Discover how Artemis uses AI-native technology to detect and stop AI-powered attacks. Learn from founders Shachar Hirshberg & Dan Shiebler on building next-g...
AI vs AI Security: How Artemis Fights Modern Cyberattacks with Intelligent Defense
The cybersecurity landscape is shifting dramatically. As attackers leverage artificial intelligence to accelerate threats and sophisticate their methods, defenders face an unprecedented challenge: to protect against AI-powered attacks, you need AI-powered defense. This is the core philosophy driving Artemis, a newly launched security operations platform that's fundamentally reimagining how enterprises detect and stop modern threats.
Artemis has emerged from stealth with significant momentum: $70 million in combined Seed and Series A funding, a team of 30 people built in just seven months, and first customers who approached the founders asking to buy—before being asked to purchase. This success story reveals essential insights about building AI-native companies, hiring in the AI era, founder-market fit, and the future of enterprise security operations.
Core Summary: Key Insights About Artemis and AI-Native Security
- 70% of Artemis customers replaced legacy security solutions entirely, demonstrating that enterprises recognize the urgency of AI-native defense as threat sophistication accelerates
- Product-market fit signals emerged in weeks, not months: customers added more team members and data sources immediately after experiencing initial value, indicating genuine product resonance
- AI-native architecture is fundamentally different from AI-enabled additions: building AI into core systems from inception enables dramatically superior performance compared to legacy platforms with bolted-on AI features
- Founder-market fit proved critical to early success: Shachar's 15 years in cybersecurity and Dan's decade in security operations gave them deep customer understanding that informed every product decision
- Unconventional hiring approach scaled to 30 people in stealth: emphasizing references, fast interviews (2 hours total), and AI fluency assessment over traditional credentials attracted top talent despite lack of public visibility
What Artemis Does: Redefining Security Operations for the AI Era
Artemis solves a critical problem facing modern enterprises: detecting and stopping attacks across an organization's entire technology stack—cloud infrastructure, identity systems, networks, and information security—before attackers can cause damage.
Traditional security information and event management (SIEM) systems operate on outdated assumptions. They assume security analysts have time to investigate thousands of alerts daily. They require weeks to months for security teams to understand their detection coverage. They place the burden of threat identification on human security experts operating within manual workflows designed in an era when attacks moved slowly.
Artemis fundamentally reimagines this challenge by being AI-native from the ground up. Rather than layering AI capabilities onto legacy security platforms, Artemis built its entire architecture around what AI systems do best: rapidly processing massive data volumes, identifying subtle patterns, reasoning across complex systems, and delivering actionable insights with minimal human intervention required.
The results in production demonstrate the power of this approach. Artemis customers spend 3-4 hours daily in the product actively using it to achieve security outcomes—a stark contrast to traditional security tools that sit unused or generate alert fatigue. This engagement level signals genuine product-market fit and customer value, not forced adoption.
Why the Founders Chose This Market: Deep Expertise Meets Urgent Problem
Building a successful company requires founder-market fit: you must genuinely understand the problem space at a depth competitors cannot easily replicate. For Shachar and Dan, this foundation was built over years of direct experience with the core challenge Artemis addresses.
Shachar spent 15 years in technology and cybersecurity, with the past decade focused specifically on security operations. He worked at AWS and Palo Alto Networks, observing how attackers continuously evolved their methods and how defenders struggled with existing tools. This experience meant he knew customers deeply—not just their stated needs, but the problems they didn't yet recognize they would face.
Dan's background at Abnormal Security (where he worked on behavioral AI-driven threat detection) provided complementary expertise. At Abnormal, Dan witnessed the earliest waves of AI transforming the threat landscape: when generative AI made it trivial for attackers anywhere to create topically relevant phishing emails at massive scale. He saw how AI-generated attacks accelerated and sophisticated faster than traditional defenses could adapt.
The timing felt urgent. In 2023-2024, threat acceleration became undeniable. CrowdStrike published research showing attack speeds dramatically increased. Entropic released similar findings. Cloud complexity multiplied. Traditional approaches were visibly failing. Shachar and Dan recognized that waiting would mean missing a market inflection point where defenders desperately needed new solutions.
But timing alone doesn't drive founders to leave successful careers. The conviction came from extensive ideation. Shachar credits First Round's PMF Method program for teaching him how to systematically think through company building. He and Dan spoke with hundreds of security professionals. They repeatedly asked: "If AI became 10x or 100x better, how would this problem change?" They examined which existing solutions were genuinely helping customers and which represented dead ends.
