Built for Trust: How VisibleThread Gives Teams Control Over AI

Hello, everyone. Thank you for joining me. Welcome to our office hour session. If this is the first time that you’re joining us, welcome. If you are, I see a few names in our list of folks that have been here before, so welcome back. Today, we are going to talk about the Visible Thread platform. So occasionally, I’ll do sessions where I jump into the Visible Thread platform and talk about some features. We’re gonna do that today, but with a little bit of a different angle. Chapter Controlling AI Through Governance So it’s not just a kind of straight up demo of the platform. We’re gonna talk a little bit about the way that the platform is built and how Visible Thread is giving you control over AI. So, ultimately, we talk to a lot of customers, and one of the things that we’re hearing most frequently are pieces of feedback around trust, around governance, around structures, around processes, and really how can we make AI work for us in controlled environments where safety and accuracy are very important. So we’re building with that in mind, and there are several aspects of the platform that focus on governance, that focus on traceability, that focus on defensibility. And today, we’re gonna walk through a few of those to show you from a practitioner level how you can trust and verify the output so you feel confident about the things that you’re creating in the platform. And then from a leadership perspective, how can we trust this rollout? How can we govern? How can we make sure that things are working just from the top down in an organization satisfying all of those areas? So I’m gonna walk through some slides in the front end. So we’ll talk a little bit about some of the themes here, show the slides, and then we will jump into the platform. So we’ll spend the bulk of today in the platform, so I won’t, kind of bore everybody with slides the whole time. But, again, thank you for joining, and let’s get started. Just our normal housekeeping. This session is a part of our, office hours. If you have not attended before, these sessions are every other Thursday. We do have a schedule change coming up with the holidays to make sure we can get as many sessions as possible in, by the end of the year. But just to give you an overview, we meet generally every other Thursday. For forty five minutes, we have q and a, and then, you can access with free replays after as well. Our last session was notes from the field. We had, our very own Kyle Peterson on with us, and he talked through some of the challenges that customers are facing and what they’re doing to adopt technology in their real world environment. So that is free for download along with all of our other sessions on our homepage for the site… For this series. Excuse me. So today, of course, we are going to talk about how Visible Thread gives you control over AI. And then our next session on the twelfth, we are really gonna dive into one of the hot topics here recently, when our first reader is a machine. So how do we need to make adjustments, and what can we do now that we know that many of our first evaluators are AI systems and our models themselves? So what can we do there? So we’d love to have you join us for that. So, let me go ahead and start on. We’ll go through this, and then I will share my screen, and we’ll jump live into the platform. Okay. So, talking about kind of the crux of our argument here, and what we hear from a feedback perspective more often than not. And I see that my camera has gone out. I apologize. I’m having some connectivity issues on my end. We have an outage in our area. So just make sure that everybody can see the screen. Give me a shout in our attendee chat if, you can’t see anything or anything goes through, and then I have folks on our end who are, who are tracking everything as well. But just give me a shout if you have any issues, and then Sarah on my end will let me know if there’s anything that we can’t see. But I apologize that you can’t see me anymore, but all the important information is still there. Okay. Chapter Trust as an Adoption Barrier So the crux of our argument here is the idea that trust is one of the leading barriers to AI adoption. Here, I say the leading barrier. In some scenarios, that is the case. So trust in the outputs as well as process bandwidth, those types of things. Ultimately, what… What’s happening is folks are using new platforms or adopting AI technology, but either the outputs are not defensible, there’s not strong traceability to where the answers are coming from. We are not getting… We’re not getting specific items from a repeatability perspective. So when we’re talking about compliance, when we’re talking about identifying requirements, when we’re talking about defending ourselves when we get into submission and orals and those types of things, that could become increasingly hard. When folks don’t trust the outputs that they’re getting, then they’re less likely to adopt and use a tool. So what do we do to combat that as an organization and as a provider like Visible Threat? Ultimately, what we wanna do is we wanna give you visibility into where the information is coming from. So unlock the black box that sort of exists when we’re using some of these AI tools, have full traceability, and then give you the path to defensibility because that is the most important part of the the process. Because we can build our compliance matrix or build our content in minutes, but if we can’t see, you know, which requirement is producing those sections of outline, or we can’t see what content we’re using, who used it, what it’s produced, Ultimately, what happens is a lot of that gets rebuilt by hand. From a drafting perspective, what we’re then seeing are downstream effects. So, ultimately, the way that way that folks are structuring things now is we’re getting to a draft much more quickly. So we’re using AI tools to do summaries and overviews, and we’re getting to a draft really quickly. But we don’t have a lot of standardization around the way that the prompts