At a glance
WHAT IT’S REALLY ABOUT
Lindy’s CEO on building agents, pivots, and execution standards fast
- Lindy quintupled revenue in six months by focusing on durable user needs—simple, fast, affordable, effective software—rather than obsessing over weekly AI model updates.
- Crivello argues AI products should be built for the next generation of models, meaning early versions may “barely work” while still validating a bigger long-term vision.
- A major growth unlock came from listening to unexpected customer demand (e.g., HIPAA needs) that revealed high-value vertical use cases like clinical documentation and SOAP note creation.
- He recounts a painful pivot from his prior remote-work startup (Team Flow) after COVID demand faded, including layoffs and a full strategic reset toward LLM-driven action-taking agents.
- His leadership philosophy emphasizes recruiting as destiny, explicit “standards of excellence,” and active micromanagement as a CEO’s obligation to ensure inputs are correct and execution matches the vision.
IDEAS WORTH REMEMBERING
5 ideasOptimize for what doesn’t change: user preferences.
Crivello leans on Bezos’s idea that even in fast-changing tech, users still want products that are simple, fast, effective, and affordable; growth comes from nailing these invariants.
In AI, a “working” first version may mean you aimed too small.
He argues you should design for upcoming model capabilities, so today’s models only barely support the product’s true vision—because model improvements can convert “almost possible” into “reliable.”
Treat compliance/constraints as signals of valuable markets.
Repeated HIPAA requests weren’t a nuisance—they revealed serious healthcare demand (e.g., automating clinical notes), guiding Lindy toward higher-stakes use cases.
Don’t build the mine before you find the gold.
He separates “prospecting” (searching for product-market fit) from “building the mine” (scaling headcount and burn); scaling too early makes pivots expensive and can force painful teardown.
Pivoting hard is a strategy, not a failure mode.
Team Flow’s COVID-driven market evaporated; seeing well-funded competitors also stall convinced him demand was gone, prompting layoffs and a full shift to the emerging LLM opportunity.
WORDS WORTH SAVING
5 quotesIn AI, I would even go as far as to say, like, if your first version works, you're not thinking big enough.
— Flo Crivello
You should build for the future generations of the model so much, and the models are improving so rapidly that the current generation of models should barely support your vision. Just barely.
— Flo Crivello
Don't build a mine until you've, you've found gold.
— Flo Crivello
Absolutely, you should micromanage. The reason why you are in a position of, of, of authority is because someone, with capital most likely, trusted you and your taste. So you, you are entitled, you are in fact obligated to micromanage.
— Flo Crivello
AI agents are basically, in the short term, going to turn us all into managers. Instead of doing individual contributor work, you're going to manage a swarm of AI agents doing your work.
— Flo Crivello
High quality AI-generated summary created from speaker-labeled transcript.
