Adept
Adept emerged from a fundamental insight: most office work involves repetitive digital tasks that follow predictable patterns across software tools. Rather than building point solutions for individual applications, the founders envisioned training AI to understand and execute these tasks the way humans do—by reading screens, understanding context, and taking action through standard interfaces like browsers and applications.
The team developed Act One, their flagship AI agent that sits as a floating dialogue interface (similar to a search bar) in the user's browser. Instead of clicking through menus and filling out fields manually, users simply type what they want in plain English. Act One then navigates to the relevant websites, logs in, performs multi-step workflows, and completes the task. In demonstrations, the system shows remarkable capability: adding a new lead to Salesforce with custom fields, creating Excel formulas from plain English descriptions, and highlighting data based on conditional logic—all requiring 15-20+ manual clicks reduced to a single natural language prompt.
The product resonated strongly enough to attract elite venture capital. Greylock and investor Scott Belsky led funding, ultimately raising $56 million to scale the vision. The founding team of 8 includes a former VP of Engineering from OpenAI, lending significant AI credibility. The product was not yet publicly released (as of the podcast date), but the team maintained a waitlist suggesting strong pent-up demand. The interface design—keeping the AI assistant lightweight and accessible rather than replacing entire applications—proved strategically sound for adoption.
Adept was actively hiring and positioned as a high-risk, high-reward opportunity for early engineers willing to bet on transformative AI infrastructure. The company's location in Nicaragua raised curiosity but didn't appear to materially impact investor confidence or the technical vision. With Act One still in limited release, the company faced the critical challenge of scaling API reliability and expanding task coverage across the thousands of enterprise software tools that businesses rely on daily.
- •The founding team's deep AI expertise from OpenAI provided credibility and technical capability to solve a genuinely hard problem that required breakthrough AI capabilities rather than incremental engineering.
- •The product-led-growth approach with a lightweight browser interface lowered friction for adoption by meeting users where they already work rather than forcing them to learn new tools or replace existing workflows.
- •The insight that office work is fundamentally repetitive and pattern-based unlocked a massive TAM spanning every enterprise software user, rather than solving for a single vertical or use case.
- •Positioning the AI as a general task executor accessible via natural language tapped into strong pent-up demand signaled by waitlist interest, which attracted elite capital before needing to prove unit economics.
- 1.Recruit technical founders or early engineers with proven expertise in the core technology domain you're building in—in this case, hiring from frontier AI labs provided both capability and investor conviction.
- 2.Design your interface to integrate into existing user workflows non-invasively rather than requiring adoption of new software, making it low-risk for users to try and reducing switching costs.
- 3.Identify a broad, horizontal problem that appears across many use cases and industries rather than starting with a vertical solution, which multiplies your addressable market and appeal to generalist venture investors.
- 4.Build a waitlist and maintain direct signals of demand before full launch, using that pent-up interest to accelerate fundraising and validate that the problem resonates before scaling distribution.
Similar Companies
247.ai
$25.0M/mo247.ai, founded by PV Cannon in 2000, is an AI-powered customer service automation platform serving over 150 enterprise customers with $300M+ in ARR. The company raised only $20M from Sequoia (2003) and bootstrap, achieving 10% net profit margins while maintaining a 12-month CAC payback period and 100% net revenue retention. Despite a security breach setback around 2018, 247.ai has recovered and recently achieved 20% new revenue booking growth in their best quarter.
iCIMS
$13.3M/moiCIMS is a bootstrapped SaaS provider founded in 1999 that dominates the talent acquisition software market as the #2 player, serving 3,500 enterprise customers with an average monthly spend of $4,000. The company exited 2017 with $160M ARR and is targeting 25%+ annual growth while maintaining profitability, recently acquiring Text Recruit to expand into candidate messaging and recruitment advertising.
Zoom
$12.0M/moZoom is a freemium SaaS video conferencing platform founded by Eric Yuan in July 2011 after he left Cisco to build a next-generation collaboration solution. The company has grown to 850,000+ paying customers across individual, SMB, and enterprise segments, generating over $12M in monthly recurring revenue with approximately 100% year-over-year growth. Rather than focusing on customer stickiness or aggressive growth targets, Zoom emphasizes customer happiness and organic word-of-mouth acquisition, which has proven highly effective in driving viral adoption.
Madwire
$10.0M/moMadwire is a comprehensive SaaS platform for small businesses (1-100 employees) that combines CRM, payments, invoicing, billing, e-commerce, and multi-channel marketing tools in a single platform. Founded in 2009, the company has grown to $120M ARR serving 20,000 customers with an average revenue per user of $500/month, while maintaining strong unit economics ($3,000-$4,000 CAC with 3-month payback) and recently turning profitable with a focus on reaching 15-20% EBITDA margins. The company is exploring an IPO within 12-18 months without having raised substantial capital beyond an initial $7.5M.
SwiftPage
$7.0M/moSwiftPage is a CRM and marketing automation platform founded in 2001 that targets small businesses. Under CEO John Oshel's leadership since 2012, the company scaled from 60,000 customers with $26.2M revenue in 2015 to 84,000 customers today with an estimated ARR of $36M+, maintaining 1.5% monthly logo churn and a 6-7 month payback period with a sub-$500 CAC.