are structured. We don’t have a lot of standardization of what the way that that looks in practice. We don’t have ways to verify where some of the content came from or own a piece there. Okay. I’m going to stop really quickly because I got a couple messages in my attendee chat that the sound is not working. Let me check with folks on my end. Is anybody else having sound issues? Okay. Okay. Perfect. So if you are having issues… I did have a couple folks in here. If you are having issues, I would recommend jumping out and then jumping back in, and sometimes that can reset it. If you… If that persists, I’ll make sure that you get the recording after. So I’m taking a note here of who had issues, and I’ll follow-up with you after and make sure you get the recording. Yep. And then I had a request for the meeting link. So let me go ahead and do that really quickly. Okay. Okay. And then Sarah is gonna drop the the link in the chat for other folks kind… Trying to join. Okay. All of our, housekeeping items are there. Everybody can hear. All good. Apologize for the delay there. Okay. So jumping back in, ultimately, what happens is is a lot of the time savings that we’re getting when we’re drafting in the front end get shifted downstream. So if we look at our entire process or our entire workflow, we are getting to a draft more quickly, but we’re finding more issues there. So that is requiring more work on the review and more verification, And ultimately, we’re not saving the time that we thought we were, and we are coming into some quality issues that we may not have had previously. So ultimately, what can we do to address this? So within the Visible Thread platform, we’re gonna approach that a couple different ways. One is visibility. Chapter Layering Generative and Deterministic Technology One is governance. And then, of course, the ability to layer technologies. I’ll show you a couple different options of this today or a couple different examples. So using generative AI to create first drafts, to create structures, to create outlines, but then deterministic technology to check our work. We need to have that layer in there where we know that the processes are repeatable and that we’re creating guardrails for ourselves. So ultimately, that is the underlying principle. That generative AI alone or deterministic technology alone are not the best approach when it comes to a complete process and a more holistic process. So we really look at a combination of those technologies and are very specific on our end of how we make that happen. So we’ll talk about kind of start to finish what this looks like. So what we’ll see in the platform, we’ll work through an opportunity, and I’ll show you a few examples within that opportunity. So how do we set the rules? First, what is AI allowed to do and when? So our workspaces, governance, guardrails, and stage gate questions, a bid has to answer before it moves. So ultimately, what can the AI access? Where do we wanna use AI as a part of the process? And where do we want to verify with other steps or other checks? That is built into the Visible Thread platform, and naturally, you’re getting those kind of safety features or deterministic checks on top of the AI. So building and verifying. So how do I know that this is correct? I think we’ve all had this experience that you have created an outline or created a matrix or created a, requirement shred using GenAI or using some type of tool. And then the next kind of natural question there is, how do I know this is right? Where did this come from? And ultimately, how can I defend this? So we’ll talk about what that looks like within the Visible Thread platform and the best use cases for the different types of technology to move us along more quickly and give us better results. And then is anyone actually using it well? Chapter Leadership Oversight and Usage Data What does this look like from the top down? So viewing this from a leaders… Leadership perspective. Usage data. Are we getting the most out of the platform, and are we using it in a holistic way? Activity tracking and dashboard. So are we adopting? Are we using the parts of the platform that are most important? Are we checking the specific sections so we can see, you know, when things have been moved, the stage gates questions that are being answered, when people are marking things complete. So you have a full view of what’s happening, when it’s happening, and what that process looks like. Additionally, we’re uncovering sticky areas of our process. Do we have things sitting in stages for a certain period of time? Do we have big slowdowns on approvals? Are things sitting in review for an extended period of time? So beyond just the technical piece of that, what can we fix or what can we do with the new visibility when we know where the bottlenecks in our process are as well? So ultimately here, we want to make sure that we are giving you, you know, complete visibility into the way that that we’re creating the setup. So the way that we’re creating the underlying pieces of the platform that are gonna deter… Determine that governance and security, what that looks like in the reports, how you can reasonably trace from solicitation document to output report to requirement, and make sure that you have… You can verify where everything has come from and make sure everything is accurate, and you have full visibility across all of the opportunities that you’re working in. So it’s not just your work, it’s the work of the entire team, and you can collaborate in a more effective way. Chapter Platform Demo and Workspace Scoping Okay. So I will pause there on the slides, and we’re going to jump into the platform. So like I said, we’re going to walk through. I’m going to first give you an idea of workspace scoping, creating our guardrails. We are going to do a shred, talk about customizations, talk about using a collection, managing the outline creation and traceability, and then talk about some of our guardrails on the back end, and then get an example of how we can layer that deterministic and generative technology