E413: How AI Will Reinvent Venture Capital

Episode Summary

Executive Summary: The conversation explores how Footwork is building an AI-native venture capital firm from the ground up, using agents to automate sourcing, CRM updates, research, and internal knowledge management. The founders argue AI improves preparedness and judgment today, will shrink VC back offices over time, and may lower barriers for new managers, while preserving human relationships and final decision-making as the last durable edge.

Main Topics: Defining an AI-native VC firm: AI-native means redesigning the firm’s operating model around AI rather than bolting AI onto old workflows. Footwork aims for agents to update data, brief partners, and proactively surface context without humans manually checking systems. Footwork’s AI stack and evolution: The firm moved from using ChatGPT/Claude, to local desktop agents, to cloud-hosted always-on agents backed by a central Postgres database that stores firm context and enables background automation. AI across the VC workflow: AI is being applied to the core venture functions—find, decide, win, help, exit—as well as firm-building tasks like finance, LP relations, and fundraising. Agents prep meetings, summarize board context, and retrieve portfolio information. Human-AI interaction and CRM automation: Footwork still uses Affinity as a source of truth, but humans no longer manually update it. Agents ingest meeting notes and emails, then update pipeline status and decision history automatically. Organizational adoption and change management: The firm emphasizes broad AI literacy, weekly sharing of use cases, and empowering people to discover their own ‘aha moments.’ Footwork also hired a dedicated AI lead to drive internal automation and experimentation. Implications for VC structure and talent: The speakers argue that entry-level analyst work and much of the back office can be automated, leaving fewer people needed overall and shifting human labor toward judgment, relationships, and founder interaction. Brand, network, and public building: Writing and building in public are framed as compounding advantages for sourcing, fundraising, and serendipity. The discussion weighs transparency against secrecy and concludes the optimal approach depends on strategy and stage.

Key Arguments: AI-native firms redesign processes from scratch; they do not merely add AI tools to old workflows. Agents can keep CRM and pipeline data updated automatically, reducing manual work and improving context continuity. A centralized context layer is essential for AI agents to function effectively across firm knowledge. AI currently improves preparation and judgment, but it is not yet ready to fully replace human decision-making in investment choices. The most valuable human capabilities remaining are relationships and judgment; these are likely to be last to automate. In venture, the ability to learn quickly has become a more important founder trait than before because the environment is changing so fast. VC will likely need fewer humans as analyst and back-office tasks become automatable, while remaining firms become more efficient. Building in public can create brand, network effects, and serendipitous sourcing, though the right level of openness depends on the business. Brand and network are the two strongest compounding moats in venture capital. Exceptional companies may require exceptions to standard fund strategy or portfolio construction rules. Starting a firm solo is becoming more feasible because AI can provide leverage across many operational tasks.

Data Points: Footwork team size: 5 full-time employees - The firm is intentionally small, which makes AI adoption easier and cheaper to implement. Number of aggressively used agents: 3 or 4 agents - Footwork currently runs a small set of agents heavily in its workflow. AI cost: Thousands of dollars monthly - Current organizational AI spend remains relatively small for the firm. Fundraising timeline for Fund One: Raised in early 2021; oversubscribed by March; closed April 13, 2021 - Used to illustrate the speed and momentum of the firm’s first fundraise. Weeks of deployment by a portfolio-company team: 5 to 6 weeks - Referenced in an example of operators deploying into portfolio companies to solve AI bottlenecks. Time since the LLM era began: About 3.5 years - Used to argue that models and workflows are still early and will improve significantly. Number of venture firms historically: About 3,000 firms - Cited in a discussion about industry fragmentation and potential manager extinction. Fundraising base: 7 endowments as LPs - Footwork reportedly attracted elite institutional capital early in its life.

Pivotal Quotes: "What it means to be AI native is that you're not just sticking AI into the firm's existing processes, hoping that something's going to change. It's actually thinking about the whole firm from the ground up with AI at the core." — Nikhil: Defining the concept of an AI-native venture capital firm. "An AI-native firm is where the humans in the firm don't even look at the CRM because that's updated agentically, it's happening in the background." — Nikhil: Explaining how AI replaces manual CRM maintenance and pipeline tracking. "The trait in founders that we look for more than ever is their slope of learning." — Nikhil: Discussing what matters most when evaluating founders in a fast-changing AI environment.

Implications: VC firms will likely become smaller, faster, and more automated, with AI handling research and ops while humans focus on judgment and relationships. Founders and managers who learn fastest and build AI into core processes may gain a lasting edge.