Vanta Spurns Token Markups, Absorbs AI Costs to Win Enterprise Agents

Vanta放弃Token加价转而自行承担AI成本，全力争夺企业级智能体市场 Christina Cacioppo认为，软件供应商必须牺牲AI推理的毛利率，而非对token进行加价转售；她押注基于专有上下文构建的自主执行能力，其生命力将远超虚高的AI估值倍数。 Christina Cacioppo详细阐述了Vanta如何从合规与SOC 2自动化工具，演变成为由全面的集成“信任图谱”驱动、以智能体为先的网络安全平台。她剖析了传统高毛利SaaS估值倍数与低毛利AI热潮之间的宏观经济脱节，解释了为何Vanta选择主动承担底层模型推理成本，以优先促进客户使用而非依赖token加价获利。探讨内容还涉及为适应编程智能体而将遗留代码库模块化的架构现实、顶尖人才传统薪酬体系的瓦解、前置部署工程（FDE）模式的隐患，以及硅谷著名的“广告牌之战”。展望未来，Cacioppo指出，针对达到机器级运行速度的持续性网络攻击，必须依靠防御性自主智能体将人类从常规修复流程中剥离出来。 Vanta拒绝在token上加价并自行承担AI推理成本，宁愿接受较低的短期毛利率，以推动智能体工作流的采用率从25%向75%的目标提升。 (10:00) 随着具备高度自主行动力且深刻理解上下文的业务骨干创造出不对称的商业价值，顶尖技术人才的薪酬在五年内激增了约10倍。 (23:45) 欧洲软件开发者正刻意采用中国开源模型而非美国API，以此消除与美国贸易限制政策变动相关的致命业务连续性风险。 (32:03) 网络安全威胁正从孤立的零日漏洞转向机器驱动的利用链攻击，自主模型通过测试数以千万亿计的排列组合，将原本无害的错误配置串联在一起发起攻击。 (33:43)

Christina Cacioppo argues that enduring enterprise software companies will survive the market's AI re-evaluation not by chasing low-margin research hype, but by transforming passive systems of record into autonomous, agentic workflows while anchoring long-term defensibility in proprietary data and verified trust. Transitioning from Passive Compliance to Agentic Remediation (5:41) Rather than treating AI as an existential threat to SaaS, Vanta transformed its compliance workflow from passive monitoring into an active, agentic execution engine. Traditional compliance software relies on alert dashboards that flag missed checks; Vanta's agentic architecture autonomously performs remediation, such as drafting custom integration scripts to revoke access when an employee departs and directly executing changes upon admin approval. This capability relies on Vanta's 'trust graph,' a specialized data foundation powered by more than 400 integrations. Cacioppo reports that roughly 25% of Vanta's approximately 16,000 customers currently operate in an autonomous workflow where agents independently resolve tasks. The immediate strategic goal is expanding that figure to 75% adoption, which Cacioppo contends will drive durable user retention and expand the platform beyond audit readiness into continuous security operations. Absorbing Compute Costs Over Friction-Heavy Token Pricing (7:39) The broader software market selloff created severe valuation divergence: durable SaaS businesses generating 80% growth at 80% gross margins were penalized, while native AI firms with 160% growth at 30% gross margins commanded 3x to 10x higher multiples. Despite this multiple compression, Cacioppo maintains that high-margin business models will outperform cash-burning AI application profiles over time. To navigate commercialization, Vanta deliberately absorbs AI inference expenses instead of passing them through as marked-up token pricing or opaque credit systems. Enterprise buyers readily recognize marked-up model costs, pushing back when asked to pay premiums for underlying model calls. Because security workflows center on targeted decisions and remediation actions rather than continuous code generation or large-scale document drafting, Vanta consumes far fewer tokens than coding tools. This allows the business to temporarily accept lower gross margins to prioritize usage, evaluating models primarily on output quality designed by domain experts rather than raw inference cost arbitrage. Modularizing Monoliths and Overcoming Legacy Design Habits (13:45) Deploying autonomous coding agents within an eight-year-old codebase required re-architecting the system. Large monolithic repositories easily overflow context windows, causing coding agents to produce broken pull requests that require excessive engineering supervision. Emulating Notion's engineering playbook, Vanta invested in modularizing its repository so agents can fully process smaller codebase subsets, likening the process to onboarding thousands of developers at once. Internally, accelerating AI adoption requires overcoming organizational inertia and fear of failure. When redesigning enterprise customer onboarding—migrating complex legacy compliance data—product teams repeatedly defaulted to traditional multi-step wizards. Product leadership had to explicitly grant permission to fail, directing teams to replace rigid wizards with conversational, drag-and-drop canvases that rely on large language models to ingest and map messy, unstructured data directly. Separating Operational Realities from Enterprise AI Illusions (30:46) Cacioppo critiques several prevailing industry narratives, characterizing enterprise fine-tuning and post-training largely as investor-facing marketing rather than durable user experience patterns. Similarly, she views the trend of Forward Deployed Engineers (FDEs) as traditional professional services repackaged to attract elite talent. While Palantir successfully leveraged deployed engineers against eight-figure contracts, Cacioppo notes the economics collapse when AI startups apply the model to standard five- or six-figure SaaS contracts. Concurrently, cybersecurity dynamics are transforming due to model persistence. Human hackers typically abandon attacks after exploring several thousand permutations, but automated models can test quadrillions of variations to string together 15 to 20 benign, low-severity vulnerabilities that human defenders leave unpatched. Countering automated exploit chaining requires defensive security agents capable of running continuous penetration tests and applying remediation patches at machine speed without human bottlenecks. As enterprise skepticism deepens and procurement scrutinizes unproven AI applications, Cacioppo positions verified trust and automated remediation as prerequisites for enterprise longevity. However, whether enterprise software companies can permanently subsidize inference costs before successfully shifting customers to outcome-aligned pricing models remains an open question.