to cover all of our bases and get the best possible views. Let me switch here just a second. I’m going to share my screen. Perfect. Okay. So I am logged into the platform, and for folks out there that have not seen VT seven recently, what we’re looking at is our home screen that every user will get when they log in to the platform. So we’re going to be greeted with some of our workflow shortcuts. So just gives us a general idea of what’s happening in the space, and then the ability to jump right to those spaces, add proposals, add opportunities, dive in, and start working. We can also check-in on activity and then navigate to the left to our different modules and configuration sections. So just a general overview of what we’re seeing and then some workflow shortcuts. So what I wanna talk about first when we’re talking about governance and security. So from a setup perspective, what can the AI see? Who can see? What? How do we segment those things? There are a few ways that we approach that from a security perspective. One of those is workspaces. So we see here that I am in my workspace in this public sector workspace. Now, a user can have access to multiple workspaces. For an example, workspaces are segmented out usually within organizations where we’re thinking about different lines of business or different divisions of an organization where the deliverables are very different, and the agencies you’re working with may be different. They could be segregated for security purposes. So security purposes, visibility purposes, we folks that do ITAR, non ITAR. They have CUI spaces, on and on and on. So ultimately, we can set up spaces in the way that works best with your business to sort of mirror how you’re working now, but also to protect information and give visibility to only the folks that need to see it. So an extra layer here. And just to show you how these are completely different, I will switch over to a separate workspace, and you see the metrics are completely different in the back end. My prompts, my my setup, my admin panels, my business profiles can be completely different as well. So here, I am seeing a completely different space, completely different activity than I did in my original. That is kind of unlimited number. I have a a massive number of workspaces, but this could be set up any way that makes sense for your team. So that is just the first kind of layer of defense and segmentation when it comes into the natural workspaces that make sense for your team, but also from a visibility perspective. So ultimately, let’s talk about where we govern. So we have our workspaces within every workspace. We’re going to see our activity and actions, and we’ll talk about this a little bit further down the line, and our dashboard and visibility. Chapter Managing Pipeline with Stage Gates So let’s talk about our stages and our stage gates. I’m going to bring us into our proposal section, actually, and look at my proposal view here. So there’s a couple different ways that I can create a view, but I’m gonna show you another way that we’re governing some of the advancement and the work within the platform. So when we first started talking about pipeline management with our customers, one of the biggest questions was, or one of the biggest pieces of feedback was that we need visibility, we need oversight and governance on where things stand, who’s involved, when they’re being moved, and what that looks like. So that is the reason for the stage gate process. Now this is completely customizable. We want to make sure that the views here are going to, mirror your process as directly as possible. So all of these different gates are completely customizable. So this is just my general process. But regardless of what this looks like, if you have additional, additional segments, different naming conventions, whatever that looks like within your team, that can be set up here as well. Now if I say, okay, I’m working on a specific, piece, and I want to assign this out. I’m gonna be working here, but I also want to assign this to Emily and Urkus. I can assign the… A specific proposal to different folks. They are then going to get specific notifications and callouts and updates on what is happening within this opportunity. So that’s just another way from a visibility perspective to make sure that everybody that needs to be involved is getting this called out for them. Now if I, say, had gone in and this is in progress, work has been done, I want to move this over to Red Team, I’m going to get this gate kind of gate progression piece pop up. So it’s telling me, I’m moving this from in progress to red team. Do I want to skip the gate questions? This is because this is set as a soft gate. So you have the option to set the soft gate, which means I could choose to skip the questions, or it can be set as a hard gate, which means these are the things that have to be answered. Now, in my example, very simple, just sections have a first draft, have we addressed the compliance matrix, and leave the status on graphics. So I can answer and then move to red team. Now, those are the things I have to answer when I move. How do you know that that happened? How do you know what the status was? And how do you know who moved this over? That all happens in our activity. So, if I come back to my dashboard and I look at activity, it’ll show that I had a stage team, that I moved in progress to red team. And you can see throughout section all of the things that are happening. So if I am creating documentation, if I’m assigning specific sections, if I am assigning different pieces of the process, I am then going to see that in the activity space. Ultimately, this is the way that we’re going to be able to track everything that happens within the opportunity. You’re also going to see that at the opportunity level. So at the opportunity level, we can see that I assigned ownership. So assigned ownership to Emily and Urkus, and then I changed the stage, and these were the answers to my questions. So those can be open forum. They can be any structure. That is a