Jake Stauch

1. Key Themes

Coding Agents for the Rest of Us: Extending Engineering's Safety Rails to Non-Engineers

Early enterprise AI adoption clustered in engineering teams because code review makes experimentation safe. Serval's thesis is to replicate that safety net for IT, HR, finance, and legal. Jake Stauch explained: "They're able to build and experiment with these tools because they have this process of code review. It's very safe for even a 19-year-old intern to build something very powerful and then go through this process before it goes into production." 00:03:34 He added: "We wanna bring that same capability to the rest of the enterprise... under the hood, it's coding agents, but then you have all these abstractions, and guardrails, and controls, so things go through approval, they go through a staging, and uh approval cycle, and an audit trail." 00:03:56

"The Boundaries Are the Product": Governance as the Moat

Serval does not build models; it sells the controls that let large enterprises adopt them. Stauch: "We are not building the LLMs ourselves. Um we are building the guardrails. We are building um the the boundaries, and the boundaries are the product for us, because we are the way that a large enterprise uh with a lot of rules, and approvals, and process, and politics, can adopt these tools safely and comfortably at scale." 00:12:51 Because Serval is also a ticketing system, it can generate access requests with audit trails and just-in-time access for deterministic workflows. 00:05:33

Least-Privilege Enterprises Make the Escalation Layer More Valuable

As agents amplify the damage any one person can do, enterprises will shrink permissions, which increases the volume of exceptions that need to be routed through someone. Stauch: "What I'm actually going to want to do if I'm a large enterprise is shrink everyone's permissions down to truly least privilege... But then you're going to have exceptions all the time. We are the layer that grants those exceptions... So by owning how privileges get escalated in the enterprise, we actually become more powerful and more essential the more you clamp down on end user permissions." 00:13:59

Code Is the Right Abstraction for Automation (and AI Finally Made It Accessible)

Serval's founding insight is that drag-and-drop workflow builders are the wrong level of abstraction. Stauch: "Code was the right way to express automation in the enterprise... these drag-and-drop workflow builders is a painful level of abstraction because it's too complex for non-technical users and it's annoyingly abstracted for technical users." 00:10:15 The bet that coding agents would hill-climb rested on the domain being verifiable: "it's a verifiable domain, and it should actually hill climb fairly well." 00:11:22

Attack the Incumbent Head-On, From Day One

Rather than building a layer on top of a system of record or going down-market, Serval targeted ServiceNow directly. Stauch: "We set out from the beginning to be the largest enterprise software company of all time. We actually made a list of all the companies. We circled ServiceNow and we said we're gonna beat them." 00:08:21 He argued the two common alternatives both cap you: "Build a small layer, play nice with the big incumbents... Or they go down-market." 00:08:56 And the choice shapes everything: "All the product decisions, how you build your team, how you think about growth, they're fundamentally different than if you go and you serve one market and then you hope to expand later on." 00:09:31

Diffusion Is the Bottleneck, So Services Become Part of the Product

The technology outruns enterprise absorption. Serval deploys "a deployment strategist... and a deployed software engineer" on site 00:07:07, then builds a recurring-services arm: "We've started to build out a team, we call it ProServal, uh or automation engineering... because there's so much that they wanna automate, and they need the help to be able to figure out what can be automated." 00:07:20 Stauch ties this to unlimited demand for work: "as we've discovered with software engineering, there's actually unlimited demand for work." 00:07:20

The Product Is an Agent That Operates the Product

Serval built Catalyst, an agent that manages the platform itself. Stauch: "We realized it would be really nice if customers could just talk to an agent that understood their requirements and then translated that into the product primitives." 00:27:45 This solves the blank-slate problem and upskills users. Sarah Guo noted it's "raising the technical competence and ability of a huge number of users... turning people into more, you know, engineers, actually." 00:28:39

Background Agents: The Moonshot Blocked by Trust, Not Capability

Stauch called background agents "the ultimate moonshot in enterprise software" that "no one's really figured out how to do... well." 00:00:00 His practical answer is incremental capability escalation: the agent requests specific deterministic tools, humans approve each, and "over time, the agent becomes pretty competent because it has a big enough toolkit." 00:31:17 The alternative, unconstrained code-writing, "is an unbounded security risk." 00:00:15

Hiring More, Not Fewer, in an AI-Native Company

Serval went from 20 to 200 people in under a year, and Stauch expects hiring to accelerate: "Even though every single employee does so much more than what they could do a year ago, we have even bigger ambitions for what they could do, and so... because every employee is so much more productive, we actually want more of them, not less of them." 00:33:40

Decentralized Product Development via Customer Immersion

Serval has no product function. Stauch: "We've decentralized product development really to the engineering team. We don't have a product function, because they're the ones talking to customers every day." 00:35:42 He compares it to "immersion in a uh country to learn a foreign language." 00:35:24

