By playing, you agree to uListen's Terms and Privacy Policy and YouTube's Terms of Service.
Summaries, chapters, transcripts, and FAQs on this page are made by uListen, not YouTube.
At a glance
WHAT IT’S REALLY ABOUT
Five AI agents PMs should deploy now to scale work
- Flo Crivello argues many knowledge-work tasks are now 100× cheaper/faster with agents, framing the shift as everyone becoming an “orchestrator” rather than a doer.
- The episode demos high-ROI agent patterns—meeting recorder with automated Slack routing, Gmail email triage, CRM/contact manager, support automation, and recruiting/outreach using Agent Swarms.
- Lindy’s differentiators are extreme customization (flow editor + natural-language agent builder), broad integrations (~6,000), and “computer use” to automate sites without APIs (e.g., Twitter/LinkedIn).
- Risk management is handled via human-in-the-loop approvals, permission scoping, and emerging cost guardrails; model choice is positioned as a speed-vs-intelligence tradeoff.
- The conversation expands into founder/operator lessons: how Lindy emerged from customer requests and rapid iteration, how to run AI transformation inside companies, and cautions on hiring (job-hopper story).
IDEAS WORTH REMEMBERING
5 ideasStart with agents that eliminate “coordination tax” first.
The meeting-recording agent that auto-summarizes and posts to the right Slack channel targets a universal PM pain: alignment and follow-through across many stakeholders.
Use natural-language “agent builder” for fast prototyping, then harden with flow controls.
Flo shows building Gmail triage from a single English prompt; the power comes when you later add conditions, loops, and routing logic in the flow editor.
Treat instructions as living policy docs, not static prompts.
A practical pattern is storing triage rules or review standards in a Google Doc that the agent consults each run, allowing non-builders to update behavior without editing the agent.
Mitigate hallucinations operationally, not philosophically.
Instead of debating LLM reliability, add one-click human confirmations on sensitive steps (e.g., sending messages, executing actions) and iterate as the agent learns from corrections.
Model selection should map to task economics: speed, IQ, and dollar value.
Use fast/cheap models for latency-sensitive tasks (e.g., phone calls via Gemini Flash) and premium reasoning models (e.g., o3 / Gemini Opus) when a task’s outcome is worth spending dollars in tokens.
WORDS WORTH SAVING
5 quotesI think people should increasingly, with AI agents and with the rise of AI in general, they should increasingly think themselves as orchestrators and not as doers.
— Flo Crivello
I just think with AI agents, it's like everybody just got a promotion to manager.
— Flo Crivello
It's like AWS for labor, right? It's like you never have to worry about scaling again.
— Flo Crivello
Don't do a grand plan. Like, the only thing I know about your grand plan is that it's wrong.
— Flo Crivello
People way overthink this shit. Like, all the networking events and the dinners and the conferences and, like, frankly, the investor coffees, like, you can skip all of it. You should skip all of it.
— Flo Crivello
SPEAKERS
Aakash Gupta
hostProduct leader and creator of the Aakash Gupta channel/podcast, interviewing PMs and tech founders about product and AI.
Flo Crivello
guestCEO and founder of Lindy, an AI agent platform focused on automations and integrations.
High quality AI-generated summary created from speaker-labeled transcript.