really easy build on your end and just to show from a settings perspective. If I go to pipeline stages, you can see the setup for the cards. So we can see all of our gates in our tracked opportunities, all of our gates in our proposal section, and the questions that are there. So we can do yes, no questions. We could do open ended. So this can really mirror what your process looks like today. So what are the questions that you’re asking, and what are the questions that your team needs to answer before progressing something? What are the things that need to be done? What artifacts need to be created? And we can work that into the process. So not only do we have to account for those things, they’re easy to check and easy to govern on your end because all of that is super visible. So talking here about how we’re managing the pipeline, what that looks like in our workspace and where those areas are from an oversight perspective. I wanna continue on this path and talk about, okay, we are now in an opportunity. What do we do there and what can we do from a security perspective and then from a visibility perspective? Chapter Deterministic Shredding and Requirements So this is the first place that we’ll talk about the combination of our technology. Ultimately, here I’m going to talk about shredding. So if we’re talking about shredding, we’re talking about compare, that’s where we’re going to apply our deterministic technology. We’re using our ditch… Our search dictionaries, and we can see every item that we find with 100% accuracy. So if I’m going to run a dictionary report, that is repeatable. Every time I know I’m gonna get the exact same thing. So if I’m looking for specific contractual language, if I’m looking for specific clauses, I’m going to be able to find that every time as opposed to some of the more analytic processes, or we’ll talk about, outline creation in a minute and what that looks like. So ultimately, it’s just the combination of the two technologies. So I’ll start here with shredding, talk a little bit about how we can customize and how we can create an experience that is accurate to what you’re doing today. So I’m going to name this, and then I can choose a template. Going to choose a basic template, but ultimately this allows us, and if you are… Have used Visible Thread in the past, this is likely new to you. This is just a new feature that came out a couple of weeks ago. Ultimately, now you have the ability to create a multitude of different matrices. Similar experience to folks that may have been running a macro in the past to change an Excel output. Now we can manage that right within the system. So I have a multitude of different examples here, requirements traceability. My cross reference matrix, and then I’m just gonna use this general compliance matrix with status, page allocations, customer dependencies. And there are a multitude of things we can do here as far as grouping and sectioning and the way that the data rolls up and on and on and on. But this just gives us the idea of what we’re looking at. So I’m gonna then choose this dictionary, choose a document. All of the documents that are a part of my solicitation package will be here. I can confirm that and then run the shred. Then when we see our output, we have kind of what we know and love from a visible thread perspective that we have all of our language, our will, shall, must, ensure, require, responsible. We have our prohibitions. We have some deliverable language. We have some compliance performance language. So this is what we’re using to break this down. The beauty of Visible Thread is that we know that not a single one of these will be missed, not a single one of them will be misplaced or positioned incorrectly. And now, we can then begin a review and start to highlight areas where we’re compliant, areas where we need to work through. We can leave notes for folks. We can call out our different users in the sections. So I can call out Emily who’s assigned to this, and then she’ll get a notification that I have a review or I have something that she needs to look at. So ultimately, we have created a situation where you do have your same deterministic output. So same deterministic output, but instead of exporting that into an Excel and having everything be manual, we can work those columns right in the platform. We can leave notes, and we can collaborate with folks. We can name check people on things. We can call them on things to make sure that they’re checked and reviewed. So we… At that end, when we’re doing all of those things in the platform, that gives us direct traceability into our activity section. So from a safety and visibility perspective, this takes away kind of the danger of we export it out of the system, everybody’s working from a spreadsheet, and it gets passed around. Do we have the most recent version? Have there been any changes and have it been recorded? On and on and on. Here, we are tracking every move that you’re making from a review perspective, from an assignment perspective, from a feedback perspective, right in the platform, and you can see that in your activity. Chapter AI Outline Creation and Traceability Now, I want to jump out of here because I want to move on to flip the switch a little bit and talk about the outline. Ultimately, I’ll show you how this is created before I jump into the full outline. So I’m going to use an example here. So if I have an opportunity, so in this sense, I have an opportunity, I wanna generate an outline. This is where Gen AI is going to come in, but then we’re gonna show… I’m gonna show you what the safeguards are there and how we are governing this in the backend. So I want to generate my outline with AI. Now, all I need is to select my document or multiple documents. You can select multiple documents, single document, just whatever is important as part of the solicitation package, and then select next. The system is going to start working. So the first thing that it’s gonna do is it’s going to identify the document. So it’s automatically identifying the document type, and that will