Vanta 联合创始人兼首席执行官 Christina Cacioppo 认为，在生成式人工智能重塑软件业估值逻辑的背景下，持久型软件公司的护城河不再依赖先发优势或虚构的商业包装，而取决于高上下文数据资产、主动执行的智能体架构以及严格验证的企业信任机制。 SaaS估值分化与长期现金流模型 (3:12) 过去一年多，传统企业软件遭遇剧烈的多重估值收缩（被称为“Sasspocalypse”）。市场情绪转向极度崇尚 AI，赋予前沿实验室和原生 AI 产品较传统 SaaS 高出 3 到 10 倍的估值倍数。Cacioppo 指出，这种现象引发了强烈的认知失调：从多年周期来看，拥有 80% 毛利率且保持 80% 年增速的软件业务，在现金流创造能力上远比毛利率仅 30% 但增速达到 160% 的高算力成本 AI 产品更健康。然而，资本市场和高端工程人才对技术前沿的追逐，迫使企业领导者必须证明自身正在前沿领域快速构建，而非沦为被私募整合机构收购的标的。 从被动合规转向主动执行的智能体架构 (5:41) 传统合规软件主要充当被动监控工具，仅在检测到未合规项时弹出警报。Vanta 转向智能体优先（Agent-First）的交互与执行架构，将系统转变为具备补救能力的执行引擎。例如，当离职员工的系统访问权限未被 IT 部门及时注销时，智能体能够自动生成自定义集成脚本，并向管理员提供“单次执行”或“始终执行”的操作选项，经批准后直接在外部系统中执行账号停用操作。在后台，该平台依托连接了 400 多个独立工具的“信任图谱”（Trust Graph）安全数据湖，并由数百个专职智能体与专业领域评估框架驱动，优先通过输出质量而非单纯的推理成本来选择底层模型。 定价机制困境：Token计费摩擦与结果定价局限 (7:39) 在探索 AI 商业化时，Vanta 发现向企业客户推行基于 Token 的定价模式存在巨大的客户阻力。企业买家非常清楚底层模型的单价成本，一旦软件厂商加价转售推理 Token，用户会立刻意识到这是中间商加价行为并产生强烈抵触。为此，Vanta 选择暂时主动承受较低的毛利率，将重心放在提升智能体工作流的采纳率上（目前约 25% 客户运行自主智能体，目标为 75%），待形成高留存与深度依赖后再调整商业化。与此同时，行业推崇的“基于成果（Outcome）计费”在合规与安全领域难以落地：若按起草文档、解决警报或注销账户等微观输出计费，会给客户带来繁琐的认知计算负担，而信用点数模式也容易被客户看穿其转换套路。 AI落地中的组织重构与高溢价人才机制 (13:45) 将智能体技术深度引入拥有八年历史的代码库带来了技术与组织层面的双重重构。面对单体代码库超出模型上下文窗口而产生低劣代码的问题，Vanta 参考 Notion 的经验对代码库进行模块化改造，使智能体能够聚焦于局部子系统并高效率产出代码。在产品设计中，团队曾反复习惯于套用带进度条的传统向导界面，Cacioppo 通过明确消除失败惩罚，促使团队转向支持直接拖入非结构化文件让大模型自主解析的无画布模式。在人才机制上，高上下文、高能动性人才的薪酬溢价较五年前激增了约 10 倍；Cacioppo 认为顶尖人才的价值在于对业务全貌的深刻理解与快速试错迭代，这种非对称贡献使得常规薪酬带完全失效。 机器持续性攻击与未来五年的自主防御 (33:43) 在网络安全前沿，攻防形态正在由“零日漏洞（Zero-Days）”向“漏洞串联（Exploit Chaining）”转移。人类攻击者在尝试数千次攻击向量后往往会放弃，而 AI 模型能够不知疲倦地测试数千万亿次可能，将十几个低严重度、甚至在人类运维分类中处于 90 天修补队列的良性微小配置缺陷串联成灾难性攻击链。面对这种机器级持续性威胁，传统的人工分类与决策流程已经失效，安全产品必须演进为自动化红队持续渗透测试与蓝队智能体即时部署补丁并行的闭环系统。在这一结构下，人类专家的定位将从逐项批准日志和警报，转变为架构系统底层规则与编写评估框架。 Christina Cacioppo 总结指出，未来的企业信任危机将伴随 AI 的普及而愈发严峻；社会对科技企业的默认信任已不复存在，驱动合规与安全从形式主义检查转变为经由实证指标支持的刚性基础设施。随着系统日趋复杂，关键挑战在于如何在保持企业稳健财务模型的同时，推动人类在安全与工程决策环路中彻底抽身，实现机器级速度的自主响应。

[0:00] Christina Cacioppo: In software, the blessing and the curse, and this is more true now, even than it was five years ago, you're really only as good as the last thing you shipped. [23:45] Christina Cacioppo: My comp philosophy is you never pay the top people enough. You should always try and you will never hit it. And there's a bunch of people you pay much more than their contribution. [30:00] Christina Cacioppo: I don't think anyone actually wants to be handed a model. [27:15] Christina Cacioppo: The reason to post-train a model is cost. Some businesses have that need. But there is not a quality reason to do it.

[0:00] Ramp is the only platform built to make your finance team leaner, faster, and better, saving businesses 5% annually on average so you can stay focused on growth. Ramp just opened for business in the United Kingdom, so if you're running a business in the UK, you can now use Ramp's AI-powered finance platform to manage cards, expenses, bills, approvals, and accounting all in one place. I run my business on Ramp, and so should you. Learn more at ramp.com slash invest. Thank you. Open AI, Cursor, Anthropic, Perplexity, and Vercel all have something in common. They all use WorkOS. And here's why. To achieve enterprise adoption at scale, you have to deliver on core capabilities like SSO, Skim, RBAC, and Audit Logs. That's where WorkOS comes in. Instead of spending months building these mission-critical capabilities yourself, you can just use WorkOS APIs to gain all of them on day zero. run on WorkOS. WorkOS is the fastest way to become enterprise ready and stay focused on what matters most, your product. Visit WorkOS.com to get started. Rogo is the AI platform purpose-built for financial institutions, serving tens of thousands of bankers and investors at hundreds of leading firms worldwide. Rogo's AI agents autonomously execute large chunk of your firm's workflows. They can screen deals, draft sims, run buyer outreach, and do diligence on data rooms, all with full security and compliance handled. stronger than the last. To learn more, visit rogo.com slash invest. So, Christina, I've been so excited to do this with you. I think it's been five years since we last did it. A lot has happened. A lifetime ago. A lifetime ago. We've been very successful in the intervening five years, so I'm excited to catch up. Can you give us like the one minute history of Vanta? I know we've done this before, but just to level set for people. When you started that, why, what it did then, how it's evolved, what it does now. Quick overview. Started Vanta about eight years ago, which feels about 8,000. Yeah, exactly. You're like in 17 centuries ago with the premise of can we help companies build and own trust and build trust is like start security programs and compliance programs or mature them and then go get credit for their customers through audits like a SOC 2 or B2. ISO 2700 and HIPAA, etc., security questionnaires, trust centers. But basically this idea of can we build a security company by helping companies to implement security because it's what their customer is asking for. So we were initially known for SOC 2, Billboard. We do have a billboard on the 101 that we were mostly known for making billboards, I joke, but it started very compliance, very audit centric. The idea was always, hey, that's just one way to get a company started on this. Since then, we've gone into different areas of security. We've gone into different ways of earning trust with your customers. Today, we're about 16,000 customers global with about five products. How did you win? This was a... new category when you started and i'm a ride to go through it my old business gets socked to compliant and certified in order to win certain accounts or whatever So it became an important thing kind of in the software era. There were many companies, I could probably name a handful that were trying to do what you're trying to do. Yes. You've emerged as sort of the only major winner. What would you attribute it to over the last, since we last talked, like how did you win? Execution. software, the blessing and the curse, and this is more true now, even than it was five years ago, you're really only as good as the last thing you shipped. You can be first and you can start it and... It doesn't really matter. And there's more true now because it's like someone can vibe code your thing in a weekend. And so you've just got to constantly be shipping new products, new features that make the experience better for customers running on that differentiation. There's some feature level. I think the other thing we really invested in when it got competitive, that was very good, was integrations, pulling in data from other tools, So now you can connect Vanta to one of 400 tools. We pull a ton of data out of those tools. And that's what powers the automation, the agent, the AI, all of that. We call it our trust graph. It's really just like a security data lake. And that is bigger, broader than any other provider, including some of the security providers, because we touch areas they don't. And I think in retrospect, that data foundation has let us build new things faster, more, Software and agent. It was that piece. And then I think on the go-to-market side, it was like just keep pushing. Good billboards. Shockingly good. Compliance that doesn't suck too much, which got really popular on r slash accounting, which none of us knew was a bucket list item in our lives. When r slash accounting likes your things, you're like, OK, something worked. And then the lived experience is just you can feel like you are going as fast as you can and you should, in fact, lean in further. [5:00] push harder. What was the scariest moment? The last 12 months, Sasspocalypse, you're like, OK, great job building a system of record. You're now screwed. You're now worthless. Yeah, exactly. You're worth like good job playing the game. Here's some steak knives. parts of that are true and parts of that are not true. In the vein of you're only as good as the last thing you shift, you're like, great. What is the agent? You have all this data and a system of record. Fine. What the heck are you doing with it? Say more about what that's felt like. Roughly, Vanta is a business that is growing very fast by historical standards with very high margins, which four years ago is considered like the best business model ever. Nothing changed about that. I think you're probably growing faster and the margins are really high. And yet the market has decided that the multiple for a business like yours is a third of what it used to be or a quarter of what it used to be. What has it been like to live through that as one of those entrepreneurs? Cognitive dissidence, you've got to hold two things in your head, right? One is, I think, at the fullness of time, a 80% growth business with 80% margins will actually be better than over the same time period, run it out long enough, a 160% growth business with 30% margins. One of these will return more cash in the fullness of time. But today the latter is valued at 3x, 4x, the multiple of the former. It's a 10x multiple of the former, right? And so there's a part where you're like, hey, guys, everyone's being irrational here. This is a good business and you have to keep doing things that make business sense