This deep exploration led them to their ideal customer profile (ICP): mid-market and enterprise companies that had already implemented traditional SIEMs and felt acute pain from their limitations. These weren't companies interested in trying new technology for its own sake. They were companies drowning in alert noise, uncertain about detection coverage, and feeling genuine pain from tools that couldn't keep pace with modern threats.
Building an AI-Native Company: Architecture, Team, and Execution
Artemis didn't just apply AI to security problems. The founders structured the entire company to operate as an AI-native organization where human and artificial intelligence work together as fundamental operating principles rather than late-stage optimizations.
AI-Native Architecture: The Foundation Matters
The difference between AI-enabled and AI-native is profound. Legacy security platforms were built around traditional logical decision trees and workflows. When companies add AI to these systems, they face architectural bottlenecks: different AI components struggle to communicate; information must be compressed into legacy data formats; the AI pieces are constrained by foundational systems designed before AI existed.
Artemis chose a fundamentally different path. Every architectural decision was evaluated through one lens: "Does this make AI systems more or less likely to be correct?" This principle cascades through entire technology decisions. The team prioritized architectural patterns that let AI components communicate fluidly, reason across complex scenarios, and maintain coherence as the system learns an organization's environment.
The practical result: engineers can work with AI coding tools at a pace and scale that would be impossible in traditional codebases. The codebase is structured so AI agents can effectively debug issues, understand unfamiliar code, and make high-confidence changes. The team runs 4-8 Claude Code instances in parallel, simultaneously shipping 4-5 features with a team that would traditionally require months to deliver one.
Hiring in the AI Era: References Over Credentials
As venture-backed security startups typically need 6-12 months to hire, Artemis's rapid growth to 30 people while in stealth required rethinking hiring entirely. The founders identified that traditional hiring processes—lengthy interviews, take-home coding challenges, credential-focused screening—don't predict success in AI-native environments.
Instead, Artemis inverted the process:
References became the primary signal. The team discovered near-perfect correlation between reference strength and actual performance. They weight references heavily and front-load reference calls early in the process. This approach reduces risk dramatically—you're hearing from people who've actually worked with candidates, not making inferences from interviews.
Interview process compressed to 2 hours maximum, split between brief team conversations and meetings with founders. The goal wasn't to stump candidates or test obscure knowledge—it was to assess capability, mission alignment, and whether the person genuinely wanted to work in an AI-native environment.
Work projects were repurposed for learning, not filtering. When Artemis hired interns, rather than requesting "complete this coding challenge and we'll review your code," they asked candidates to "build a website that accomplishes X and send us a link." They didn't examine the code—didn't care if candidates used no-code tools, frameworks, or pure code. The assessment was simple: can you build something from nothing? That matters far more than any technical detail.
AI fluency became a learnable skill rather than a prerequisite. Early in the process, interviewers asked candidates about AI tool usage. Nearly everyone claimed extensive experience. But deeper investigation revealed that most people hadn't used these tools remotely as intensively as Artemis expected. The insight: prior AI experience is a poor predictor of how someone will leverage AI in the future. The team looks for builders with genuine curiosity, intellectual humility, and willingness to learn cutting-edge capabilities.
The hiring process moved with remarkable speed. Offers extended within 1-2 days of initial conversation. Candidates moved to the front of decision queues because the process respected their time. The company culture emphasized that the best people always have multiple options; making their choice easy—through excellent candidate experience and speed—became a hiring advantage rather than something to optimize after selection.
Building Culture in an AI-Native Organization
As a small company scaling rapidly, Artemis was deliberate about defining culture early. The team identified five core values, kept that number intentionally small, and reinforced them weekly through specific callouts in all-hands meetings.
Customer obsession sits at the center. Not as a buzzword, but as genuine operational commitment. Founders text customers directly when issues arise. Every person in the company can escalate customer concerns to leadership immediately. The commitment isn't polite—it's genuine: customer happiness is "the single most important thing we are focusing on because if that works, everything else will follow."
Ownership means individuals drive projects from conception through delivery without deferring responsibility. When you see something broken, you fix it. You don't say "that's not my job." In an AI-native environment where tools expand what any individual can accomplish, ownership becomes even more powerful.