inform some of the work that it’s doing to search for requirements and creating the structure of the outline. Now what it’s doing is it’s it’s detecting and starting to extract the requirements. You’ll see that when these start to come on the stream… Screen that it’s going to extract requirements as well as type them. Again, the extraction structure and the typing all goes into what the final outline structure looks like and how the requirements are grouped. So you see we’re automatically starting to pull out requirements and then typing them, technical management, past performance requirements, on and on and on. So this can… This process will continue. I won’t make everybody sit here while it does, but this process will continue until the outline begins to be structured and then is supplied. So that’ll run-in the background. I will jump in here back to the opportunity we were working with and look at one that’s already completed. So this outline’s already completed, and I already started to do some of the work on it to show you what that looks like when it’s a little bit populated. Now looking at our left hand side, we have our section names, and all of our kind of sections and levels throughout. So we have 64 sections. We have five levels in this specific outline. Now you have complete control over this. Again, when we’re talking about governance, we’re talking about control and safety when we’re working with AI. So this was generated by JetAI. You have complete control over the structure. So you can edit this. We can change levels. We can change orders. We can add sections. We can delete sections. So you can manage this on your own right in the platform. This is nothing that’s set on stone, it’s very, very easy to change and modify, particularly as you move forward and your approach may change. So I want to cancel that edit, come back in, and see what we’re looking at. So we talked about the structure here a little bit. Now I want to talk about the traceability. So this is a huge part of the puzzle when we’re thinking about safety and defensibility. Now, I see my requirements. When we were… The requirements were being extracted on that front page when the outline was being created, we saw that the requirements were being typed, traced, and grouped. This is where that comes into practice. So I will select one of our trace requirements groups, and what this is telling me is that these are the requirements that make up that specific section. So these are the requirements that make up that section. These are the requirements that need to be addressed in this specific section or deliverables that we have here. You can also, if I select one of these requirements, jump straight to where we’re seeing that in the solicitation. So we went from a generated outline, which was I’m going to have Jen and I create an outline for me. That… You’ll see that in a lot of platforms as well, that we’re gonna create a general outline. You get back, you know, a word document with general structure based on, you know, section l and requirements, we’re taking that quite a step further because we want, number one, for this to be live and active and for you to be able work in the system. But number two, I’ve gone from a generated outline immediately to my source material in two clicks. So that allows me to ensure that the requirements that are making up a specific section are accurate, that I don’t have any misses. And then if I wanna open it back in my coverage section, I can see this for my entire outline. So I can see the entire outline, all of the sections and segments of my solicitation, of my source documents, where things have been marked as out as… Excuse me, as requirements, and how that shows up in the source material. So you have direct traceability to your source material all the way up to your outline and delivery items. Let me jump back to my outline and we can talk just a little bit on what this looks like. Okay, now back on my outline. I’m going to zoom in here a little bit for us. There we go. Sorry about that. Okay, zoom in so you can see this a little bit better because I know there’s a lot of information on the screen. Okay, so looking at our outline. So we talked about our section and our structure, traceability of requirements, and then where do we go from here? Our next step, obviously, once we feel good about the requirements, feel good about the structure, is to begin, you know, begin assigning out and creating content. Once again, you can do that right in the system. And being able to do that here does give you another level of visibility and control over the status of certain items, any type of issues, and the ability to give directives all in one place. So I can say, I want to assign a specific section and we’ll look. So I’ve assigned this section, say I want to assign an additional writer here, and then I want to leave a directive. I’m going to have Anne write on this, but I also want to leave her a note. I want to say. Now I am asking her to reference a specific collection and can give directives on positioning or documents to reference or other pieces to reference in this situation. So I can do that right in our section. Chapter Content Drafting and Style Templates I can then say, okay. We’re going to start drafting, move this into space, and then I can start writing. So what I wanna do is I want to start writing. We have a couple options here in the platform or right within Word with the add in. I’m going to select this so you can see the templates. Another thing… Another way that we’re sort of controlling our outputs are with the style templates. So style templates are your own word templates. So when you create the content, it’s coming out in the, you know, correct formatting, spacing, font, etcetera, etcetera, So you can use your templates here and then begin to work right in Word. It’s gonna create that for me. Let me open it up. And hold on just a second, Okay. So this is what the Word document looks like when we start working in this part of the platform. I’m just going to sign in to my add in, and then