and doing the things that customers want, which may or may not be token pricing models that put you more in the latter category than the former profile. And then there's like a very clear like game on the field piece. Well, if you need to be financed, again, you're going to be a You're looking at investors today, many of whom are kind of AI maximalists. Talk to firms that are like, yeah, we'll invest in the labs and we'll basically invest in tractors in Iowa. Nothing in between. robo tractors in Iowa at a lab and like everything else. And you're like, oh, interesting point of view. And from a like talent perspective, You hire ambitious people who want to do things. And if it feels like They're not in the right boat. They're going to change boat. The talent war part is very real, too. And so you've just got to live with the fullness of time. I think this will... all sort itself out and you have to live for the fullness of time. So there's this funny fear that now like bending spoons is the grim reaper in the hallway, you know, meme or whatever. No one wants to be a bed spoon. Love bending spoons, by the way. I think it's an interesting question of the siren song right now is sell tokens via agents for security and compliance, I'm sure, or something like this. Lots of thoughts on this. I'm just so interested by companies like yours that were sort of winners of this last era that now have to make a set of decisions. look at the 10 year rate, the cost of capital is going up because you can buy lots of money, a really high return in AI stuff for the time being. Someone looking at your business is like, well, I want higher returns or the same returns. And I get that through agents or something. Talk me through what it's been like to make the decisions. What are you thinking today? Do you think of yourself as an AI business? What's that? So we do. When we go benchmark where we are from an agent in our product perspective against the quote unquote, AI applications. We have yet to find someone who is further along. You could be like, we're doing a bad job looking is the obvious critique there. We are like constantly going to the set of names you imagine when you think AI applications and generally finding they're not yet available. where we are, where an agent is doing most of the work. It's chat-based, but there's like long-running agents. Our agent is auto-generating UI on the fly because sometimes you want chat, right? And building UI, user can do something, comes back in the flow, keeps it going. very much optimized around the like keep the user going in the flow loops. So I make sure I understand, meaning you are like a watchdog that's always on watching the security. So we are definitely that. But in some ways, you're like, you could do that in the past with just like monitoring and alerting. This is more like the Vanta product experience is now oriented around an agent that is bringing you as a user things and often saying, hey, I went and did this for you. Are you good with it? Then I will go do it. OK, great. The next thing I think I should What's an example? Just like ground that in an example. Someone just left your company and you didn't get them out all of their accounts. I just wrote a custom agent for you that will go do that. Do you want me to go turn this on? Great. Do you also want me to go remove accounts, access from these three people? Great. So it's not just here's a red alert that someone didn't meet the monitoring step. It's like, would you like me to go do this? We're taking a lot from the AI labs in their UI, but it's like, you want me to do this once? You want me to do this always? Great. The idea is like the agent will just be doing more of this work with the user's permission, but it's like, if they tell us to go do it, we're just going to go. In the course of that, we end up using a lot of tokens. We have talked to our customers a lot about token based pricing, and I think what we've generally heard is like people are real smart. So if you try to charge them based on tokens, they are very quick to say, oh, you've just marked up a Claude token 4X. [10:00] Are you 4x as good as Claude? I don't know. Let me go try it out. You actually put yourselves in a competitive position against... the models that the public no application really wants to like we're gonna talk about harnesses You don't want to get there. What we've seen on token pricing is, If you're trying to mark up at all, users are very smart and you will lose on that comparison. You can like not mark up. And this gets to the margin question. But anyone who is marking up cloud tokens for X and trying to pass that off to their users, I think. So what are you doing? That. We're eating the token cost right now. We're just taking lower margin. We're taking lower margin and really concentrating on usage. That experience that I just talked through where the agent is going off and doing things, we just fundamentally believe that is a much better user experience. And we've got. 25% of all of our customers in that sort of flow right now. We can get that to 75%. One, I think we'll have much better retention. much better customer experience. And then if we need to go sort out pricing later, we should go sort out pricing later. This is... usage phase and like a deployment phase of this technology and just go nail that. And you can figure out commercialization after. I'm also so curious that you said marking up a cloud token, but what about marking up like a Kimmy token or like something that's much cheaper? What's it been like deciding what to use, what's good enough versus cost? This is something we started concentrating on at the beginning of the year is basically evals. task specific or agent specific evals. From a product perspective, we call it all the Vanta agent. And it's one thing, of course, behind the scene, it's like hundreds of agents. But that all have specific tasks and use tools and have computers and that whole setup that I think everyone's converging on. but in the beginning of the year really invested in Using our subject matter experts, so the folks who like ran security and compliance programs at other companies to write the eval frameworks and then eval their evals. and using tooling for that. When we think about model swapping, it is mostly quality. It's not cost. I also think some of it to this is because we don't use as many tokens as coding. Coding is, and coding agents, lots of token use, maybe legal and like document drafting next. And then when you get to more of the... decision action world, which is where we are. Relative to coding, we just don't use that many tokens. So that lets us optimize for quality versus being as concerned about cost as some of the other companies. When you talk to other founders, both of your era and of this new era, what are they doing here? Are they also eating the cost and discharging a software fee? What do you think the end state is going to be? I think everyone wants to get to outcome pricing. Everyone wants to get to that. In another life I was a microeconomics professor. Like, I love this conceptually. You would get paid for outcomes and the alignment of incentives. Love it. The hard part is what are the outcomes that, support. Ticket closed makes a lot of sense. Actually, to that, maybe differentiate outcome and output as I understand it. In support, the ticket closed is an output. The outcome is actually my Sonos now works. But even like the ticket closed, that is... relatively unabductionable. I think a lot of... companies, Ranta included, are trying to figure out what's that like neat unit of work that we can all agree on. And you can say, oh, it's a document drafted. It's an alert closed. It's an account off-boarded. And you can kind of do that, but then... There's just so much cognitive overhead. Cause you're like, I just named three things I could name. another 47. And then our user has to be like, oh, well, if I do this with the agent, I paid this. But if I do it, you know, and you're just like, no one wants that. Everyone I talk to wants to get to outcome output pricing, but it's really not clear how you get there. Charging tokens themselves. really puts you up against Claude. Then people try to obscure that by charging credits and having some like credit token exchange rate thing. Which I think, again, users are smart and they're like, I see what you're doing and I see you've obscured it. Fine. Are you similarly far along in If you compared yourself just as you did just now to companies and how you're building product and the business using these tools. What are you doing that's interesting there? in the vein of eight-year-old company and eight-year-old code base. It is so much easier to throw the coding agents at a small code base on a large one. And that's like a token input thing. And they're getting better and better. But one thing we've spent a lot of time also investing in in the last year, and I know Notion has spent a bunch of time in and they were like nine months ahead of us on. is modularizing parts of the code base so that the agent can just understand smaller parts of it and understand it fully and then