A 26-year-old venture capitalist at OMVC in Singapore has discovered that AI tools costing $120 monthly outperform a traditional intern in both speed and consistency. Marc Palet, who invests in over 12 AI companies, built custom AI agents using Claude Code and Replit that now handle meeting summaries, portfolio tracking, and content creation, freeing him to focus on deal flow and founder relationships.

How Did AI Replace Routine Operational Work?

Palet's transformation began when he shifted from using general-purpose AI tools like ChatGPT to coding-focused platforms. He started building automated workflows that pull data from multiple sources, including Gmail, Slack, call transcriptions, newsletters, and YouTube podcasts. An AI agent then parses this raw information and organizes it into structured portfolio company pages covering fundraising, financials, and commercial updates.

One of his first projects automated meeting summaries. Using Fireflies for recording, Palet receives structured summaries covering the company's problem, solution, and market information without manually listening back to each conversation. This single workflow eliminated hours of administrative work per week.

What Specific Tasks Did AI Agents Handle?

• Portfolio Intelligence: AI agents automatically ingest information from multiple channels and update portfolio company pages, rewriting content to maintain coherence rather than simply copying and pasting new data.
• Quarterly Reporting: One workflow automatically creates limited partner reports by summarizing portfolio activity during each reporting period, eliminating manual compilation work.
• Content Production: AI agents for LinkedIn posts increased Palet's output from one deep-dive article every two months to one every three weeks, directly generating more deal flow.
• Audit Support: Another agent gathers information from portfolio companies to streamline compliance and audit work.

The key insight, Palet noted, is that these agents don't just automate tasks; they learn his preferences and execute work the way he wants it done. Unlike an intern who requires weeks of training and typically stays only three months, AI agents maintain consistency and improve with each iteration.

How Does This Compare to Traditional Hiring?

Palet's monthly AI subscription bill of $120 covers Claude, ChatGPT, n8n (a workflow automation platform), and transcription APIs. For a small VC fund operating as a solo operator outside the US, this represents a fraction of what hiring even a junior employee would cost. More importantly, the time savings compound. By automating tedious operational tasks, Palet increased his capacity for in-person founder meetings and strategic work, the activities that actually generate returns in venture capital.

The turning point came when Palet read about OpenAI founding member Andrej Karpathy's approach to building self-updating knowledge bases. Inspired by that concept, he built a similar system for venture capital using Claude Code. This agent-driven approach transformed how he manages information flow across his portfolio, turning what would have been hours of weekly administrative work into automated processes that run continuously.

For Palet, the math is straightforward: AI plus his expertise is more powerful than him plus an intern. The tools cost less, work faster, and scale without the constraints of hiring, training, and retention. As AI coding platforms like Replit, Claude Code, and Cursor mature, this model may become increasingly common among small teams and individual contributors who need to punch above their weight.