CNBC Reporter Recreates Monday.com Using Anthropic Claude AI in One Hour

CNBC reporters Deirdre Bosa and Jasmine Wu used an Anthropic Claude product to “vibe-code” a working monday.com-style project manager in roughly one hour, connecting the prototype to a Gmail account and a calendar and spending an estimated $5–$15 in compute credits. The hands-on experiment produced multiple boards, team assignments, status dropdowns, calendar integration and reminders, and it prompted immediate market and community reaction.

Deirdre Bosa publicized the experiment on Twitter, writing at 4:39 PM on Feb 4, 2026, “ok WOW. Woke up this morning and said, for fun, lets try to recreate monday. com w Claude cowork. it wont work or anything, but we can just show our audience that its plausible. 1 hour later... I literally have my own monday. com that's plugged into my calendar & gmail and” The tweet registered heavy engagement in one replay of the experiment, showing 389K views, 144 replies, 94 reposts and 1.79K likes in the metrics reproduced by one outlet.

The reporters began by telling Claude to build a project-management dashboard with “multiple project boards, assigned team members and a status dropdown,” then asked the model to research Monday.com and reproduce core features. Techbuzz Ai summarized the result: “The AI agent spit out a working prototype in minutes.” CNBC’s write-up detailed how the AI autonomously added calendars and other functionality, and the clone surfaced an overlooked calendar invite and added reminders to book tickets and sign a waiver for family events.

Time and cost estimates are consistent across sources: the build was reported as “under an hour” or “in less than an hour,” and compute-credit costs were reported in a cluster around $5–$15, with some reports saying simply “less than $15.” Jasmine Wu conceded the low bar for expectations in the experiment, writing, “We didn't expect to get anywhere,” in the published account of the hands-on session.

Markets reacted quickly. One analysis tracked by observers said Monday.com’s stock traded down about 6% after the tweet, then later fell another roughly 10% before a 2% bounce, and estimated that roughly $300 million of market value was erased in a 30-minute window; that same write-up described the company as generating $1.3 billion in annual revenue across 250,000 customers. Other coverage framed the episode as part of a broader SaaS selloff, with one outlet calling it evidence that “the software sector's worst nightmare just got real,” and citing a 30% selloff across enterprise SaaS names in the week’s moves.

CNBC TechCheck and other reporting flagged which vendors look most exposed and which may be more resilient: Atlassian, Adobe, HubSpot, Zendesk and Smartsheet were listed as tools “that sit on top of the work,” while cybersecurity firms such as CrowdStrike and Palo Alto Networks and systems-of-record like Salesforce were described as harder to clone quickly. The segment included an interview reference with Vercel CEO Guillermo Rauch on the shifting landscape of developer tools.

Online communities responded with a mixture of amazement and skepticism. Reddit threads quoted users worried about security and scalability — “I'd love to hear how secure this vibe coded app is...” — and others mocking the demo’s polish: “I'm sure its a 1:1 comparison.. totally sure.. AI definitely didn't add extreme jank anywhere.... totally not.” The experiment also became part of a broader labor-market conversation that noted recent layoffs and AI-driven restructuring at firms cited in commentary.

The experiment did not claim production readiness for enterprises, but it crystallized a practical question for Monday.com, competitors and investors: if reporters with no engineering backgrounds can spin up a working workflow app in about an hour for a handful of dollars, SaaS companies must clarify differentiation, security posture and enterprise resilience as they defend valuations and customer trust.