Video: Vanta MCP 101: Getting Started | Duration: 3520s | Summary: Vanta MCP 101: Getting Started | Chapters: Welcome & Introduction (16.575s), Welcome and Introductions (137s), Session Agenda Overview (230.06s), Comfort Level Check (354.44s), Meet Sam (411.23s), Live MCP Demo (543.255s), MCP Core Concepts (778.505s), Use Cases & Examples (1016.67s), Why MCP Matters (1276.865s), Connecting to MCP (1555.025s), Prompting Best Practices (1914.01s), Prompting Best Practices (2439.065s), When and Where (2633.17s), Session Recap and Completion (2888.275s), Next Steps Announced (3133.69s), Next Steps (3246.74s), Wrap-Up & Next Steps (3365.605s)
Transcript for "Vanta MCP 101: Getting Started": Hello, everyone. Welcome. I'm laughing at the GIF in the chat. Welcome to MCP one zero one. We're gonna give it probably, like, at least a minute or two for those that may still be rolling in. So, yes, please share all the all the GIFs and emojis in the chat, and we're so glad you're here. But we're probably gonna give it about one more minute before we get started officially. Hello, and welcome if you're just joining us. We are getting started here, shortly, but I'm just giving a a minute or two of grace period. But welcome to Vanta MCP one zero one. So many great gifts, and emojis coming through in the chat, so please keep sharing. And if you wanna take a second to introduce yourselves, I would love to know who's here, where you're joining us from, if you wanna share the organization you're at. That's always great to see as well. But, yeah, let's give it about one more minute, and then we'll officially get started. Alright. Should we jump in? Hello, everyone, and welcome to Vanta MCP one zero one getting started. I'm Geralyn, the program manager of technical education here at Vanta, and I am so glad you're attending today. Thank you so much for being here. I see a ton of folks in the room, hopefully, more even joining us, here soon. So great to see you all here today. Special shout out to our VantaXchange community members who are here. If you don't know what that is, you will know what it is by the end of today's session. But, I do think we have some special folks from there today, so I just wanted to make sure to give them a special shout out and warm welcome. Let's take a minute to introduce ourselves in the chat. So, you know, say your name, where you're joining us from, your organization if you feel comfortable. I love to just know who's here and to see where in the world you all are. I'm in Baltimore, Maryland. So any DMV folks, please make yourselves known. We'd love to to see, where you're at. Okay. Great. We have some Atlanta folks here. We have, London representing. Wonderful. That's so exciting. We do have some Vanta folks in the chat who all kind of give a special shout out, to them here in a minute, but great to see them. Greece. Wow. Amazing. San Diego, San Francisco. Great. All over the place. This is wonderful. I'm so glad you all are here. Okay. Let's quickly chat about our plan for the next hour. This is truly a one zero one session and is designed for those beginners in the room. But if you do consider yourself more intermediate or advanced, I'm also really glad you're here to help, you know, share ideas, insights, best practices, and more. So all are welcome. I also wanna call out that you don't need to connect or be connected to get value out of today. You know, like I said, this is designed for people that haven't connected or used our MCP yet. So if you can connect along the way during our session, that's wonderful. If you learn things to take back to IT or legal for approval, also wonderful. And if you're just gonna observe, take notes, ask questions, save things for later to try once, you connect, that's also great. Here's our agenda for the session. We're going to answer four questions in order. Very one zero one friendly here. What is this thing really? Why does it matter to your GRC program? How do you use it well? And when and where should you actually reach for it? So every section is gonna answer those questions before we move to the next one. And since we do have a big room today, we have a ton of people registered, a ton of people here, which I'm so excited about. I'll be driving for most of the hour, but I built in some moments where you'll be able to participate in the chat and answer questions, share ideas. So please make sure to engage. It seems like we have a crew ready to engage in the chat, which I'm so excited about, because that makes training, so much more valuable. So Please make sure to participate along the way. We'd love to hear from you. A special shout out, we do have some customer education and digital success folks from Vanta in the chat here. You'll see KC is here. Coming through is Caitlin Morris, Jacob, and Jaquez is here. So make sure to DM them or reach out to them if you want to connect with them, ask questions. They're here to help and to see you. Okay. Let's do a quick pulse. So this is just a little pulse check before we start. You can be a 100% honest. There's no wrong answers. On a scale of one to five, how comfortable are you with the Vanta MCP right now? So one would kind of be like, I register for this because I have no idea what it is or how to use it. And five would be, I already use it every single day. So go ahead and put a number in the chat and share. These are exactly the numbers I was hoping for. This is amazing. Okay. This is a perfect session for you all then. We'll see if those numbers change by the end. Hopefully, they do, but I'm really excited. You are in the right place. Before we dig into our first question, which is gonna be the what, I do want you to meet someone here who is a nondescript, member of our GRC community. This is Sam. So Sam runs security and compliance maybe as the whole function, maybe as part of a small team, which means Sam is, let's be honest, probably similar to many of you in this room. Sam's gonna have a scenario that we utilize throughout the whole session. So let's