be really effective. Whereas if you chuck it at, again, eight year code base site unseen. It just doesn't have the context window to understand everything and we'll do silly things. Then you need a high-context engineer to babysit it, which is helpful, but sort of invalidates some of the point. Companies have always had this problem. They've just talked about it as, oh, when we onboard a new engineer, it takes them a while to figure out the good base. Agents have made it like, well, great. Now we want to onboard 9000 new engineers tomorrow. So it's just like heightened that problem. [15:00] I give her a lower grade on that. we are like in the process of using... agents and people to make the code base easier to understand too. You've had to grade yourself one through 10. about, wow, this changes everything. Building this company from this point forward is going to be totally different. Is it 10? Oh, yeah. Like nine. Yeah. You are like eight, nine. I mean, I think the principles are still the same. Everything is hard. People matter. All the tropes are still the tropes. But the way you do stuff, it should be quite different. Do you feel like there's a separation happening? Again, if I took your cohort of the successful... a year ago era companies, Do you feel there is a bifurcation where some have gotten the memo and are doing really well and some are just going to Bending Spoons? Kind of. I think there are just structural parts of some markets. And I think Bending Spoons is very good at identifying these where... You can put any entrepreneur in the world in that structure. This is kind of like the like this market or founder win Andreessen Horowitz thing and you're like market wins. I think that is still true. And I think this AI technology shift has made what some markets that used to look really good look really bad. The best person in the world can run into that will still run into a brick wall. And I think there is a founder motivation piece where it's like you've got to be in it and like push in this era. And if you don't want to do that, totally fine, but the rest of the company will not self-organize to do so. Get started with $1,000 off at Banta.com slash invest. Ridgeline is the first end-to-end system of record with embedded AI for investment management firms running portfolio accounting, reconciliation, reporting, trading, and compliance on one unified platform. Firms are moving off legacy technology and onto Ridgeline because of how far ahead Ridgeline's AI features are compared to anything else in investment management software. I've been hearing some unsure where to even start, and others convinced they can build their own order management system over a weekend. The reality is that running an investment firm will always require governance, controls, and a single source of truth for your data. And no amount of AI enthusiasm changes that requirement. Ridgeline is built on exactly that foundation, which is why I believe that the firms that come out ahead in the AI era will be the ones running on Ridgeline's unified platform. demo at Ridgeline.ai. So given that you've been at the edge, what's the Adopting AI 101 approach? If you're back to being a professor, what have been the most important things that have been hard to do but have proven fruitful? Just getting people to use it, which sounds obvious, but definitely use it yourself. Build for your company personally does not matter. understand where the models are, where they're good, where they're bad, what harness means. do that at least every two weeks because it just changes so quickly. You have to have a visceral understanding of what works and what doesn't. So that when somebody comes and they say, we're going to, fire all of our SDRs because I built one agent and you can kind of be like, did you now? Where is that true and where the false? So there's like a B in the mock. You want to get as many other people in that as possible. broadly generalized, it's very black and white. Some people just go do that and some people don't. If you'd asked me a year ago, you'd been like, oh, well, there's some intrinsic motivation and the people who aren't separate problems go deal with them. I think I'm slightly more empathetic now in that we have seen very strong people not go dive in and you have to know. Nudge them more, but try to figure out what's blocking them. And something can be blocking them and you can get people into the muck. even if they don't dive in initially. So if you're a small company starting out on Vanta, great. We've got this set of controls for you. You adopt them. You may be messed with them on the edges, but you're like off to the races. If you're a larger company, you clearly run a security and compliance program. And you come to Vanta because you want that. You want the automation and you want the monitoring. You want the real-time dashboards. You don't want the Jira ticket thing. Fine. Well, the hard part is you've got to get all that existing stuff. The controls, the risks, the data, the policies, the procedures, all that stuff in Navanta. classically the way you would do that, SaaS company is like, oh, we're going to make an onboarding wizard. It has four steps on a progress bar and we will like give you confetti. You're going to do that whole thing. we were setting out to build better onboarding this year. And the first couple product reviews the team kept bringing, onboarding was hurts. I'm being like, stop with the onboard. Like, no, go use an LLM. Stop. But they kept breaking onboarding with her. And this was like a. [20:00] motivated team talented people they cared and kept bringing onboarding wizards we finally got the experience now they shipped and it's working It's basically clawed. Big canvas. drag all your messy files in, drag in all your screenshots. We will like make sense of it. ask you questions, and then you are set up and off to the races. And it works really well. Why were these smart, motivated people bringing onboarding wizards? So you go out, like, finally. And the thing was, they were like, well, we don't know. if we can get the drag your files thing to work. It might not work. And we know the onboarding wizard will work because we've built those before. we basically don't think we're allowed to fail on this project. Oh, that's the thing? Great. We're just going to bet that the LLM flow is going to work and we'll make it work and we'll hill climb all we need to. But burn the boats. and just make this thing work. And then they were like, oh, okay, great. Do you want a next product review tomorrow? But it was this like funny thing where you're like, I would not have guessed that was their blocker. And you saw that and they were like, go on. You mentioned the talent war thing before. There's this big trend there. low margin, high growth thing where from an investor's perspective, The funny example you gave of a lab or a tractor, Everyone wants something that can be that thing. Tell me about how you're going to be a hundred billion dollar business or a trillion dollar business. Somehow now every business needs to be that big. What's that like to face down where you may not be building a business that should be a hundred billion dollar business. There's like a hundred of those in the world. There's not that many. Yeah. And it's not to say Bento won't be. It's funny for me, it's like, oh, should I adjust the pitch and be like, well, inevitably, the baseline here is Fanta's 50 billion. The real question you should be considering, if you join, can you take us on a journey to like 250 billion? Should I just start saying that? Probably not. But the answer is not. Definitely not. Totally. Because you're like, well, that is... the pitch that people who work at Vanta are getting from other companies. The market now is saying that so much. You do seem unambitious. So what do you do? There's a Vanta principle, value of a principle of do what it says on the tin. Do what you say you're going to do. And look, you can put ambitious things on it too. And it's not that. Don't over promise and under deliver. What we do is give some of that. Here's what we do as a company. Good, bad, indifferent. Here's the context. Now, let me talk to you about the business metrics that, And we'll generally just show a candidate to like show everybody everything. It's fine. They're good. We don't really want them to leak, but if they leak, It's not the worst. Walk them through that and what we're doing and how that grows over time. Try to root it in all of the facts and all of the data. And so in some ways, it's, counter the vibes with data and trustworthy vibes versus YOLO vibes. So what is it like in those talent wars if the most talented people are being told this thing's going to go to a trillion? Basically, every other company that's trying to hire them. Well, the better one is, well, we're six billion today, but Sequoia is going to do a 10 billion dollar check in a month. And then somebody is going to write it up to 25 next year. But we can get you in at six today. That's a good one. That one's tough. So you're faced with that. Like, what do you do? Do you not want those people? Maybe