everything that we have in the platform is available in our word add in. Now several aspects of the content creation platform when we’re talking about the way that this is managed in Visible Thread are governed by some of our underlying systems or the organization has influence here. One of the biggest things that we hear as a piece of feedback is that a lot of organizations are using AI in some capacity, and a lot of times that’s we have access to Copilot or access to internal GPTs or different platforms where where, you know, our writers are using Gen AI or our SMEs are using Gen AI to create content. What happens is is that historically, before anything else was involved, and this was a manual process, it was a challenge to put together a multi contributor response because we have different authors, different tone of voice, different acronyms, on and on and on. That problem can be amplified a bit with Gemini because if people are prompting in different ways, they’re creating different structures for their outputs. You are creating content more quickly, so you’re scaling some of that dysfunction. Chapter Standardizing Prompts and Collections So that historic problem has kind of been amplified with, generating created content. So within the Visible Thread platform, we work with our folks to create their prompt library. Prompt library means that I have my outcome shortcuts here, so I want to create a draft and RFI response. I want to do a requirements response, do a rewrite, create an opening paragraph. Paragraph. I I wanna wanna do a, you know, capabilities overview, a quick executive summary. These are outcomes or these are actions or things that I want to create. Well, if I’m doing this within the platform and I am creating a capabilities overview, or one of my colleagues is creating a capabilities overview, if we’re all using the same prompt from the same library that we trust the output for, then we have at the beginning already jumped ahead from a structural perspective. So it’s a way that we can standardize the content that’s being created in GenAI as well as test test what’s working best for us and upscale other users and use this to, you know, the best capacity possible. On top of that, we are going to ground in a collection. So let me go ahead to highlight this, and then I’m just going to respond to this requirement. I wanna respond to this requirement, but I want to use a collection. The collections are groups of documents pulled in either from your local environment or SharePoint. Any SharePoint folder that you have access to, you can pull straight in, and it allows us to group documents in ways that are going to be really effective to use as source material. So I’m going to continue with this collection and send. So I am referencing my collection. I’m referencing my documents. It’s gonna ask me a couple questions. I want this to be brief because I don’t wanna spend too much time here, for demo purposes. But ultimately, what’s happening now is the system is using Rag to identify areas in our source document that are relevant and then using the LLM to shape and structure our response. What we’re going to see when we start getting the output is that these are all referenceable. You’ve probably seen this type of thing if you’ve used other platforms to generate content, but everything is referenceable down to your source document. And we can select those reference items and go right to the spot in our source documentation where we pulled that information from. So you’ll see these as they are coming through. And when I click on those, it’s going to pull me straight to the area in my capability narrative documents that the information was found in. So we’re attacking this in a couple different ways. One, we are using standardized prompting and a standardized workflow to create content. We’re not just free prompting in the chat box and turning that into a back and forth. We are using a standardized system to create content, and then we are grounding that content in our trusted data to get a high quality first result, and then giving you referencing to make sure that these items are verifiable and defensible. So protecting safety in a few different ways. Then the last piece of that puzzle is… Oh, and let me go back. Sorry. I’m going to go ahead and insert that. So the last piece of that puzzle is the readability spot of it. Chapter Quality and Readability Checks So let’s talk about reintroducing some of those deterministic pieces. So we’ve created this content. Does it meet our standards? Now, as an organization, we are going to decide how… What’s a long sentence? How much passive voice is too much? What kind of readability we’re going through? What grade level are we writing at? Are there specific terms that we want to ignore? Are there specific terms that we want to include or things that are red flags for us or language that we want to use that’s customer language or industry language. So all of those things are factored in here. So ultimately, we have the ability to leverage GenAI in a really great capacity by using, Gen AI to create content directly related to what we need from our outline. But then we also are grounding that in our content, and then we’re layering in a quality layer to make sure that the the Gen AI content is meeting our standards and is matching from a language and a structure perspective. And, again, all of that in your templates and your your structure. Okay. I’m going to jump back into the platform and talk about a couple more things. And then we’ll finish up. Okay, perfect. So the last thing that I want to touch on is we’ve talked a little, we’ve touched on a few different things when we’re thinking about the combination of technologies and creation. I wanna touch on one more example of the layered technology, and then we’ll talk about the dashboards, kind of end on dashboards and visibility. So let’s talk a little bit, and I’m going to do a… Chapter Document Comparison and Material Changes Here demo. And we’re going to choose a document… Choose two documents. If you’re familiar with this