go ahead and jump into, what Sam is going through and learn more about that. Might sound familiar. Okay. It's a Monday morning. Sam is looking at a screen that has Vanta in one tab, his email in another tab, Slack is blinking, there's a spreadsheet open from last quarter. And before the first coffee's finished, four asks are already coming in. Okay? One, leadership wants a SOC two status update before a board call. Depending on your role, that ask might be coming from your CEO, your auditor, your security team, your biggest customer. Two, a handful of employees are overdue on security training again. Three, Sam still needs a plan for the week. And four, the security lead wants the biggest open risk identified and logged in the risk register before the quarterly risk review, which normally means taking through findings and then handwriting the risk entry. This seems like a very stressful Monday. So I wanna know, like, what in this scenario is sounding familiar to you? Who's asking you for these things? You know, who's sliding into your DMs first thing Monday with, like, a, hey. Quick question. You know? Is I see someone shared already. This sounds very familiar, so that's great. Anyone else identifying with maybe one or all of these problems that are hitting Sam's computer? Wonderful. Great. I figured this might be, something you all could identify with. So, you know, it's tough to answer all of these asks and get the work done that you need to do, right, which is where we come in. None of these asks are hard necessarily, of course. Right? That's the frustrating part. Everyone is, you know, look something up, turn it into words, send it to someone. But, you know, every single one of these might cost, you know, fifteen to twenty minutes if we're lucky of tab hopping. And it's not just us. Right? Statistics show that industry wide 92% of teams pull audit evidence from three or more tools and most spend five plus hours a week on manual compliance work. Here's what I want to do. We have the whole hour to kind of work through this list together in addition to what we're gonna learn about VantaMCP. So let's see if the MCP can maybe knock one of these out for us right now, which selfishly is a sneaky way for me to demo the MCP for you before we jump in, to show you how efficiently we can tackle this to do list. So let's start with number four, the risk one, which I think is the neediest, and so we can start there. Yes. This, GIF about herding cats is so so accurate. Okay. So watch this. I'm I'm gonna do something with Sam's morning that I couldn't have shown you even six months ago, and then we'll jump into the what, and I promise we'll roll from there. So I'm gonna switch over to my Claude screen. I will be working in Claude throughout this session, so you will, see that. I know Claude was having a little bit of an outage earlier, so we're gonna just pull up my preloaded because I was planning ahead just in case something was going wrong. So this is my prompt. I said, look across my Vanta account. What's my single risky riskiest open item right now and why? Just to call out, this is a demo account, so it has a lot of problems. Hopefully, yours doesn't look quite like this. But it did give me one, which was saying that I have no root user access keys exist, for AWS. So it's kind of telling me why it's flagging this, things that are making it worse, what it's compared against. So it's answering that prompt. Right? But then, you know, I found it. So now what am I gonna do with it? Right? Which is sort of the second part of what Sam has to do. I said create a risk scenario for it in my risk register with an initial likelihood and impact in a short note on treatment, then give me the link. So I ran this, and it did. It did ask me a follow-up question about what risk register to put it in, and then it gave me the link and kinda gave me some insight on it. And if I pop over to Vanta, I can show you that it now lives here. I have it all pulled up, ready to go. All of that happened using my Vanta MCP in Claude. And it's it's ready. It's good to go. This whole sequence, even though I preloaded it just in case there was an outage, probably took maybe two to three minutes. It was really fast. So it didn't just find my riskiest item, it filed it. The risk entry now exists in my actual Vantorisk register created from a conversation in about forty five seconds. You know, I can delete it now, of course, because it's a demo account. So, you know, I still have control over that. So, you know, that's kind of what we're here to learn. Right? Is look at all these tasks we have on our plate like Sam does, and how do we start tackling them with the Vanta MCP? So Sam's Monday is off to a way better start now with this workflow. That's ask number four, done. The biggest open risk identified and logged in the register in a minute or two. That's a tool we're here to learn. The other three asks on Sam's List, we're gonna get to those soon, but let's go ahead and jump into the what to learn more about the Vanta MCP. Okay. So what is this thing, really? MCP stands for model context protocol. It's an open standard, and I want to stress standard. Right? It's basically how the entire software industry now is letting AI tools talk securely to other software. Think of it as a universal adapter between an AI tool and the software where your real work lives. Your AI tool can hold many MCD connections at once, like your docs, your ticketing system, your chat tool, Envanta. So you are able to kinda bring all your tools into one central location. The Vanta MCP is one of those connections. It is a secure permissioned line between your AI tool, whether that's Claude, Cursor, Perplexity, or another MCP compatible tool and your Vanta account. So here's the distinction that matters most because some of you may have already gone to Chat GPT or Claude and just ask a general compliance question. Right? Which is great. That's a great use of it. An AI tool without the Vanta MCP answers from general