it's a filter. Sometimes I think if you go to the person and be like, what are you optimizing for? Why did you entertain them? But not in a pokey way. Just what's annoying you, Advanta? One of these in specific, very strong leader, still at Fanta. It was basically this person felt like they couldn't make decisions that would put Vantel on a similar trajectory. And that would be their preference. But they're like, oh, but this new place is promising me full decision making authority. on this trajectory... Totally fair feedback. What do you need to feel like you can make this? Like, just go make these decisions. Go. There's a money part and a FOMO part for sure. We've definitely gone up on comp and Comp bands are kind of a joke now. The comp band is... Whatever data you have is wrong. It is what is being offered to these people. How much more do you pay the top person at Vanto than you did like as a multiple versus like? Oh, I think we should have way more confidence version than we do. So that's interesting. But if I go five years back and that person made X dollars, what multiple of X does the person make? Now? Total comp. It's probably like a factor of 10. Wow. I think it actually should be higher. My comp philosophy is you never pay the top people enough. You should always try and you will never hit it. And there's a bunch of people you pay much more than their contribution. Say a bit more about what those top people are doing that's so valuable, typically. High agency. Get shit done. That sounds so milquetoast and that's worth millions of dollars. But like, yes, actually it is. a high context person who understands the company and the customers, And the domain making generally reasonable decisions and sometimes bad ones and then iterating on their bad ones and just moving. Is so valuable. Do you think it'll be 10x again in five more years? Or is that starting to get too crazy? is the rate at which [25:00] So the most valuable people are getting more valuable, staying the same, accelerating. Accelerating. The counterforce. At what point will investors get more concerned about dilution and stock-based comp? which they weren't throughout the 2010s. late 2010s, We had more of a like stock based conversation as an industry in like 2022, 23. Now we're definitely not talking about it. That's the counter side. Without that push on like dilution, I think people just get paid more and more. That's what they probably should. How much do you care about? your valuation and multiple and the engineering of those things. I've historically... not cared and sometimes even done things to like demonstrate my not caring and I care about ARR and ARR growth. I pay a ton of attention to those and those in the metrics. I think I've now been through enough valuation environments, like 2010s, 2021, 2022, seen the like multiple roller coaster. Or I just don't put a bunch of stock in that or I used to not. I think the thing I care a bunch about that then will get me to caring about these multiples is the sentiment of people who work at Vanta. And they feel like they're at a company that is growing and growing and then both in general growing and then growing relative to other there are other opportunities in the market. There's a part of the market that is... raising money every six months at 5x the prior valuation and That's a real talent competitor. And what about things like tenders? Have you done this historically? We have not done them historically, but The game on the field is you're doing a tender every time and – You got to do them. Does this feel unhealthy? Totally. My opinion about it don't matter. I can have all the feelings I want in whatever direction about tenders. They are normalized and people expect them and they expect it as part of their comp. And if you are not giving it to them, you got to give them a bunch of other stuff to make it up. How does all of this affect your exit ambitions to go public or to be sold or realize some end game for vanta not at all not really which is again a very founder privilege statement but the end game is how long can we keep this going for And 99% of the big independent software companies in the United States are public. Yes, IPO in that sense. I don't have it as a bucket list item myself. It's more of like means to an end. Lots of people join companies and want the bucket list item. So found a privilege caveat here again. But Goal is a big independent standalone company. On all this stuff, tools for building, ways of building, all these kinds of things I'm asking about, what do you think is overhyped and underhyped? Underhyped is... knowing the company and the customers, Relative to knowing the skill set, it is still really hard to hire, train, onboard, make effective people who join a company. Company Marines make this better and easier and put information at fingertips. But there is still totally something to be said about someone who's been in a company, knows how to navigate the company, knows the customers, all of that. I'm at a phase right now where we value that skill set more than like the vertical skill set of you're good at sales, you're good at customer success and would take a worse salesperson on sales skill who knows what. VVANTA pieces more. There's something in that. And I think especially as a lot of the labs and fast-coding AI companies are building out go-to-market teams, I think they're actually very much evaluating you're a good salesperson. That's slightly off. And then I think as we think about building agents for all of this, the company context piece. The customer context piece is probably being relatively undervalued to like you are a good salesperson. The less valuable skill slightly today than it was. Four years ago. What about overhyped? token-based pricing, low margin for the sake of having low margins. custom training models for the marketing value. We'll just say we post trained a model on our stuff. Explain why. Go talk to the inference providers. And sometimes give them, depending on the provider, give them true serum. They will say like, look, the reason to post-train a model is cost. Some businesses have that need. But there is not a quality reason to do it. There's kind of a marketing reason to do it right now in this environment. And I think there's a lot of post training for marketing reasons for investors and for customers, mostly for investors, actually. If one of those people were pitching the other side of this, the value of post training, what is the pitch to you? When these companies comes to you and is like, we'll do X, Y, Z for you, what are they saying? And what do they say the benefits are? You are putting a stake in the ground of being a real AI company. customers will notice. I think that's the primary one. I do think there's a secondary one in some use cases of like a form factor thing. And I think especially with large enterprises, you want to deploy this piece of software wall-to-wall that requires integrations and 17 people to sign off. Oof. That's the whole thing. Why don't you just give [30:00] them a like model specific to them for this use case. I don't know, semi reasonable. - So if I was one of these people and I came into Vanta, What does post-training literally mean? What would it mean in your context? What are you doing? What are the steps that you would go through? So we would need a bunch of data from that specific customer and then go... write a bunch of eval, then change model weights based on that customer specific data. The process would be they're trying to achieve something. They've got some data. You tune a model to achieve the thing, increase the eval, and then sell them that. Exactly. On an ongoing basis. And your point is that that's overrated because it doesn't actually improve their... It can. I think I'm skeptical in a couple ways. I was like, does everyone really want to be handed a model or there's so much around that from user experience perspective, It feels a little like, you know, in the social media or mobile apps, when we hadn't figured out, pull the refresh yet. There's something here. I don't think anyone actually wants to be handed a model. The same way it's like hitting a refresh button. I don't know if that sounds silly, but that's what we did in 2011. There's some version of that here. We haven't figured out how this actually works. And handing models, I think, is not the right reasonable iteration, but like not the right solution. Is there anything else like in the popular discourse that you think is crazy? FDEs? I kind of think FDEs are crazy. I think it's a great talent arbitrage, the same way like chief of staff is. What FDE often is is professional services. But I think if you call it a professional services role, that's a very different talent profile than FDE in the current discourse. And I think you're not going to get the Harvard CS grad who's high agency and wants to go do things and learn how to build a business to go sign up to be like, pro-serve person number four or number 12 at a company. I think there's a real job there. But I think this FDE... wrapping around it. is probably not going to stay. The company that we