2. Contrarian Perspectives

Personal Agents Are Too Weak for Large Enterprises

The prevailing narrative is that personal AI assistants will transform work everywhere. Stauch argues the permission structure of big companies hobbles them: "In a large enterprise, there's so much... process around approvals... the permission sets of any individual are so limited that even if you were allowed to bring your agent to work, which a lot of cases you're not, what it can do is so incredibly limited. It can barely even read your emails." 00:04:51 Serval's workaround is elevated service-account permissions wrapped in approvals: "it's a little bit more annoying than the magical personal agent experience... but it's what has to be true for a large enterprise, especially a regulated enterprise." 00:06:17

Smarter Agents Mean Tighter Controls, Not Looser

Many assume that as models improve, oversight relaxes. Stauch argues the opposite: "The intelligence getting better just makes our product more powerful and also makes our customers want to actually clamp down even further on those boundaries." 00:13:14

The "SaaS-pocalypse" Hits Point Solutions, Not Systems of Record

Against the narrative that companies will vibe-code their own software, Stauch says: "All the companies I've talked to that tried to vibe code these tools internally, especially the main system record tools... have discovered that it's quite expensive to build and maintain enterprise software." 00:23:44 The casualties are the long tail: "A lot of companies have hundreds and hundreds of SaaS applications. They probably need a couple dozen, total... I don't think that you're going to truly get away from the core system records, the ERP, the HRIS, the ITSM, the CRM." 00:24:35 Elad Gil added that "all the middle ware, too, I think, is just going to go away." 00:24:51 But, Stauch noted, a new generation of incumbents will form: "we are displacing those tools." 00:25:22

AI Spend Will Keep Rising Despite the Token-Maxing Pullback

While many read CIOs' pullback as a sign of waning demand, Stauch sees a cycle: "I do see our CIO customers trying to pull back from that initial token-maxing era, but... it's going to continue to be a pretty dramatic rise, because most companies are still figuring this out for the first time." 00:21:53 The real constraint is ROI proof: "I don't know that anyone's really figured out how to measure what you're getting from the tokens that you're buying." 00:22:33

Don't Build a Layer on the Incumbent

Guo and Stauch agreed that the standard startup playbook of an AI agent on top of Salesforce "has never worked" / "rarely has happened," with Elad Gil citing Veeva as the lone backward-integration exception. 00:09:04 Stauch: "It really caps you unless you find some niche to kind of like expand into later on." 00:09:14

3. Companies Identified

Serval

Enterprise automation and service management platform using code-gen agents; competes with ServiceNow. Mentioned as the subject of the episode, two and a half years old with 200 employees, up from 20 in January. Stauch: "We were 20 in January, so it's been an incredible ramp here." 00:33:19 Its stated ambition: "to be the largest enterprise software company of all time." 00:08:21

ServiceNow

The legacy leader in IT service management and workflow automation. Mentioned as the incumbent target, respected but vulnerable. Stauch: "They're a great company. They've done so much so well," 00:01:36 but "it's always kind of running 6 to 12 months behind where you want it to be... and then it's very expensive, and it keeps getting more expensive over time." 00:02:13

Fox Corporation

Large media company, one of Serval's largest customers. Mentioned as proof of the career-upside sales pitch. Stauch: "Fox Corporation adopted Serval company-wide, and the person who led that migration at the on the ground level, actually got nominated to be the MVP of the entire company." 00:15:12

Verkada

Physical security company where Stauch previously worked. Mentioned as where he discovered the IT/ServiceNow pain point. Stauch: "I discovered this category actually working at a previous company called Verkada, uh where we sold enterprise physical security products to IT teams." 00:01:36

Veeva

Life sciences software company that began on Salesforce and became its own platform. Mentioned by Elad Gil as the rare backward-integration success story. Gil: "Maybe the only one that worked is Veeva or something, right? 'Cause they started off on Salesforce and then they migrated to become their own platform." 00:09:06

Salesforce

CRM giant. Mentioned as a canonical incumbent and as the platform on which "AI agent on top of Salesforce" startups get capped. 00:09:04

Workday

HR/finance software incumbent. Mentioned as a target of down-market competitors. Stauch: "A lot of competitors to Workday, competitors to Salesforce. They just say, 'We're gonna be that, but for the lower end.'" 00:09:14

Rippling

HR/IT platform. Mentioned by Sarah Guo as an example of the down-market strategy. Guo: "This is Rippling or HubSpot the first 10 years, yeah." 00:09:29

HubSpot

CRM/marketing software. Mentioned alongside Rippling as a company that started down-market. 00:09:29

Claude Code (Anthropic)

Coding agent. Mentioned as the kind of tool Serval deliberately does not just hand to IT. Stauch: "We don't just kind of hand them Claude Code and tell them good luck. We actually build a platform that's dedicated to their use case." 00:26:17