knowledge. Right? Like, if you say, am I ready for my audit? It might give you, like, a competent generic answer from about audits in general. Like, to be ready for your audit, you should check these things. You know, it'll kind of give you that very generic, would be relevant for anyone answer. It's guessing kind of. Right? But the same tool with the Vanta MCP answers from your controls, your tests, your people, your documents right now. So it's if it isn't connected, the tool's guessing. Right? If it is connected, it's reporting, which is a much different, more powerful way to use your tools. So what can it actually reach? This is a map to kind of give you a sense of what all it can touch. Spoiler alert. It's basically everything. So on the same side, your frameworks and controls, tests and evidence, policies and documents, people and access, vendors and risk, you know, questionnaires and privacy, audits and vulnerabilities, integrations and issues. The short answer is if it's in Vanta, you can ask about it. And from there, it does five different kinds of work with that access. So you could do status questions like how's our ISO 27,001 going, summaries in any voice for any audience, cross referencing across your whole program in one ask, drafting and acting, turn your data into emails and reports, and increasingly writing back into Vanta. You saw one of those rights in the opening demo. And for those of you that are really new to MCP, when we say it has the capacity to read, it basically means it can go look at your Vanta instance and then answer questions about it, report back to you, provide these summaries, these status updates, or it can work with you to create reports and and such. When we talk about right, it means your tool can actually go back into Vanta and complete actions. So just wanna make sure everyone's on the same page about that. Finally, the last thing to note is it can do cross tool chaining. So if your AI tool is also connected to Jira, Slack, Google Drive, Notion, etcetera, Vanta findings can flow straight into a ticket or a post or a page. Vanta kinda becomes one node in your AI tool's whole graph. For today, we're staying inside Vanta, but that chaining is what the weekly builder workshop teaches, so consider joining that if you're interested in learning more. Okay. So that's kind of the general map, but I always like to plug a little bit of, like, concrete use cases. So these are five favorite things for me that I I like that it can do. And we know some customers are using these exact prompts today because they share and talk about it in our community. So one would be the contradiction check. So this might be a prompt like, do any of my policies contradict each other? So, yes, it could actually answer that for you. It would read across your entire policy library and flag where documents disagree. No one ever expects this to be possible, so it's a prompt that people usually get pretty excited about. Two is like an auditor answer. You could prompt it with an auditor just asked me x y z question, pull every policy, document, and past answer I have that addresses it. Your whole knowledge base will be interrogated in this one sentence, which is amazing. And this kind of leads into our, you know, part two of this MCP training series, which is in October. So a little bit of a teaser there. Three would be a subprocessor list. So one of our community members built their complete subprocessor list by asking the MCP to cross reference their vendor list against their privacy and data management policies. You know, that's a task that's usually, you know, so time consuming, a lot of spreadsheet archaeology, and then it's done in one prompt. Amazing. Four, the evidence pre check. So this could be something like check this evidence against the request before my auditor does. It's kind of your own pre audit safety net. So I really like that one. And then finally, the audit day snapshot. This is another community favorite. So this one would be a prompt like pull my whole program into one dated Excel file, risks, controls, vendors, people. Auditors really love that point in time evidence. Right? And this is a dated copy of everything refreshable in minutes before every audit. And honestly, it's also just your data in your hands anytime you want it, which is exactly how it should be. So these are five favorites, five completely different kinds of work. And just to note, a couple of these lean on specific product areas. You know, the auditor answer obviously needs questionnaire automation. The sub processor list needs your vendors in Banta, but that's kind of the pattern. Right? The MCP will mirror your Banta, including what's in your plan, and we'll talk about that more in the how section. Let's keep rolling. Okay. A few capacity notes that we should cover kind of let in from my last sentence there. So the Vanta MCP is permissioned always. So every role connects with its own access and sees only what your Vanta permissions allow. So security team asks you about that, you can share this with them. It reads everything and it writes a growing set. So over a 100 tools today, you watched a right action in that opening demo for the risk entry. Rights now span risks, policies, vendors, privacy records, issues, and there's more rolling out all the time. It's also bounded by your program. So if you ask about a product area you don't use or haven't configured yet, you you'll get an empty answer, which is by design. So, you know, that's functioning correctly. And just remember, of course, that it's a draft or not an authority. So answers are always grounded in your data, but you should always verify before you forward to anyone, and we'll chat more about that in the how section. Okay. So any questions, thoughts, ideas, how are we doing in the room? You know, give me an emoji, thumbs up to let me know you're still with me. Wanna just check-in how we're doing. Love it. Thanks for sharing the emojis. I like to make sure everyone is still here and engaged. Great. Okay. Let's let's keep rolling. Okay. So that was the what. We've covered the what. Now we have to move to the why. So why does this matter? Why should you care? Maybe a more interesting question than the what. So I wanna answer the why with kind of thinking back to our Sam scenario a little bit instead of giving you a feature list. So there are really five problems that this can solve for you, and I'm guessing at least two to three are on your plate right now. You can throw a number in the chat list now. So one is the reporting tax. This is a problem that it can solve for you so easily. Somebody is always asking how it's going. Your CEO, your board, your auditor, a customer security team. Sam is dealing with this as one of his problems on Monday morning. Each answer requires you to gather, screenshot, reformat, translate. Remember those stats we talked about? 