From Pain to Platform - Building Merge, with Shensi Ding
Listen to episode
About this episode
Shensi Ding co-founded Merge after watching her own company lose deals it should have won — not on product, but on integrations. In this episode, she traces the path from teaching herself to code at 12 (building emo band fan sites in Boston) to building the unified API powering integrations for thousands of companies. She and Nate dig into why integrations went from "nice to have" to table stakes, why accounting was the hardest category Merge ever tackled, how founder-led sales carried them to $10M, and why she believes openness — not data silos — is the only durable moat in the AI era.
A candid conversation about selling pain, building infrastructure customers can't afford to have break, and the pivotal decision to make AI adoption non-negotiable across the entire company.
In this episode:
- (01:08) Learning to code at 12 — and why being "a loser" building band fan sites led to a CS degree at Columbia
- (02:14) Diverging paths after school: Gil to SF as a top engineer, Shensi to investment banking and growth investing
- (03:09) Chief of staff at Expanse (acquired by Palo Alto Networks) and a front-row seat to enterprise sales
- (04:05) The origin of Merge: losing deals purely on integrations, and why the market was about to get much worse
- (06:48) The first integrations — ticketing systems like Jira and ServiceNow
- (08:28) Why they launched with HRIS and ATS, and the TAM problem with starting in ATS alone
- (10:26) Accounting: the brutally hard third category (and why "hard to build" became a moat)
- (13:18) Why Merge has to hire genuinely sharp reps — and trains them across every category
- (14:05) Fundraising: $75M raised, and how the pitch shifted from "how is this different from Zapier?" to "can they expand beyond HR?"
- (15:19) Building ahead of the AI wave — knowledge bases, file storage, and being ready when the market arrived
- (17:05) Go-to-market evolution: founder-to-founder grinding to $10M, then a real SDR team and enterprise motion
- (18:29) Spotting early customers by "PMing" the prospect — and how Ramp became their first expense-management customer
- (19:54) Moving upmarket after Series B and why Merge segments enterprise by employee count, not customer count
- (22:35) Working through closed HRIS platforms and shifting the whole category toward openness
- (23:50) Open vs. closed in the AI era — and why siloed data is only a temporary moat
- (24:50) The consumer nuance: social platforms blocking bots vs. wearables like Whoop and Oura staying open
- (26:34) Why the categories never end — plus new bets on agentic tooling and LLM routing
- (27:31) How Merge's LLM routing works: smart fallbacks, white-labeling, and enforcing customer data contracts
- (29:26) The pivotal moment: making aggressive AI adoption mandatory across the entire company
- (30:33) The end goal — and who Merge is hiring across SF, NYC, and Berlin
More AI podcast episodes
Browse all →Want to find AI jobs?
Join thousands of AI professionals finding their next opportunity