Listen now on YouTube, Apple Podcasts, or Spotify.
👤 Guest Bio
Andrew Pignanelli is the Co-founder and CEO of The General Intelligence Company of New York, where he’s building Co-Founder, an AI chief-of-staff that helps individuals and teams run their businesses autonomously.
Before GIC, Andrew built Velvet, a retrieval-augmented intelligence system for private equity firms. Now he’s working on making it possible for one person to operate a billion-dollar company, powered by intelligent agents that know you, remember context, and take action.
PRESENTED BY AUTOSKILLS
Take your team from AI-curious to AI-ready in months days
Autoskills helps teams to go from AI-curious to AI-ready with:
→ AI acceleration sprints
→ Fractional AI automation engineers to build AI workflows
→ Custom AI transformations
Teams that work with Autoskills cut hours of repetitive work, identify high-ROI use cases, and leave with the confidence (and playbook) to scale AI responsibly.
Limited to 3 clients per quarter - book a free AI Readiness Audit today!
🎙Episode Intro
About a month ago, Andrew’s LinkedIn post launching Co-founder blew up. This week, we sat down to unpack the demo that sparked it and the product thinking behind it.
The conversation ranges across the product promise (an agent that remembers company context and runs long workflows), the two-agent architecture, and memory layers that make that possible.
We also dig into the organizational rhythms and experiment-first engineering culture that supported the build, the role of “taste” and examples in shaping agent behavior, onboarding and trial mechanics, and the broader idea of agents enabling new ways of working.
⏱ What’s Covered
(01:15) Andrew’s journey from Velvet to The General Intelligence Company
(02:27) The idea behind the Co-founder
(03:55) The viral launch and early traction
(05:30) Why “memory” is the last step to general intelligence
(12:04) How GIC structures agents
(15:30) Knowledge graphs, retrieval accuracy, and grounding memory in real data
(19:18) Letting engineers experiment
(23:55) Context engineering and “taste” as core product principles
(28:43) The building of Co-founder
(35:10) Why GIC runs vision-first
(41:27) How Co-founder makes entrepreneurship accessible
(48:42) AI, jobs, and ownership
(54:56) What’s next for Cofounder and the path toward true general intelligence
💡 Key Takeaways
Context is everything. AI tools that don’t understand your history or goals can’t make good decisions.
Build experimentation into the workflow. It’s important to give engineers room to experiment. Ask outcome-focused challenges (e.g., “make the agent 10% faster”) and let the team prototype quickly
Context engineering is a new discipline. The bottleneck for long-running, capable agents isn’t just retrieval, but deciding which 20 paragraphs of context matter for a task
Knowledge graphs are the foundation for real AI workflows. Instead of vector search alone, GIC builds structured relationships between people, projects, and data.
Taste matters as much as tech. Taste now serves as a product value — the human sense of what feels right in a response, interface, or decision.
Vision-first beats incremental. While others chase small automations, GIC bets on a broad vision that attracts top talent and deep believers.
📚 References & resources
LangChain — Popular framework for building agentic workflows and retrieval chains.
Devin — An AI software engineer / pair-programmer agent for engineering teams.
Claude (Anthropic) — Anthropic’s family of assistant models used for long-context reasoning and agent workflows.
GIC blog — Deep technical and product writeups from the GIC team on memory, context, and the architecture behind their agent.
🔗 Where to Andrew / Global Intelligence Company
X/Twitter: https://x.com/ndrewpignanelli
The General Intelligence Company of New York: https://www.generalintelligencecompany.com
👀 P.S. They’re hiring: https://www.generalintelligencecompany.com/careers
👉 If you found this episode useful, share it with a teammate or founder navigating AI adoption.
And subscribe to AI Ready for weekly lessons on how leaders are making AI real at scale.
Until next time,
Haroon