92% of teams pulling evidence from three or more tools, five plus hours a week of manual work, and 62% say their evidence gathering is at least occasionally error prone. But this ritual is one of the biggest recurring time costs in GRC, and custom reporting is the single most common thing we see customers actually do with the MCP today. It helps to solve that problem. Great. Two, the context switching tasks. You might have five tabs open to answer one question. With the MCP, the question goes to one place and the data comes to you. And, you know, these aren't hypothetical solutions. At our Vanta Delivers launch this month, how many of you watched our Vanta Delivers, saw it on LinkedIn, know about Vanta Delivers? Thumbs up. How many of you have been plugged in to Vanta Delivers? Great. Yes. Love to see it. Perfect. At our Vanta Delivers launch this month, Clay showed a workflow that pulls dependency vulnerabilities from Vanta through the MCP, attempts fixes automatically, and escalates only the hard cases to a human. Hard Rock Digital also shared they combine their MCP with the Git based system to accelerate threat modeling and audit evidence collection. So those are real teams running on exactly the connection you're hopefully about to make or will make soon and solving these problems. Okay. Third, cross referencing by hand. Right? Test to controls to frameworks with two windows and a spreadsheet in between. Now it's like whole program synthesis, and that's what the tool is genuinely best at. For chasing down people, now you have the capacity to find who's behind on training. You know, you can send a a reminder email, Slack, especially if you're using MCP for those along with Vanta's MCP in your AI tool. One prompt returns a list of who's behind, and it can even draft the message for you, and then it could even send it for you if you've connected to those other tools. Five, the blank page problem. So even with the data in hand, someone has to then turn that data into words or the deliverable or whatever the ask is. And now the MCB can start drafts for you, and take you from zero to done in minutes. So chat check-in, the line here you see on the bottom of your screen for our chat poll, which of these five was your Monday yesterday? So did any of these problems come up for you yesterday on your Monday? If so, share the number in the chat. We'd love to hear. Great. Yes. Yeah. Chasing people, man. That's a that's a tough one. Great. Okay. So you can relate. Awesome. Let's see if we can keep helping to solve these problems for you. Okay. And the bigger why. So that was kind of the in the weeds why. Right? Like, the problems that it solves. But the bigger why is that, you know, beyond your own hours and your own sanity, any one of those problems sits between you and noticing something, you know, you noticing something and then having the capacity and time to fix something. You know, when answers take minutes instead of afternoons, gaps get found sooner, fixes start earlier, and people ask more questions of their program because asking is easier. So it's not just about speed. Right? It's actually providing you with a much healthier program, which is the ultimate why that we all really wanna address and this can help get you there. Okay. Now we're moving on to the how. We answered the why. So the how has three parts. So if you've kind of tuned back out, if you're checking Slack or something, tune back in. This is a really important part because this is all about your connection. So we'll talk about how to connect, how to prompt, and how to use best practices for the best results. And we'll especially dig into more of Sam's scenario here, in this portion of our session. Okay. First, some great news for you. Anyone on your Vanta team can now connect. So it used to be admins only, but as of late August, editors, collaborators, custom roles can all go through the same flow. They can connect to the Vanta MCP, and each one will be scoped to exactly what their permissions allow. So if I'm not the admin was your reason for not trying this before, that reason is now gone. So you should hopefully be unblocked and able to connect. Now since you're all on different AI tools, it doesn't really make sense for me to demo one tool's connection flow to you. But I am gonna kinda show you what your resources are for getting connected and just kinda speak generally to the map of how to do that. So in our Vanta help center, which hopefully you know and love well, thumbs up. Everyone visits the Vanta help center, hopefully, uses it often. Yes. Great. Love it. It's great. We love the Vanta help center. Okay. So there's this article called connecting to Vanta's MCP, and it is also linked in the docs section of our webinar, and I will also drop it here into chat. Okay. It has a dedicated section for each tool, Claude Code, Claude Cowork, Cursor, and Perplexity, and then it has general steps here on the bottom. Right there, you can see my mouse, for connecting to anything else that's MCP compatible. So it also has the regional endpoints of your companies in The EU or Australia. So you can find your tools section and then follow