How To Use AI In Your Startup

YC Spring Batch & Strategic Directives

• The application deadline for the first YC Spring Batch is February 11th.
• Accepted startups receive a $500,000 investment and access to Y Combinator's network.
• Founders are advised to consider relocating to the San Francisco Bay Area (temporarily or permanently) to access real-time market intelligence and community insights.
• Temporary presence of 3–4 weeks in SF (attending hackathons, visiting neighbors) is recommended if permanent relocation is not immediately feasible.

AI Strategy & Fundamental Principles

• Founders should not pivot a business solely because it is fashionable to use AI; fundamental value creation requires solving real customer problems.
• However, almost every new company should have LLMs at the heart of its product or internal operations.
• Leveraging AI is now as essential for startups as adopting the cloud was in the early 2010s.
• Switching an idea to simply make API calls to OpenAI without deeper customer insights or environmental changes will not alter a startup's trajectory.
• Big companies are moving slowly; despite ChatGPT being two years old, legacy products (e.g., Alexa) have not yet evolved significantly, leaving a window for startups.
• Founders are encouraged to observe friends working in target industries and analyze their repetitive tasks and screen usage to identify automation opportunities.

Case Studies: Success Through Pivot & Location

• WAPI (formerly InvestBetter): Founded in 2020 as a financial platform, pivoted in early 2021 to a Zoom productivity tool, then moved from Toronto to San Francisco in early 2023.
• WAPI stopped working on its previous product to build a voice AI solution, growing from zero to powering a significant portion of YC and non-YC voice AI companies within 15 months.
• The company's success was driven by embedding itself in the SF community, which allowed founders to identify the shift toward Voice AI earlier than remote competitors.
• Conversely, companies that pivoted to AI without changing their operating environment, community, or customer insights (e.g., generic customer support agents) struggled.

Specific Use Cases & Industry Opportunities

• Replexica: Automates UI localization by translating user interfaces into different languages using AI.
• Gecko Security: Acts as an AI security engineer, allowing software engineers to manage security independently within the release cycle.
• Medicare Advantage Pilot: Built by former insurance tech employees to create an AI co-pilot for agents, addressing obscure but critical workflows.
• Pre-auth Automation: LLMs are used to summarize doctor data and automatically generate pre-authorization requests for payment portals, replacing manual data entry between legacy systems.
• Patient Engagement: Voice AI is deployed to call patients between visits to check on health status and schedule follow-ups.
• Healthcare Admin: The US healthcare system spends ~$1.3–1.4 trillion on admin out of $4 trillion total; much of this involves humans manually moving data between legacy software portals.

Market Dynamics & Historical Parallels

• The current AI wave mirrors the 2007 mobile shift; founders who missed the mobile cycle (due to early timing or lack of visibility) may struggle to see the current AI trajectory.
• The opportunity lies in replacing legacy on-prem software with new, AI-native versions, similar to how Workday emerged from the shift from PeopleSoft to the cloud.
• Proximity to leading AI companies (e.g., "next door" neighbors) allows founders to learn best practices instantly, a barrier for those operating remotely in locations like Chicago.
• The "state of the art" in AI changes monthly; remote observation often results in a one-year lag in understanding what is actually possible and useful.