Edward Kim, co-founder and CTO of Gusto, recounts how a failed first startup, a half-marathon run on a bet, and a childhood watching his mother run the back office of his father’s medical practice converged to create a $9.5B payroll platform — and how, after a decade as a manager, AI tools pulled him back into writing code to build “Co-Founder,” an AI agent that automates the same back-office work his mother once did by hand.
Childhood in the back office seeded the problem Gusto would solve
Edward’s parents were Korean immigrants; his father ran a private medical practice in Upland, California, and his mother handled every back-office function: payroll, scheduling, insurance billing, taxes, compliance.
After school, Edward and his sisters would sit in the back room doing homework while his mother finished the work that kept the business legal and running.
He internalized that running a small business meant a massive burden of administrative work unrelated to the core craft — a problem he later recognized as “much bigger” than the utility apps he was building.
A winding path to Gusto: failed startup, Android apps, and a half-marathon bet
Edward studied electrical engineering at Stanford after a serendipitous roommate introduced him to CS; he had originally wanted to study film.
His father once handed him a printed Paul Graham essay (“How to Start a Startup”) in the car, which lay dormant until years later.
First startup (2008, YC batch with Gary Tan) shut down; Edward spent 1.5 years building Android apps solo, peaking at ~$700k/year (notably “Screenshot It,” a $7 app that let early Android phones take screenshots).
He realized these apps solved only temporary, shallow gaps that the platform would eventually fill — not the deep, enduring problem he wanted to tackle.
On a whim, he ran the San Francisco half-marathon untrained, limped the last three miles, and randomly ran into college friend Josh Reeves at the start line.
They reconnected over coffee at Crema on Fourth & King, discovered both had exited first ventures and wanted to build something meaningful, and decided to start a company together.
Serendipity brought the third co-founder, Tomer London
Josh had been paired with Tomer (a new Stanford PhD student from Israel) through an alumni-newcomer networking program; they became roommates in Palo Alto.
Josh introduced Edward to Tomer; they shared values and upbringing, and the three agreed to co-found Zen Payroll (later Gusto).
Edward reflects that none of these connections were manufactured — luck provided the meetings, but readiness and openness let them become a founding team.
Explicit co-founder role negotiation prevented later founder conflict
Immediately after YC acceptance, the three sat down, acknowledged everyone secretly wanted to be CEO, and honestly assessed skills and desires.
Edward advises early teams to have this uncomfortable conversation upfront; ~70% of startup failures he sees stem from founder implosions, not market or product issues.
The technical co-founder’s ego trap and the long road to management
In the first YC batch, Edward (technical) coded 16–18 hours/day while the business co-founder pitched investors over steak dinners — a perceived unfairness driven by ego.
Gary Tan (technical himself, later YC partner) helped him see that contributions even out over years if trust exists; you cannot judge a single point in time.
~1.5 years into Gusto (~10 engineers), Josh confronted Edward: the company needed him to manage, not code. Edward initially resisted (“you don’t understand the bugs”), then accepted.
He sought mentors (Aditya Agarwal, CTO of Dropbox; Steve Herrod, CTO of VMware), read books, and forced himself to learn management — discovering a virtuous cycle: competence bred enjoyment.
Practical ego-management tactic: recognize your physical trigger (shortness of breath, heat, cold) as a fight-or-flight signal, pause (e.g., “I need the bathroom”), reset, then choose a proactive response.
AI reopened the door to hands-on building after a decade away
In early 2024, Edward experimented with OpenClaw (an open-source AI agent) on a Mac Mini; he loved its Telegram chat interface and “heartbeat” (recurring proactive tasks) but didn’t yet see the full hype.
On a delayed return flight from Madrid (stuck 5.5 hours at Heathrow), he opened Claude Code and prototyped an idea: let Gusto customers describe a custom web app in chat and have it built inside Gusto using the Gusto design system.
AI let him ramp back up on 10 years of platform changes (DataDog, Tableau, internal tooling) without embarrassment — “the least judgmental technical mentor.”
He showed the prototype to a small group of engineers/designers in March 2024; they whiteboarded in Denver and formed an 8-person “AI builder” team (including a designer now writing production code).
Co-Founder: an AI agent that runs back-office workflows on a schedule
Co-Founder is not a GUI app builder but a workflow automator: users describe a recurring business process in natural language (via web, SMS, WhatsApp, email, Slack), and the agent executes it proactively on a schedule.
Example: a massage spa exports therapist data from MindBody, calculates complex pay (per-minute rates, CBD oil bonuses, hot-stone bonuses, pooled tips) in spreadsheets, then manually enters results into Gusto. Co-Founder can read the spreadsheet, compute everything, push into Gusto, run payroll, and report the total.
Target: any small business (tech startups, coffee shops, churches, spas) — the same segment Gusto already serves (300k businesses).
The team shipped a customer-demoable version in 5 weeks, incorporated feedback from 10 customers in NYC, and launched a public waitlist 5 weeks later — 10 weeks total with 5–6 builders.
Returning to the trenches reshaped Edward’s view on organizational speed
No roadmap, no Jira boards; the team moves fast because everyone is an “AI-native builder” writing and committing code daily.
Edward now believes every leader should stay hands-on: doing the work reveals nuances (build friction, developer experience, customer pain) that reports and dashboards obscure.
He sees this as completing the picture that pure management left incomplete.
AI is creating a step-function drop in small-business formation friction
Census data shows new-business formation at an all-time high, a trend starting pre-COVID and accelerating with AI.
Building blocks (Stripe, Shopify, Etsy, Gusto) removed capital, labor, and compliance barriers; AI now automates the remaining operational heavy lifting.
Small-business owners are inherently entrepreneurial and quick to adopt AI tools that let them focus on their craft (coffee, pets, massage) instead of back-office drudgery.
Five-year vision: Gusto as the default “yes” to “I want to start a business”
Edward wants any person with a dormant business idea to think of Gusto and feel zero operational, capital, knowledge, or legal blockers — “I can absolutely do this.”
Personal philosophy and habits
Life philosophy in one sentence: “Be positive.”
Positive mindset as a practical tool: it keeps you open to serendipity and willing to walk through doors that appear.
At 40, he committed to health (gym 3–4×/week, strength-focused) to extend the life he enjoys.
Surprising trait if shadowed: relentless positivity, hard work, humility, and a persistent beginner’s mindset.