it top to bottom and get connected. Also though, just to jump back here, because every tool is the same, there's really, like, three general steps you have to do. Honestly, most people connect can connect in, like, one to two minutes. So the first step would be to find your tools connectors or integration settings. Two, you can add the Vanta MCP server URL. And then three, sign in with your normal Vanta credentials and approve the permission screen, which is really asking for nothing beyond what your role already sees, and that's it. Three things across any of your tools will generally get you connected. Alright. Here are just a few tips for getting connected. These are from the support trenches. So these are kinda like cast on from the support team just sharing, like, typical things that people might run into. So first, Regent tends to be the number one gotcha. So just make sure you're paying attention to those different endpoints. Second, the permission screen isn't asking for special access. So it is your roles access, so you can feel really comfortable with that. Third, the am I actually connected test is simply running your first prompt, which we'll do together in a minute here. Fourth, if your connector ever shows inactive, you can just reauthenticate it. You don't have to reinstall. Then fifth, if your company restricts a specific AI tool, remember the MCP is tool agnostic and that flexibility is actually how one of our customers got it through approval when cloud specifically was restricted. Engineers and SREs, if you're here, you can connect through Cursor, Cloud Code, and the terminal, which are all first class for everything, you wanna do. So I definitely wanted to call that out. So if you can connect right now, I would love for you to try. So I am gonna give about, like, sixty to ninety seconds to give you a chance to at least go into your tool and kinda try that flow out. So even if you don't connect in this moment, that's okay. But go try and see if you can at least follow the steps in the article or follow those general three beats of how to connect, and then we'll jump back in. I'm going to take a little sip of water, give you about sixty seconds or so to connect, and then we'll jump back in. Okay. I'm not gonna give a full sixty because I'm just gonna talk while you're trying to connect and answer Mike's question. So keep trying to connect. I do want you to be at least attempting to connect. But I am going to oh, great. Shannon already answered Mike. So thank you so much, Shannon, for that. I appreciate it. Shannon, Sierra too from the customer education team. So, thanks so much for jumping in there. Okay. If you were able to connect, great. Or if you are connected but just haven't used it yet, great. If you are gonna try to connect later, then that's also fine. You're just gonna kinda follow along and watch as we go. Let's do our first prompt together to test the connection. If you are connected, you're going to take this prompt over to your AI tool and I'm going to drop it in chat so you have a really easy way to just copy paste it. You're just going to drop that into your AI tool and run it. Let's see what happens. I'm going to go to Claude. I'm going to show you what it returns for me, which is scary because it's from a demo environment. But regardless, I'll show you what I got. Then I want to see what it does for you. Go ahead and take that and run it. Let's see what happens. Okay. So when I take this to my tool, I said, what frameworks am I working toward in Vanta? And roughly how far along is each one? Claude gave me some unsolicited SaaS, which says this account has a lot going on. Yes. I know. Thank you so much. So it gave me the furthest along, which was 60 to 75%, and it lists those out here. Then it gave me mid progress and about, you know, roughly 50%, early stage, paused. So it returned all sorts of great insight to me here that I can kind of just briefly check on the status. And, again, this is just to kind of check your connection. So, we're just making sure you're there. Okay. Yeah. So I do wanna call out that it will start to give you a lot of homework and a lot of tasks once you start running, which is helpful. But, yes, unfortunately, it is, it might give you some stuff you have to do. Go ahead and drop one framework name from your results in chat. Look at your AI tool and what returned to you and go ahead and drop you know, a framework that returned in your results into the chat. Or if you aren't connected, maybe drop the framework you would want to ask about first. I would just want to see what the room is working towards. A quiet chat? No one's working towards any frameworks? Oh, thank god. Okay. Great. Love it. Perfect. Anyone else? What are we working towards? What did you what did Claude return to you? Or what did perplexity return to you? Or what are you what do you want to ask it about? Great. Okay. Awesome. Alrighty. Well, if your frameworks are coming back, then congrats. You're connected. And if not, you can still connect later. So keep keep tuning in here, and let's roll. Alright. So we've tested the connection. We're all live hopefully or at least taking notes for later when we do connect. Let's circle back to Sam's Monday tasks and our running scenario. We've already taken care of one, which we can see. We did our risk register. Now let's work on the leadership asks. That one is this first box here. Something to consider before we do too much prompting is we need to ask well. So as we work here on Sam's first task, the leadership update, we're gonna raise the bar a little bit. So we're going to consider context, task, and format in our prompt. So context is who's asking and what data to use. Task is a clear verb and what to include. And then format is what the output should look like, and that can be a a step that everyone skips. And then, honestly, sometimes the return can be really unhelpful because it's not giving it to you in the way that you need it. So for our risk one, we could be a little bit