all think does as well is Palantir, which I think they probably do. They have eight figure contracts. There are some companies that have eight figure contracts. I think we're using FDEs in a place where there's like five and six figure contracts. And those economics probably are upside down. I mean, again, era of 30% growth margin better. And so that's what it works. That doesn't feel sustainable. When you're at like a dinner in San Francisco with other founders, what are you talking about? Are you talking about Models and which ones to use. What are the... actual topics of conversation between founders right now? There's definitely like talent war pieces. There was a time, this is more nine months ago, but it was the joy to go around the table and complain about Cursor. They're taking of everyone. There's definitely that piece. There is a bit of an ELTIC FTE. Do you actually do this? Is it working? Where are you on this? Right now, personal agents. Which ones do you have? What are actually useful? When you think about the open source stuff, Chinese models, this, that, and the other. What have you learned talking to people around the world that are grappling with these same questions? I was with our customers in France a couple of weeks ago, which was lovely. One of the things that I was very surprised to know is a lot of the French software companies in Paris are using Chinese models predominantly. I initially thought it was nationalism, maybe anti-American, you know, just something. there and stupidly suggested that. And someone very quickly corrected me and said, absolutely not. We basically all switched to Chinese models when Trump closed Fable off to non-Americans. And we just kind of realized there's a huge business continuity risk. We can't trust that. that if we build a business on one of these American models, it will always be available. So we stopped doing that. If you dream the most about where the business could go over the next several years, what's the version of Vanta that could look like that? We think of Vanta as a security company. Started with compliance. The best version of compliance is proving the security you have. You've got to have a good security. We think we can and are moving into a bunch of security domains and sort of a platform pitch of like, look, we've already got the data. We've already got the intelligency of the agent. We can start stacking products on top of that. and more security focused products. As we understand the security of all these companies, the pressure on them to go prove they're legitimate and industrial grade is just going to go up. That's one of the founding theses of Vanta eight years ago, but it feels even more true now. whether they're working with enterprises, Governments. Regulators, lawmakers, lawmakers. You just have to. prove your software is industrial grade in lots of ways and that will happen more and more in more and more ways in the future. And Vanta is the thing that connects... both sides of that marketplace. So what's the direction of security? This is, of course, an incredibly interesting topic. Early internet, everything broke, everything got hacked. This will probably be that times. [35:00] 10 or times 100. Does security mean you compete with CrowdStrike and places like that? What does security mean and what's its trajectory in this crazy era? These models just are so much more persistent than people. They will just... try nine quadrillion things the way a person would give up after, I don't know, 9,000. And then in the course of that, they find things and chain things together. A lot of the discourse is around zero days, and there's definitely some of that these models are finding that we didn't find before or didn't publicly find before. I think what's more so happening is these models are able to chain up. 17 benign-ish or benign on their own things together and then really find something. If you watch the OpenAI talk at RSA about hugging face, it's basically that. Persistence, putting together things that on their own, people would have been like, that's low severity. I'll fix it in 90 days. But that update, which I think... is tough and we have to fight with machines because it takes what was oh, that's important, but low severity and again, 90 days. And like, I'm going to fix this other stuff first. It just turns a prioritization on its head. And saying it makes everything critical isn't helpful because you're like, well, what do you do then? But I think what that ends up looking like is you've just got to fix stuff as fast as you find it. And some of the prioritization does need to go out the window. But if you're fixing things at machine speed, you don't have to prioritize. So that's what I think we're going to get to. But I think it'll be rocky as we get there. So as I visualize that, what that feels like to me is a constant change. penetration test? Yes. Where an agent is trying to break in like a white hacker agent all the time. And then every time they find something, it's got a buddy agent that's a patch. Let's fix it. Yeah. Is that like the ultimate product vision then? For some parts of Vanta, yes. I think we can do that in very compelling ways for the bulk of our customers. I think CrowdStrike will do it better for an enterprise. This is where choosing where you play and what market segment is most important there. Is there a firm that's doing something right now that you start to encroach on? Like of the traditional security firm? Different markets. I mean, take a... ongoing penetration tester and remediator fixer. We think we can do that for the cloud providers, like an AWS, any other misconfigurations there. Classically, Wizz does that. Wizz will do it for its customers. Wizz's customer base starts at a multi-thousand person company and their price point starts at a multi-thousand person company. We're more thinking about the platform as where are we already embedded with customers? Where are they unlikely to go paid? $150,000 for a whiz, but very well could pay 40 or 50 for something suited to them. Thinking through where are we already strong and where the competitors focused elsewhere, but probably on a more out market segment. How has your personal motivation changed? You set out to build this thing. I remember talking to you the first time about this very charming way you built the original product, building an Excel first. And this seems very old school. Looking back on it. There's always like the dog that caught the car thing. We're like, yeah, he built like this huge software business. Amazing. How is your own motivation evolved or changed through all this? Why keep going? It's interesting and challenging and it changes all the time. I really like my job. Sometimes I feel like I'm good at it. Often I feel like I'm bad at it. That's fine. That's in some ways not the point. But it changes and it's different. And when I feel like I've gotten good at it, usually something happens. It just changes and then I feel bad at it again. And I really like that dynamic. In the beginning, I had a lot more of a feeling of people think I can't do this. I think I can and I'm going to show them. And I still have that. We even did the like, You don't think that is going to be 250 billion? Like, let me show you. I still have it, but I think there's like less of a. dark edge to it. It's still a little dark edge, but there's less of it. It's, oh, cool. We got here. Now we're just going to see how far we can run. in a way that feels a little bit more like fun and joyful. It's my day-to-day. It's not always joyful. I don't mean that, but It feels more, oh, cool. Let's see how far we can go with this. And I think we can go pretty damn far. Whereas I think in the beginning, I was more like, you don't think I can get off starting line. Let me show you. Your finance team isn't losing money on big mistakes. It's leaking through a thousand tiny decisions nobody's watching. Ramp puts guardrails on spending before it happens. Real-time limits, automatic rules, zero firefighting. Try it at ramp.com slash invest. As your business grows, Vanta scales with you, automating compliance and giving you a single source of truth for security and risk. Learn more at vanta.com slash invest. We'll be right back. [40:00] understand your process. Rogo is built by finance professionals. It's connected to your data, learns your process, and quickly produces client-ready outputs. Check them out at rogo.com slash invest. The best AI and software companies from OpenAI to Cursor to Perplexity use WorkOS to become enterprise-ready overnight, not in months. Visit workos.com to skip the unglamorous infrastructure work and focus on your product. How surprised would the eight-year-old version of you be at some specific skill that you've developed today? And the way it's like the most surprising one that you're now good at that you hold, you would have said no way. I like go-to-market more than I ever would have guessed. you get to talk to like, personable, friendly people all the time. And I was like a terrible public speaker. I would get the most nervous in front of my own team because I cared the most about