Solving the Blank Canvas Problem: Gusto's AI Co-Founder

Gusto Co-Founder is designed to automate recurring work such as running payroll, approving time off, and reminding staff to submit timesheets. Kim says it can also handle tasks beyond Gusto, such as checking the weather and notifying customers of a tour business when rain is forecast. Its suggested automations build on payroll, HR, and scheduling tasks customers already do, addressing the blank canvas problem of open-ended AI tools.

After setting up OpenClaw, Kim found it useful to text an AI but struggled to decide what to ask it to do. During a five-hour layover, he used Claude Code to prototype a prompt-based tool that built custom web apps in Gusto’s design system. The idea shifted toward automating recurring workflows, using Gusto’s customer and industry data to suggest relevant tasks that run on a regular heartbeat or schedule.

Customers readily understand automation when it addresses work they do every week. Kim describes a massage spa that moves data from Mindbody into Google Sheets to calculate commissions, tips, and hours before running payroll. Gusto Co-Founder can send a summary for approval by text or Slack, and Kim says the product had 500 customers at launch after feedback from a 20-customer Small Business Council.

Kim sees saved time helping owners focus on growth, and wants Co-Founder to flag tasks businesses may be missing. He gives the example of identifying a possible R&D tax credit and preparing forms for review; Gusto found $50,000 in credits for Cabana Pools. The roadmap includes more messaging channels and industry connectors, and eventually support for people starting a business before they have a Gusto account.

A team of five, including Kim, built the product over 10 weeks using AI coding tools; a designer and engineers shared coding and design work. The team minimized meetings, documents, and planning rituals, and often built prototypes before discussing them, even when the code was later discarded. Kim says fast experimentation still requires discipline about which features reach customers.