Intentionality requires thoughtful decision-making about tradeoffs. Even in high-velocity environments, people think through implications rather than moving recklessly. The team distinguishes between one-way doors (major, difficult-to-reverse decisions requiring deep thought) and two-way doors (reversible decisions made quickly with post-decision reflection).
Velocity and ** high standards** form a paradox that AI-native operations resolve. Traditionally, shipping fast meant cutting corners. With AI tools dramatically reducing build time, Artemis proves you can move at exceptional speeds while maintaining excellence.
These aren't values printed on walls. They're reinforced through hiring decisions, conflict resolution, and constant reinforcement in company meetings. New team members see founders and leaders exemplifying these values daily.
Acquiring First Customers: From Design Partners to Enterprise Revenue
Most startups struggle to find even one customer willing to bet on unproven technology. Artemis had a different challenge: the first three customers approached the founders asking to purchase before being asked to sell.
This outcome wasn't accidental. The founders spent months working with potential customers as design partners. These weren't casual conversations or exploratory pilots. Shachar and Dan brought security leaders deeply into the product vision, asked for feedback, iterated based on genuine insights, and built products that directly solved their partner problems.
The trust earned through this process was crucial. Security teams at large enterprises don't lightly connect sensitive data sources to new vendors. They require confidence that value will flow in exchange for the political capital required to enable integration. Artemis earned that confidence by making integration as seamless as possible and delivering immediate insights.
The conversion pattern that emerged revealed genuine product-market fit. Customers would connect a few relatively easy data sources—initial setup taking hours, not days. Within days of receiving first insights, they'd add multiple team members and request connections to additional critical data sources. This escalation from initial exploration to deep commitment happened rapidly and repeatedly.
This pattern differed dramatically from traditional security sales where pilots stretch for months with vague outcomes. Here, the question was simple: "Do you immediately want more access and more features?" Affirmative answers came fast.
By late January 2024 (just 4 months into building), the product had reached sufficient maturity that usage metrics skyrocketed. The product was solving real problems at a depth customers immediately recognized. Engagement metrics that had been steady became parabolic. The founders knew they had product-market fit.
Go-to-Market: From Founder-Led Sales to Scalable Process
The first sales came from founder-led effort—hundreds of customer conversations where Shachar and Dan learned the market, understood pain points, and built trust. This direct engagement remains valuable. Shachar texts with every customer daily. He's accessible for escalations and relationship-building at a personal level.
The challenge: founder sales doesn't scale. With the company exploding in growth, Artemis needed to build repeatable sales motion without losing the relationship depth that drove early success. The team recently hired their first account executives, sales engineers, and go-to-market engineer to build scalable processes.
The challenge is real: conversion rates will likely decline initially as new salespeople lack the deep conviction and credibility that founders bring. But the founders aren't retreating from customer relationships. The goal is that customers five years from now feel comfortable texting leadership directly when issues arise—that customer access remains a competitive advantage even as the company scales.
One customer expressed concern that founder attention would fade as the company grew. The response was unambiguous: "If anything is wrong, you should tell us. Your happiness is the single most important thing we're focusing on." This commitment to customer success at scale, not just in the early days, is rare and valuable.
AI vs AI: The Future of Security Operations
The broader threat landscape that drove Artemis's creation continues accelerating. Attackers are using AI to move faster, generate more sophisticated attacks, and exploit vulnerabilities at scales previously impossible. Defenders who rely on human-speed response simply cannot keep pace.
This reality drove Artemis to market now rather than waiting for optimal conditions. The founders observed that attackers will go after financial gain and will use the latest technology. On the defender's side, you must adapt—it's not optional. Without adaptation, you will be compromised.
Interestingly, this urgency has translated into market acceptance. About 70% of Artemis customers have already replaced legacy security solutions entirely with the platform. They've decided that traditional tools cannot possibly defend against modern threats. The other 30% use Artemis to augment legacy systems while planning complete migration within 2 years—recognizing that current tools "won't cut it" but needing time to manage the transition.
The future looks different. In three years, the founders believe most work—across industries but especially in cybersecurity—will be directed by humans but executed by AI. Attackers will deploy AI agents that reason through multi-stage exploitation at machine speed. Defenders need equivalent capabilities. This isn't a possibility; it's a necessity.