process, if you’re a Visible Thrive user, you’ve probably seen a compare before. If not, what we’re doing here is we’re taking two documents. In this instance, we’re going to do a solicitation, and I’m going to do the amendment of that solicitation. Now, these can be two versions of a response. These can be two versions of internal documentation. Anything that, you know, we’re going one to one on, we can use here, but here, I’ll use it for this instance. I’m going to run this report, And again, this is deterministic. It is beautifully structured and really easy to navigate. Now I can… I see all of my changes. I can drill into specific change types. I can begin reviews and mark these as, high, medium, low. I’ll start a review. I can mark these. I can leave comments. Review. And then, you know, if I’m leaving comments for folks, they get notifications and get notifications that they have specific things they need to review. We can work through that process and go kind of one by one and take a look at all of our comments and all of our changes. Now what we can also do is here, of course, we have our one to one. So we have paragraph to paragraph, perfectly structured, leaving room for deleted spaces, every change that’s in this document. However, we’d like to get there a little more quickly or get a general overview. This is where I can start to layer. So instead of having to go down and check one to one, I can then bring in Visi everywhere. So it’s gonna give me some suggestions here. Anything change? Any blah blah blah. But I just wanna ask what are the material changes? And you can see in these, it’s going to suggest items. So, original and revised wording, so we can jump in, jump in different spaces. So, essentially, I asked are, what are the material changes? So, what are the things that I need to know that are most important? And it’s calling out for me exactly what those things are, the exact type of things that you need to see. So, the set aside changed, the NAICS code, and size standard changes, our response date changed, and the period of performance changed. So the contract type staffing, there are some very specific changes in this amendment. So ultimately, this allows us to layer and get the best and the most out of technology without losing our accuracy. What we have on the left, our traditional compare is the source material, and this is the 100% accurate report of what has changed from one to the next. Here on the right, we can then drill into what are the things that actually changed, and then start to ask questions about what does that indicate. Are we seeing trends over time? Should we change, something in our response, or should we change something from, a positioning perspective? All of those things are things that you can you can drill in and do. So you’re using Gen AI where it’s helpful, and you still have your guardrails, which are which are our outputs that are showing us our one to one changes. Okay. So I’ll go there. Last thing that I’ll talk about… I’m going to exit this, exit and discard. The last thing that I wanna talk about before we finish up today are quickly to touch on our dashboards. Chapter Dashboards and Activity Tracking So the dashboards are really important when we’re talking about that visibility. So everything that we showed today is going to bubble up to either dashboards, notifications, or activity. So here, we can see actions. You have full visibility into the way the platform is being used, different types of activity, how often we’re using Vizi, what our token usage looks like. So across the board for every workspace, everybody in your workspace, can see the workload distribution, you can see how things are being used, and you have full visibility into what that looks like. From a notifications perspective, we saw that I went through and did a few things. I assigned some things to some folks. I commented and mentioned them in those areas, and then they will get notifications that they have outstanding assignments and outstanding comments. So that is going to bubble up into our notifications. And then finally, our dashboards. So the dashboards are really powerful for a couple different things. Number one, yes, we see what’s in our funnel. We see what we’re managing. We look at best fit. We look at all of our timelines. Those are all great. We also see who is owning this. How long has it been sitting in stage with our archived opportunities? Do we… How many no bid disqualify one loss? What does our what do our per win percentages look like? All of these items are here, and all of them can be drilled down into to take a look at number one, what’s happening. Number two, where are the issues? So if I have things sitting in stage too long, where are they getting hung up? Why do I have a bottleneck? Is there a process I need to fix? So you’re getting visibility into that whole process. And then finally, looking at the proposal dashboard, same idea, but it’s going to drill down even further. You saw me in the outline assigned some sections. So I assigned some sections to some writers. That is all, managed here as well. So we can see section assignments. We can see section status. So I can see that my proposal’s in process. I have one hundred and eleven secondtions. We’re only 6% complete because we have a lot of stuff that’s still in review. So that’s a good question. I have a lot of stuff sitting in review here. I also have a really disproportionate assignment from a section assignment perspective. So are those two things related? It’s just the data can roll up to give you visibility into some of the issues that you might be having. And then, of course, due date timelines, activity notifications, and of course, here we have our archive proposals as well. So we can see our win loss spaces and drill down drill down on those a bit further. And we can do over time custom ranges. So you have a lot of control over what you’re seeing here. Beyond this, there are a multitude of other things in the platform that can be customized and that can help you surface items that are gonna be important