OpenAI

Referenced by Elad Gil in the context of agent incidents where agents do unexpected things to accomplish tasks. 00:30:18 Also referenced for workforce-agent products. 00:04:40

Accenture and Deloitte

Consulting firms that have owned much of enterprise AI implementation. Elad Gil asked how Serval's forward-deployed motion compares to "the Accentures and Deloittes, and the like of the world, kind of owning a piece of this, doing the implementation for people." 00:06:30

Instinct and Muse

Personal agent tools Stauch cited as examples of the "magical personal agent experience we love." 00:06:17

4. People Identified

Co-founder and CEO of Serval; former Verkada leader. Mentioned as the guest and operator behind a direct assault on ServiceNow. His track record: "Spent many, many years talking to IT teams, figuring out what to build for them." 00:01:36 He credits his own willingness to reinvent: "One of the things I'm proudest of at Serval is how many times that we've decided to erase things we've done in the past that were very committed to." 00:37:22

Alex (Serval co-founder)

Jake's co-founder, who originated the code-as-automation conviction. Stauch: "My co-founder really had this idea early on, that code was the right way to express automation in the enterprise." 00:10:15 And on origins: "It was two people in my co-founder, Alex's, living room." 00:08:21

Elad Gil

Investor and co-host. Mentioned as the sharp interviewer who framed Serval as "coding agents for all that internal stuff" 00:03:08 and identified the Veeva exception. 00:09:06

Sarah Guo

Investor and co-host. Mentioned for framing the product as "direct attack, replacing enterprise systems of record, including for large enterprises." 00:07:59

Fox employee (unnamed IT lead)

The person who led Fox's migration to Serval and was nominated company MVP. Stauch uses this as the canonical career-upside story: "got nominated to be the MVP of the entire company." 00:15:12

5. Operating Insights

Sell Career Upside, Not Just Risk Reduction

Replacing an embedded incumbent is personal risk for the buyer. Serval reframes it as a promotion path. Stauch: "We have to convince ourselves and our customers that Serval is the way you get promoted at work. It can't be the fact that Serval is a risk that you take to hopefully get something on the other side... People built their career on ServiceNow. You're going to build your career on this next generation." 00:14:52

Never Build Bespoke Features; Promote Repeated Customer Automations to Platform

Serval's embedded engineers "never build custom features." 00:20:00 Instead, the platform doubles as an observation deck: when customers repeatedly build the same automation, it graduates to a first-class feature. Examples: detecting conflicting or out-of-date knowledge base articles, and reclaiming unused software licenses. Stauch: "It is a way for us to kind of test the waters of what should be really first-class features versus what should stay as automations." 00:21:22

Price on Workflows Run, Not Tokens

Stauch: "We try to price on workflows run because I think it's a better reflection of actual work being done versus just tokens, which is more like verbosity, um, and inefficiency." 00:22:59 This aligns price with customer-perceived value and sidesteps token-cost scrutiny, while customers' real ask is "Help me prove that this was worth it." 00:22:59

Treat Buyers as Adults and Design for Multiple Personas

Stauch: "A lot of the products that have been out there have actually kind of been pedantic to IT of like, 'Oh, you don't actually understand this stuff, so you need this very stripped down... version.' Actually, IT appreciates the complexity and appreciates the power of these platforms." 00:27:05 Serval designs for personas from IT engineers who read code to marketing users who never want to see it, while keeping code visible underneath. 00:26:42

Push Product Decisions to the Edge and Hire Product-Oriented Engineers

Without a PM function, engineers join customer calls (most on at least one a day) and hold decision rights. Stauch: "You have to be willing to push a lot of the product decision-making down below, and you have to be willing to move fast and also undo and revert... and kill your darlings." 00:36:58

6. Overlooked Insights

An Agent Platform Is Secretly an Access-Governance and Ticketing Business

Stauch mentioned in passing that Serval is "also a ticketing system," which is what makes just-in-time privilege escalation and audit trails native. 00:05:33 That combination positions Serval as the choke point where every elevated agent action must pass. As agents proliferate and permissions shrink to least privilege, the exception-granting layer accumulates data, workflow, and compliance lock-in far beyond what a pure automation tool would. This is the structural reason the company claims it grows "more essential the more you clamp down." 00:14:14

Customer-Built Automations Are a Free, Continuous Product-Discovery Engine

Because every workflow is generated code on a shared set of primitives, Serval can observe in aggregate what enterprises actually want automated. Stauch noted that customers' "password reset workflow for one company is gonna be very different, shockingly, than the password reset for another" 00:20:19 yet the platform sees the repeated patterns that signal real demand, such as license reclamation and stale knowledge base detection. 00:20:57 This inverts the usual customization-versus-product tension: heavy customization generates, rather than fragments, roadmap signal. It also hints at an eventual data advantage in understanding enterprise processes that incumbents with rigid workflow builders do not capture.