Founder Lessons: First-Time Founders Navigating Rapid Growth
Both Shachar and Dan were surprised by how natural the 0-20 founder phase felt. Coming from larger companies (Dan from Abnormal around 200 people, Shachar from AWS), they expected the scrappy early stage to feel radically different. Instead, the principles remained consistent: build the right product, serve customers obsessively, hire excellent people, move fast.
What was genuinely surprising was discovering their own alignment with the founder role. Shachar notes: "I feel like I was built for this, and I'd always felt like I was built for this." The conviction that company-building was his path proved accurate rather than merely aspirational.
Both describe the workload as genuinely challenging—extended 100-hour weeks—but rewarding rather than draining. Dan notes he's "having the best time of my life, living my dream." Multiple founders claim they "can't wait to sell the company and get an easier job," yet both Shachar and Dan wouldn't trade the experience away.
For others considering founding, the advice is straightforward: if building a company is genuinely what you want to do, pursue it. The timing can vary—some people should found immediately after college; others benefit from years of experience in larger organizations first. But when you genuinely feel ready, hesitation becomes the greater risk than action.
Areas for continued growth include learning how to lead teams with dramatically different skill sets and functions. Shachar has managed software engineers but never sales professionals at this level, or product managers with distinct workflows. Learning to lead across diverse functions while maintaining company culture represents a multi-year development journey.
Dan focuses on an even broader question: what does leadership look like in an AI-native company? There's no playbook for this. As Artemis grows and builds executive teams over the next 1-2 years, the leadership model itself will be novel. Dan thinks deeply about how AI will change work, how companies should be structured to leverage AI effectively, and how to position Artemis as a leading example of AI-native company operations.
Broader Implications: Why AI-Native Companies Have Structural Advantages
The Artemis story illuminates a fundamental shift in startup strategy. The question isn't whether to use AI in your business—it's whether to build your entire business around AI from inception or try to retrofit AI into legacy structures.
Companies that choose the retrofit path face permanent constraints. Their foundational systems were designed without AI in mind. Adding AI capabilities later means building bridges between legacy systems and new AI components. Information must flow through old data formats. Decision-making still defaults to traditional logic when legacy systems require it. The result is perpetual compromise: you move faster than before AI, but not nearly as fast as truly native systems.
AI-native companies built from zero can operate at scales that seem impossible to legacy-focused competitors. Artemis's engineering team ships 4-5 features simultaneously using AI coding tools in ways that would require 2-3x larger traditional teams. The codebase structure itself enables this through architectural decisions made with AI capabilities in mind.
This advantage extends beyond engineering. If your company's core functions—hiring, sales, product development, customer success, operations—are designed around what AI does best, you can accomplish exponentially more with smaller, more efficient teams. Artemis grew to 30 people in 7 months while in stealth, acquiring $70 million in funding and signing enterprise customers, on a trajectory that would typically require 60+ people and visible market presence.
For security specifically, this shift is existential. Attackers using AI can reason through exploitation scenarios, modify attack patterns in real-time, and operate at velocities where human security analysts cannot possibly respond. Only AI-native defense systems can reasonably keep pace. Legacy security vendors bolting AI onto existing platforms are functionally obsolete—they're trying to address 2025 threats with 2020 architecture.
Conclusion: The Future Belongs to AI-Native Builders
Artemis's emergence from stealth with $70 million in funding and immediate enterprise traction represents more than a successful security startup launch. It demonstrates that the most significant opportunity in 2025 isn't using AI to improve existing businesses—it's building entirely new companies where AI is foundational rather than supplementary.
The security industry needed this disruption. Enterprises were drowning in alert noise from tools designed for slower threat landscapes. They needed AI-native defense that could operate at machine speed while humans directed intent. Artemis delivered this.
For founders, Artemis models how to think about problem selection, team building, and execution in the AI era. For investors, it shows that founder-market fit, combined with AI-native architecture, can compress typical startup timelines dramatically. For enterprises, it signals that legacy security solutions will struggle increasingly as threats accelerate.
The AI versus AI war is no longer theoretical—it's the current reality of cybersecurity. Defenders who move to AI-native operations gain structural advantages over legacy competitors. The next three years will likely see dramatic consolidation as enterprises recognize that traditional tools simply cannot defend modern environments. The winners will be companies that were built for exactly this moment.
Original source: Inside Artemis' "AI vs AI" war | Shachar Hirshberg & Dan Shiebler (Co-founders, Artemis)
powered by osmu.app