what they thought about. And I was like, oh, my God, I'm representing you. I want to do this so well. Exposure therapy was helpful on that one. You're like, you just do it and you do get better at it. But I think the, like... thought of being at kickoff or an all company get together in front of a thousand people on a stage. And I think it would have been like, oh. shit, like, I'm going to be bad at that. And I think the people around me have been like, yeah, she'll be bad at that. How much coaching can she get? It's like anything else. It's like, yeah, if you work hard at developing a skill, you tend to get better. You just put in the work. In that one in particular, I'm very interested in that one. It seems always high leverage. What matters? You've given however many hundreds of these things. What makes one good? The trope of like, no one remembers what you say, but they remember how you felt. It's so annoying, but it is so true. I do have to believe in the content. And if I just get up, I feel like it is content less. I will also feel that way and it will be obvious to everyone. You kind of have to figure out your version of it. Which for me, I try to put real content in and then also hold in my head, no one's going to remember any of these facts. So that's not the hard matters more than mine. So much more. Yeah. And the key to good heart is what? Believing in it yourself. Authenticity. One-on-one, on a stage. People are pretty good at reading one another. when you feel like the person you're talking to, is like going through the motions or doesn't really feel like it or is misleading you. Maybe they're also misleading themselves. People can tell. We can't entirely articulate what's happening. something seems off versus someone who is authentically into it. even if they're into something kooky. That just comes off. What are the big trade-offs that you struggle with right now? You can't do everything. You've got to make decisions. The stakes are higher than ever. It's easier to build than ever. It seems like a cool time to be doing what you're doing, but there must be big trade-offs that you struggle with. What are those? Yeah, I mean, one conversation... have had and have is we have tried over the last year to build in such an agent forward way. What does agent mean? I don't know. You get 2x revenue multiple if you say agent. It's like 3x if you say harness. I could say harness more. Prompt file, like minus 3x. We've tried to build such a great harness. Such a great harness for our customers. But actually. And... not build a bunch of SaaS UI. Oh, you want reporting? We'll have our agent go make charts for you on demand based on all our data. We're not going to put Tableau inside of Vanta, which is probably what we would have done in 2022, 2021, if we had been building a set. And we believe in this. In like a short term customer outcome sense, but also in a medium to long term, like this is what software will look like. And no one is going to be putting Tableau inside their products. So there's an aspect of this diffusion thing is just scary because maybe the tractor company takes 10 years. And in the meantime, you were right, but early and if we're wrong. And if we're wrong, you're kind of betting on the time horizon there. The other version of this big enterprises were working with the US enterprise trying to close. And they basically were like, hey... Mainly want your AI features, maybe we won't, but can you turn them off? For now. I don't go, ugh, why? The answer is because it's a way easier to get a non-AI tool through procurement. And we want you to turn it on after or after. Right. And we don't really have an off switch on, you know, you're like, oh, did we go build that? But you're like, gross. We don't want to build that. Hopefully we don't have to use it. We'd so much rather take that time and effort and build something actually useful. Anyway, we're not going to build this off switch, but it's just one of those. trade-offs. If you were an investor, which you used to be, where would you be focused right now? Both. Do not envy the job at all right now. Why? Why don't you envy it? As much as... I can critique and make fun of on the company side, the like, oh, I think people are miss thinking about some of these businesses, you know, like very short term way. I think the FOMO... And what are your returns? [45:00] I saw this tweet. If you haven't invested in a lab or an inference provider in the last four years, you should just shut down your firm because like, what are you doing? It feels like more acute on the investor side than the company side. However hard I think my life is. My life's not hard to be clear. Whatever these tradeoffs are that we're facing, it feels like more echo chambery, more judged on incremental returns and paper returns on the VC side than the company side. So what would you do? Where would you hunt? the labs and the inference providers apparently know those feel a little well we can debate how they played out but like high entry places it feels like the discourse and kind of the certainty around the open source model providers and then like the tooling and the companies around them feel sort of disconnected it's oh we've all decided the open source model will be very real and then the ecosystem around them is so much less built out than the big labs and that just feels off to me you probably what do you mean by ecosystem around them if you think that a lab or is it long like is that what it's called like these were basically just going to go deploy this technology and put AI in these businesses can you build one of those around the open source providers of course yes but can you really grab one of those models the way open AI is grabbing its own stuff and just replicate that playbook Basically a bunch of the lab playbooks seems like they should apply to the open source models And so there's more to do there. Exactly. Yeah, yeah. What am I missing on that? I think it's just like a pace thing. Anything where these things can go. It's kind of like, ironically, private equity or something. What private equity promised to do, which was to shape businesses up. Now they can much more easily. And so probably everything will get way more efficient if it can. And the return will come from higher margin, more efficient operations, whatever. That's like an efficiency thing, not a generative thing. I guess what I'm more curious about as you watch your peers and people around you is where are the green shoots on the generative side? A bunch of people starting security cyber companies, both because the environment and like I see proportionally more of those than other things. a bunch of people attacking rev tech, like go-to-market technology, which I think makes a ton of sense. Some of the old SaaS rev tech providers, they should be building all this stuff, but it doesn't seem like they are. And actually they're running more PE, Bench spoon sort of playbooks, which is again, a choice, like maybe a reasonable choice. Robotics is probably underplayed somewhat hot and hypey That not a novel area but it still definitely underweighted compared to all the software stuff The models are also getting real damn good there, to state the obvious. So probably spend time there. What else do you feel like you know because of your unique path that others don't yet fully appreciate? Over the past couple of years, Vanta has spent more time with government, federal, state, local levels inside the U.S., outside the U.S. There is an appetite for modernization, big stated priority of the current U.S. administration, but again, exists outside of the U.S. too. The AI pieces created more FOMO in government officials, where they may have always known they did not have modern processes or modern software, but it feels more of an imperative now than it used to. There's something there. Here's a broader answer. all of the conversation around AI and the technology has changed the priority and upped the priority of a lot of modernization. Sometimes it's within governments, sometimes it's within enterprises You see this with actually a lot of the cyber stuff and the CISOs we talk to their jobs are changing because of AI And they saying oh given AI and like the risks from the agents I would like to overhaul my enterprise risk program And I would like budget for that. And I think if they said, I want budget for my enterprise risk program three years ago, people would be like, we'll talk about an AOP, maybe. And I think because of the attached to AI, they are much more likely to get that budget today. So there is more of a buying motion. What would be your big pitch to come work at Vanta right now? Oh, it's a fun stage. You get 16,000 customers using a product, using agents. When you ship things, they are in use by real people getting real feedback instantly. It has product market fit. You're just like going and experimenting with like fast iteration loops. But early enough that you could still. You can move. Yeah, yeah. We won't let you move. Yes. Because it's such an amazing Silicon Valley microcosm. Can you tell like the most detailed version of the billboard war