more open ended. Now for this task that we're working on with Sam, we need to be a little more dialed in, and we're still learning the how portion of our training. So how to prompt. We wanna consider these three things. Let's take this over to our Claude and I will drop it into the chat so you can just copy paste it. Great. If you want to run it alongside me, there's a prompt for you and I'll show you what this looks like. So here we go. Prompt. My leadership wants a one page status view of our SOC two and ISO programs. Build a red, yellow, green table by control category, one plain English line per row, and end with the three items most at risk. So here, it's giving me kind of like a headline. It has broken down that stoplight format for me, given me the three items most at risk. So the value of the way we prompted it is that I have exactly what I need in the format I need it, and it's even prompting me to take this next step, right, where it says, do you want me to turn this into a formatted one pager PDF? I probably do want them to do that. So this is the value of getting a board ready stoplight view. Then I say this with love. The Vant UI doesn't produce in this format. Right? And this is, you know, a really common thing real customers do with the MCP. They pull their data and reshape it into the format their audience wants. And you can in this scenario with Sam, you can swap the asker for whatever yours might be. Right? Board, auditor, enterprise customer, etcetera. The prompt shape is identical no matter what. Okay. So for those of you that are connected, did you run this in your AI tool? What did it return to you? And if you are not running it, maybe drop the category you suspect would turn red if you are running it. So, again, our chat prompt is if you did run it, you know, drop one word in the chat for your reddest category. No judgment. You saw mine. And for everyone else, maybe drop the category you suspect might come back red. That kinda gives us a sense of what we might wanna work on in Sam's scenario. No one wants to share their red, Did anyone try to run this with me? How about that? Thumbs up if you tried to run this prompt alongside me. Curious if anyone in the room tested it out. Love to see it. Great. Okay. Alright. Let's keep rolling. So, really, just for this portion of our how, how, I wanna reinforce a couple of things before we move on. So one, once we get that return that you just saw, I'm not gonna forward that table yet. Right? I want to verify the reds. So we don't wanna just collect it and immediately forward it on, especially in these, you know, elevated kind of really important asks. We always wanna verify and then send. And then the other thing we wanna remember is if the format's not right, don't start over. So when I'm in my MCP, if it doesn't return the right thing to me, I could just say something like make it more formal, collapse it into this amount of rows, or change how you approached this wording. Like, there are ways you can iterate. So don't always feel like you have to start over, start over, start over to get the perfect thing. Really iterate on it and that can that follow-up is often where the magic can be in getting you what you need. Oh, thank you, Mario, for sharing. Love to see it. You are in a much better spot than my demo tenant. That is for sure. Okay. So this is looking at Sam's second task. Okay. We just did the leadership one. Now we're looking at Sam's second task, which is the overdue training chase down, which it sounded like a lot of you might be doing with chasing down people. So this time, we can really build the prompt together. So what a beginner or early MCPU user might type in would be check on our policies. Right? Okay. That probably would work. But with what we've learned about prompting and what we're trying to get with our data, it's probably not gonna give us exactly what we need in the way that we need it. So something better might be, which three of my active policies have the lowest employee acceptance rates? Give me a table of policy, acceptance rate, and outstanding count up to 25 names and emails per policy, and one short friendly reminder written as plain text. So we're giving a lot of context. Right? We're telling what format and we are just being a lot more specific to make sure we're getting what we need. So when I take this to my AI tool, k. Prompt is right there. I use the same thing again. Here are my three lowest acceptance active policies. Okay? You guys like this 00%, 0%. Fun thing about using a demo tenant or so. It was lots of data to record on. And then it went ahead and flagged my outstanding employees. And then down here is a reminder draft that I can use to get them to go accept their policies. Fun thing. Right? If you have Slack connected or your email, your next prompt could be great. Can you reach out to all of these folks on Slack and send them this message? You know, that is how easy that process can be. Now you have the insight, the list, and the email all from one prompt, same tool, same data. The only thing that changed was the question here. So it's a great way to kind of start to tap into that chain MCP chain tooling that we talked about a little bit ago. Alright. So best practices to remember. Vague prompt will equal a vague answer. So always remember to think about context, task, and format. Two, asking beyond your products or your your role will likely yield empty or potentially scoped answers, which is honest and by design, but you'll wanna keep that in mind when you're prompting. Three, you know, always spot check. So, you know, trust and verify. Right? Always spot check before sending somewhere else. Four, asking for everything at once will probably be too overwhelming and will get you exactly what you want. So sometimes three small sharp prompts can beat one giant one, so just kinda keep that learning curve in mind. Alright. Look at that. Three things off the list, and we still have some time left to go to tackle our last one. So Sam's Monday is