen be successful within the platform.

[00:15:04] Speaker 2: Yeah, that's, that's great. Actually, as mentioned, created a document that I think is similar to the prompting guide that I originally made is associated to the different types of ways people are viewing our platform, or at least when they're looking at a data set. I created a questions guide that is, you know, hopefully going to sharpen those questions that are associated to, you know, what outcome you're looking to get. So inconsistencies again flag every place. Any statement contradicts them with the body cams, corroboration, which statements of the officers interviews corroborated and within the body cam compliance, if we were looking from the internal investigation side, or, you know, internal affairs and looking to see in regards to the standards. One of the main use cases my teams like to also do is bring us handbooks and utilize that as something that is thrown into that data set and say within my state, say here in Texas. This is the state standard for our department, and I want to review these videos and allow for this data set to be utilized and have those policies reflected in my outcome, or in my, my answer, as well as report writing, or, you know, case handoff and so if anybody's interested in, in this document and want to find more ways or questions that they can get the most out of our platform, would love to have. Other questions that you're finding successful shared to me so I can add them to this document or, you know, love to share this to you as well.

[00:16:39] Speaker 1: Perfect. Thanks, Dylan. The next question that we have is, what's the quickest way to get files into the platform.

[00:16:48] Speaker 2: Yeah. So as mentioned, we have, we have the ability for you to drag and drop any files you can bring a zip files at any moment. The other pieces, especially in the law enforcement space, most people forget that we have a mobile app. And so, when you're interviewing a suspect or interviewing a witness, having a mobile app that allows you to record directly and upload directly into the platform is one of the quickest ways to get interviews into our platform. It transcribes it immediately allows it the second that it's done to be fully transcribed and added to the work, the case that you're already looking at. And that is a perfect place for you to start or speed up that process of getting files into the platform. We also have integrations into Doc, Box, OneDrive, and Google Drive. And so, a lot of that being a wherever your documents are sitting that allows for us to speed up the process and hopefully, you know, live out of rev when it comes to your evidence evaluations moving forward.

[00:17:48] Speaker 1: Thanks, Dylan. The last question that we have is what kind of security does Rev have?

[00:17:54] Speaker 2: Yeah, so Rev is a security first platform and focused on understanding that whatever you provide to us, you know that it's not being trained on it can be deleted at any moment. And anything is, you know, sitting in servers based out of the US. We are CJIS compliant. Depending on the state that you're in, we would love to support and make sure that we match to the state requirements that you have. We also can be HIPAA compliant, as well as anything provided into our platform is not training or sharing out to external models. And so, knowing that whatever sensitive data that you have or your team deals with, it is safe within our platform, and you can bring it to us and work off of it without being concerned like you would be on any other platforms out on the market.

[00:18:44] Speaker 1: Perfect. Thank you. Well, that concludes our Q&A. If we don't have any other questions, we'll be sending the recording of the webinar shortly. So keep a lookout for that. And if anything else comes up, don't hesitate to reach out to Dylan. But thank you all for joining and have a great rest of your day.

[00:19:04] Speaker 2: Yeah, thank you so much again, everybody for joining. And as mentioned, here is some of my contact information. If y'all have any interests, please shoot me an email and would love to chat.

[00:19:15] Speaker 1: Awesome. Thanks, y'all. Have a good one.

[00:19:17] Speaker 2: Take care.