to see. The last thing I’ll touch on is the activity. You’re getting activity at every level. So activity at the the proposal segment level, activity at the actual opportunity level. So if I’m assigned to this and I’m working, I can come in and I can see everything that’s happened on this opportunity, all the reports that are generated, all of the information that’s there. And when I come back, I can see in this list view, I have my opportunity. I see I have 30 documents. I have nine shreds. I have an outline. I have 13 proposal documents in process. I’ve done an acronym report. I’ve done some compares. So I can see everything that’s there. That’s another piece of the puzzle from a visibility and collaboration perspective that we’re not doing rework, that we’re working off of the documents that are that are kind of our master documents. And then you can see from an activity perspective who has done what and what has progressed. So I will stop there. I’m going to jump back to my slide and just wrap things up here. Okay. Okay. Chapter Summary of Practitioner and Leadership Benefits So let’s talk about kind of what we just saw. So thinking about the… From a trust perspective, a practitioner wants to know, you know, can I trust this result? So the… All of the things that we just talked about kind of come together to make sure that you can trust the result and that you can verify it when needed. You know, your gate questions are gonna keep the bid inside the process. So we make sure that we’re hitting everything, and that there… We aren’t skipping any of our steps. The traceability shows you which requirement produce each heading, and where the requirements are sitting within our segments. Our collections and approved prompts in our prompt library are gonna ground the draft in your content as well as standardize the way that content’s being created from a structural perspective. And then, you know, the… Using our compare, using our watchwords, dictionaries, readability, that’s going to make sure we’re verifying the same way every time. There are so many parts of the process where when we bring generative AI in without the guardrails, the processes can sort of morph. They just kind of Frankenstein a little bit, and they’re often not repeatable. So making sure that we are using that combined approach to verify that we’re doing the same checks every time, the same structures every time, and that this is something where we can use GenAI for what it’s good at, but then everything has a guardrail or a safety check. So from a leadership perspective, the question is, you know, can I trust this rollout? Can I trust this to do the job, you know, I brought it in to do? So guardrails enforce policy in the tool. So this isn’t something secondary to the tool from a governance perspective. All of that, every action in the platform is being bubbled up and traced to make sure that we can enforce governance at the source. The usage data shows who’s adopted it and who hasn’t. So that’s giving you everything from a usage perspective and from, you know, a training opportunity perspective. There’s a lot of information we can garner from that usage data. The activity tracking really shows what the AI is actually being used for. So in the beginning, we make a determination on where we want the AI to be a part of the process and where we want to rely on some of our deterministic pieces, and then you can verify with activity that that is what’s happening, and we’re using things the way that they’re intended. And then the dashboards are going to show you the results per pursuit. So we know where to expand. We know where to train. We know what’s working. We identify bottlenecks. We identify sticky parts of our process and really can optimize our investment. So ultimately, same kind of views, but different viewpoints. So this satisfies things from a leadership perspective and a practitioner perspective. Because of course, you know, leadership wants things to work and they want it to work effectively. But from a practitioner perspective, if you don’t trust the results and don’t trust the quality of the content that you’re creating, then that is, you know, we’re not even gonna get to the the leadership point and the real, engagement of the platform. Chapter Platform Visibility and Governance Summary So that is kind of everything I’m covering today. Kind of the takeaway is the fix for trust here is visibility, and it has to work for everybody. And that is a key part of what we’ve built into the platform. Governance where you need it, oversight on all of the work that you’re doing, the ability to communicate, and the ability to use the platform as a source of truth so you don’t have silo… Siloed work and everybody can see what’s happening on a specific bid. Now the last thing that I will cover… Do we have any… Let me see here. Oh, and I did. I had a question. I know we’re over, so I will follow-up directly on the question that I had come through. And if you have any additional questions on functionality… Now I know there are a lot of pieces of puzzle kind of jumping around talking about the governance. There are a lot of, working pieces and customizations that we didn’t get to show. So if you have, any follow-up questions, would like to continue that conversation, always please connect with me on LinkedIn. I would love to chat. Love to meet with you and answer any questions that you have. And then, of course, our content library and our events page with webinars. Lots of great information there as well. And then, of course, if you are interested in seeing more about the platform, we can accommodate. So please reach out to us. Reach out to me with any questions. Chapter Upcoming AI Evaluators Session And, again, thank you so much for joining. Like I said, we’ll be back in a couple weeks, and we will be talking about AI evaluators. So writing for AI evaluators and how some of these pieces of the puzzle and some of the deterministic pieces can help with that. So, again, have a wonderful rest of your Thursday, a wonderful weekend, and thank you for joining us.