looking a lot better already just in, forty five minutes. And we're on our last section here, which is the when and where. So when and where. This is basically a story of both and. So reach for the MCP when you need an answer, a summary, a draft, or a cross reference, and do that when you need it fast. Right? So in a scenario of someone just reached out to you or the meeting is in ten minutes kind of kind of moments, the MCP steps in really powerfully in that type of situation. But reach for the Vanti UI when you're configuring, assigning, approving. You need the full detailed picture. That is and will always remain home based, and it's not going anywhere. So, really, this is an answer of both for the, you know, the when. But I actually wanna turn this question over to you. So tell me, you know, would love for you to drop some ideas in the chats for what you use the MCP for or maybe are hoping to use it for and what you still use in the UI beyond what I recommend here. Does anyone who's a little more familiar with our MCP wanna chime in in the chat and share what you turn to the MCP for and what you turn to the UI for? Would love to hear from a couple folks in the chat. Great. Yeah. Vulnerability management for sure. I know there's a question here about the MCP and a comment to test. I actually don't know the answer to that off the top of my head, but I'm sure one of my folks that are here, from digital success can track that down. If not, we can follow-up with you after. I don't know the one off the top of my head. But, yes, vulnerability management. Love the idea of using it for speeding up reporting and notifications. That's so cool. Oh, and you created a skill. That's awesome. Love it. Great. Thanks for sharing, you seasoned folks here on the chat. Okay. So that was really the when and now where. So kind of figuring out, like, how does this MCP fit into the rhythm of your workflow. Customers get the most out of the MCP when they kind of figure out what that rhythm and cadence looks like. So this is just an idea. These are suggestions and ways for you to figure out how it fits into your plan. So daily could look like a pulse check. Right? Like, what needs my attention today? You know, show me my failing tests or overdue items. Give me the top five. That could be something really brief that you use every morning. Maybe weekly, it looks more like a brief, right, which is related to Sam's third task. So we're actually gonna do that one momentarily. So pause on that. Monthly, maybe a report. You could do that stoplight table again that we ran a few minutes ago. That could happen on a schedule. Maybe during audit season, you can use this for cross referencing. So failing tests, map to at risk controls, prioritize by audit exposure. That can really help you move from split screens into one prompt. And remember, it's a two way street. So you watch the right action in the very first demo related to the risk register. Right? Rights span risks, policies, vendors, privacy, and issues with more right actions landing constantly. So work you do in your AI tool can publish back into Vanta. So that is kind of the other part of this rhythm that you might not see here that you want to figure out how that fits into your workflow to actually use it to go take actions in Vanta. For engineers, your version of the weekly brief might be something like show me my failing tests and dependency vulnerabilities and then fix them. That clay workflow from the y portion of our session runs on the Vanta CLI, which is generally available in open source. Runs on the Vanta CLI, which is generally available and open source. We also have the Cloud Code plugin, which can help you with, you know, doing infrastructure fixes, opening pull requests straight from your terminal. So this is just some stuff for you to noodle on, some ways for you to think about how all of this can fit into your workflow rhythm. But, you know, definitely feel free to screenshot, and consider how you might wanna test this out. Okay. Our last Sam check-in here, we only have one more thing to complete. So let's close this out with this last prompt that says build my week. Look across my Vanta account for overdue or soon do tests, open evidence requests, access reviews, and progress, and critical vulnerabilities. Give me a prioritized to do list with a one line why for each. Drop them in the chat so you all have it. So you can if you're connected, you can take it, run this, see what comes back to you, and let me show you what it looks like when I do it. Okay. Here we are. So same prompt you just saw. Right? Build my week. Here's your week built from a live sweep of the account. So priority one, newly due soon tests. Priority two, overdue tests. Priority three, access reviews. Priority four, evidence request. And then it says, you know, it's already prompting me for that next action. Right? Like, do you want me to drill into any of these? Look at them a little bit more closely. So would love to hear, you know, either through an emoji, maybe you're excited about this prompt or, you know, would you use this prompt? How would you use it? How might this prompt be helpful for you? How could you iterate on it so it's more customized to your workflow? Would love to just kinda get some general reactions about this last prompt for Sam's last final Monday morning crisis. Yay. Thanks for the engagement. Love a thumbs up. Love to see people would use it. Oh, yeah. Great. Brian, I love that idea. Yeah. Take it. Run it first. I mean, it's Tuesday. It's basically still the beginning of the week, so you can use this today. Love it. Love it. Okay. And look at that. Sam's done. That's it. In under an hour, we've addressed Sam's scenario along with learning about Vanta's MCP from the one zero one level. We've answered the what, the why, the how, when and where on top of doing all of these tasks. So we've done everything and more. And you now as an attendee are moving away from beginner status to regular user status, hopefully. To put