story that you've ever told? I was always very much a billboard skeptic because it seemed like who knows if they work. The tracking is terrible. Founder ego is why someone would do a billboard. It's like because I feel cool because my company has a billboard. So I was always very anti-billboard. Mid-summer 2021. Yeah. [50:00] SASTR conference. It felt like a big deal at the time. Like, thank God we've all moved beyond that. our marketing team decided to try out of home. I was like super skeptical. It was like, fine, whatever. Try it. And we're going to put something on the billboard. And Sarah, who was our PMM at the time, came up with compliance that doesn't suck too much. pitched into our head of marketing. who didn't like it because we were like negging our users. And it's basically like, no, don't. than Carter, much nicer, but like than Carter. And Sarah kind of believed in it. And so she went to me, which usually does not. She is not a like go around her boss person. But like really believed in compliance that doesn't suck too much. And so took it to me. And I read it. And I like immediately left, actually immediately texted Andrew Reed, our investor, and I'm As like, what do you think? You have good internet meme taste. She loved it. And they're like, OK. Let's try it. Let's run it. So like, what's the worst that's going to happen? People will think we're obnoxious and then probably forget in six months. So we stand the billboard up. Somewhere like near Burmigame. It goes over so well. We tweeted the tweet goes viral. The tweet then is still the next couple months, but like ends up on our slash and counting and the final day, right? You're like, oh, this is a gem. And I think it really worked because like it's our brand at our best, our marketing at the best, which our attribute is with a wink. We take our work seriously, but we don't always take ourselves seriously. And so I think there was something to that that resonated. It was a little obscure, but it was sufficiently obscure where if you got it, you felt like you got a secret. Anyway, and so like compliance, it doesn't talk too much, just works. 22, market crashes. Turns out there's a lot of billboard inventory available on one-on-one. So we get this great spot where like 101 and 280 intersect and it's on the... Great spot. for like not. that much, especially in Rotterdam's Fact. And we joke and we're like, oh, yeah, we've locked it up for years. We were known for this billboard in this spot. I think even on podcasts, I said the contract was locked up for years. It was not locked up for years. But that billboard got famous. And so People would contact us and Sarah all the time and be like, how do you do billboard? How do you do your billboard? And we're like, oh, we work with this agency. We work with this guy. And so we referred this guy, all of this business. So we'd just be like, it's him, it's him, it's him. millions of dollars of business. Great. Turns out in the background, we did not have a multi-year contract, which I should have shown. Certainly if I was going to say that on a podcast, not really. Do it as other than 10. And... The contract was basically... You put the billboard up. That's it. This is like not a contract at all. Multiple years combined. Sarah's seeing one of her friends for dinner who works at a startup. And her friend was like, oh, it's so sad. You guys gave up that billboard spot. Why'd you do that? And Sarah was like, pardon? And they're like, oh, yeah, we just signed for your spot. And Sarah's like, what? We go back. Or like contract, contract, contract written in crayon. There is no contract, basically. It's a policeman order. Call the guy. And he's like, yeah, you know, it's a placement order. They paid more. So we like swapped it out underneath us. I go to founder of said company and be like, hello, how much can I pay you? And I am like an extremely... in Elastic Fire. Like, what can I pay you for this spot? And he's like, nothing. And I'm like, how about I double? I think he was paying like a million and a half. And you're like, how about I give you three? Just get the spot. And he's like, we're actually cash flow positive. So like literally I'm an electric seller. And so we lost the spot. Good punchline. Things that made me think... Where I was like, oh no, Vanta's screwed. Honestly, losing that billboard, I was like, we did probably just lose a billion of market costs. Yeah, yeah, yeah. You're both bad omen. There is some vibes market cap decrease that just happened here. Yes. Is there anything I've missed that has your curiosity, your attention... that you think is really important. Right now. You see it in the data center conversations. To some extent in the personal agent stuff, it's like we just don't trust software companies, tech companies. the way we used to, and that is getting worse year over year. And it's kind of hard to imagine that inflecting back up. and getting to the useful optimism of 15 years ago. in that companies showing that they are trustworthy and industrial grade and secure through lots of metrics and actual definitions is more important than ever and will just get more and more important. And again, to a host of customers, regulators, lawmakers, auditors, etc., Vanta at $250 billion is in the center of all of those transactions. helping companies meet those high bars or wherever those bars are set. and then conveying that. helping them go convey that and accelerate their business. It's another way of saying that like AI has ushered in this extremely high excitement, right? but low trust. Yeah, that's good. Era. Yeah. [55:00] And Vanta's job is... Preserve the acceptance, but increase the trust legitimately. So not just on vibes. What will satisfy the. The market. to trust again. In a B2B context, businesses, I do think there's a vibes piece, but there's a very real data piece. And I would like to see... specific things or specific things. improvement curves or make promises with maybe contracts behind them. That one's quote unquote logical, rational, the way like enterprise markets often are. On the consumer side, it is vibes. That one seems very hard to turn. I do not envy that. Do you trust them? Use instinct, etc. Big companies now. Why not? It's not that there's anything particularly interesting in my email or chats or whatever. My life's not that interesting. But I would rather just Google or just Apple or just insert company here have that information. What's security like five years from now? Like if you just wrap it up with like, okay, here's where we're going and we want to be the person at the center of this. Describe that state for us. Security from five years from now is... faster, even more intense, more offensive, hopefully, in that, yes, we'll have more attacks, but I think you need more. white hat pushing back on all of that. And just get a speed that we haven't had. And I think you've got to get humans to get to that speed. You've got to get some amount of the human decision making out. Or humans are like architecting systems and writing rules, but not making decisions. every single decision. It's like falling from that. Writing evals. Maybe that's all we're doing in the future. We're all writing evals all day long. If you had to define... what the next five years of Vanta will be in one word. What word do you pick? expansive. I think we have this incredible base of customers of data. And now we get to go. Build things on top of that. Eight years to get here. We could go just build stuff on top of it for the next five or eight. It's so cool that we did this five years ago. Here we are again. Thanks so much for your lunch. Patrick O'Shaughnessy is the CEO of Positive Sum. All opinions expressed by Patrick and podcast guests are solely their own opinions and do not reflect the opinion of Positive Sum. This podcast is for informational purposes only and should not be relied upon as a basis for investment decisions. Clients of Positive Sum may maintain positions in the securities discussed in this podcast. To learn more, visit PSUM.VC. Outro Music. You know how small advantages compound over time? That's true in investing and just as true in how you run your company. Your spending system is your capital allocation strategy. Ramp makes it smarter by default. Better data, better decisions, better economics over time. See how at ramp.com slash invest. As your business grows, Vanta scales with you, automating compliance and giving you a single source of truth for security and risk. Learn more at vanta.com slash invest. The best AI and software not in months. Visit workos.com to skip the unglamorous infrastructure work and focus on your product. Ridgeline is redefining asset management technology as a true partner, not just a software vendor. They've helped firms 5x in scale, enabling faster growth, smarter operations, and a competitive edge. Visit ridgelineapps.com to see what they can unlock for your firm. Every investment firm is unique and generic AI doesn't understand your process. Rogo does. It's an AI platform built specifically for Wall Street, connected to your real outputs. Check them out at rogo.ai slash invest.