a finer point on it, we addressed the leadership ask, the overdue training, the week with no plan, the open risk. In this, you know, one hour, fifty minutes, you watched you know, fifty. We're not even in an hour yet. You watched every one of those get handled. Right? We did the stoplight table for leadership, the overdue list with the reminder already drafted, ready to push to send to Slack, the weekly brief built, the riskiest open item found, and a risk scenario created. Sam is now ready to walk into that board call and you're ready to. You hope you know, it seems like based on our opening poll, you kinda came in with a vague sense of, you know, I probably should try this thing or I wanna try it. But now you're hopefully actually leaving with some concrete ways to get started. And that's a really great way to start the week, and I'm sure Sam thinks so too. Let's do the same poll again. Maybe our ones have moved a little closer to to two. So how are you feeling? Are you feeling a little more comfortable with the Vanta MCP right now? Would you move yourself up on the scale? Give your number in the chat again or a thumbs up or an emoji or something. We just kinda love to see where everyone's at. We walked in with a lot of ones and twos. So maybe we've moved up. Oh, we have moved up. Oh, great. I love that. Oh, thank you for engaging and sharing. Oh, this is great. Love to see it. You guys have been such great attendees. This is perfect. Okay. So this was actually part one of three. So we're gonna do three MCP focused trainings. And, you know, this one was, of course, our one zero one about getting started. In October, we're gonna do one that's focused on questionnaires and answer library. So if you've ever lost a week or more to security questionnaire, this one is for you. In December, we're gonna do one that's focused on third party risk. So registration links will be shared in the exchange, which we'll talk about momentarily, the wider Vanta community. And also, I'm gonna send you, me, or CSM or Vanta itself, will send you an email, to remind you to register since you've already attended the session. So you kind of already have, like, a secret in, ahead of the game. So I'll make sure you all know when the next ones are live for you to register. So I hope you can come to part two and three. And if you're kind of ready to move past beginner status before this October session, I do run the weekly MCP builder workshops, so highly recommend you all registering for those as well. Let's talk about this week. So to keep the momentum going, connect this week if you haven't already. Check out our help center documentation, which has a prompt library, by the way, a really good one. You can go deeper with the Banta CLI and Cloud Code plugin. You can reach out to your CSM with feedback about these tools, share what you need more of. And then we want you to do one other thing, which is to join the Vanta community. So Vanta community is our, you know, go to spot for GRC practitioners to engage and share, connect. There's so many great ideas in there about how folks are using the MCP and just Vanta in general. So we have a couple of ways for you to join. You can join from the product, which you'll see over here on the right hand side at the top. It's giving you a little screenshot of what that looks like. Or you can scan this QR code and join that way. Great VantaCon shout out in the chat. I will be a VantaCon. I'm doing a training there. So if you all are coming, please come see me, come join the training. Would love to meet you in person. Just a shout out, if you join community, you can get a Vanta Builder badge for attending today's session. And we do offer some swag up for people who share what they built. So we'd love for you to, get involved there. And we have a special area of the community, an exclusive area called the exchange. So we do exclusive build workshops. We have office hours. We have AMAs with wonderful people from, you know, all across the Vanta organization and other GRC experts. We share different badges and swag. So we would really love for you to join the exchange as well. So please this is popping off in the chat I see, but also, there is a QR code here for you to scan. Keep the conversation going in the exchange. Please post what you ran today, what you learned, what you loved, what you're gonna be testing out. You can ask us questions. I'm in the community too so you can always shoot me a question there. Would love to hear from you all. And one last engagement ask from you. If you love today, please tell people about it. We would love for you to post about your experience in the LinkedIn. You can tag Vanta. Vanta loves to comment on LinkedIn posts. So I'm sure our Ilma Llama will pop up there to to, give you a congrats. You can tag me. You could use our hashtag build with Vanta. This QR code will take you directly to Vanta's LinkedIn. And also, I know sometimes looking at a blank LinkedIn screen can be frustrating, so I put some copy here that you could adapt and post if you want to do that. So you can feel free to take a screenshot here. But other than that, I just wanna say, would love for you to complete a survey about the training today. If you go to the docs section of our webinar, you'll see training survey. If you click to open that, you can leave me some feedback about today's session. Other than that, thank you so much for being here. You all were really great. You engaged so much. I love hearing from you all in the chat. It was wonderful. I hope that you'll be at our part two in October. And, again, connect with us on the community. Reach out to us. Post on LinkedIn. We'd love to hear from you. And thanks for being here. I hope you're excited about your one zero one level and ready to go to the next one. Thanks so much, guys, for all the emojis, comments. It's great. I'll probably stay on for another minute or so just in case there's any questions, but you all are good to go start testing out your MCP. Hope you have a great week.