Check out the conversation on Apple, Spotify and YouTube.
Brought to you by:
Today’s Episode
AI agents are all the rage right now. So I brought in the CEO of the hottest AI agents startup in the world.
Flo Crivello was a Head of Product at Uber. Now, he’s the CEO and founder of Lindy.
If you saw my episode on the 7 AI tools I use in my business, then you know I personally use Lindy.
It’s a fantastic tool for building AI agents. And they’re latest launch, the Lindy agent to build your AI agents, makes it easier than ever to deploy AI agents for your custom workflows.
Whether that’s email writing or processing Gong calls, Lindy can help you.
In today’s episode, we demo how you can use Lindy, we go deep on AI agents, and we cover how PMs can become founders of successful companies like Flo.
If you’re a PM aspiring to be a founder, or anyone who’s interested in building AI agents, this episode is for you.
Your Newsletter Subscriber Bonus:
For subscribers, each episode I also write up a newsletter version of the podcast. Thank you for having me in your inbox.
(By the way, we’ve launched our podcast clips channel as well and we’re going to post most valuable podcast moments on this channel, so don’t miss out: subscribe here.)
What We’re Covering
From PM → Founding Lindy
5 AI Agents You Should Build:
Email Triage Agent
Meeting Notes + Action Agent
Spam Filter for Twitter Mentions
CRM, Fully in Agent
Recruiting Agent
1. From Uber → Founding Lindy
Flo has the classic arc: software engineer → PM → Founder:
Here’s how he made the transitions:
SWE → PM
He became a technical lead and started to work on more on the product side. As he did that, the company naturally wanted to transition him to PM as he took on more that work.
PM → Founder
His advice for PMs who are considering becoming founders:
Just do it.
The PM skillset is a great preparation for becoming a founder. But nothing’s going to prepare you like doing it.
Flo’s first startup, Teamflow, did well, but it took a pivot to AI agents with Lindy for his startup to really take off.
Now, they’ve 5.5x’d in the last 12 months.
2. 5 AI Agents Workflow Demos
You can build an AI agent to do anything. But here are 5 insanely powerful one’s we recommend:
Agent One - Email Triage Agent
We all waste so much time each day just responding to emails.
That’s why everyone should create an agent that reads new emails, classifies them, and acts:
How to Build It:
Trigger:
Use Lindy’s Gmail integration.
Set the trigger as: “When new email received”.
Instruction:
Use plain English. Example prompt:“If an email comes from my manager, mark it urgent and ping me on Slack.
If it’s a newsletter, archive it.
If it’s a client, tag it ‘VIP’ and leave it unread.”Customization Tip: You don’t need to write code or precise prompts — just describe the behavior you want in simple language.
Test & Adjust: Run the agent on recent emails. Check how well it classifies. Tweak logic like:
Add fallback rules for unrecognized senders.
Tell it how to handle forwarded or looped threads.
Your inbox becomes a filtered to-do list. The agent does the sorting so you can start acting.
Agent Two - Meeting Notes + Action Agent
Great ideas die in meetings if nobody records or routes the decisions.
That’s why you should build an agent that joins meetings, writes notes, extracts action items, and delivers them to the right people/tools:
How to Build It:
Trigger: Connect to Google Calendar + Zoom/Meet.
Trigger: “When a meeting begins”.Instructions:
“Record the meeting, summarize key points, and extract action items.
For each action item, route it based on who’s responsible:
If it’s for me → Slack DM
If for team → Post in #action-items channel
If task → Add to Asana”
The best part is Lindy stores all the information and whenever you want, you can retrieve it without remembering anything. Think of querying:
“What was our exact in Q4 and what did Mike say about it?”
With this agent, you can do that 🫠.
Agent Three - Spam Filter for Twitter Mentions
The Twitter API is pretty locked down. It seems impossible to build on it.
But with Lindy’s computer use you can.
For instance, do bots flood your mentions with praise spam? They do mine and Flo’s. You can build an AI agent to block repetitive spam accounts by mimicking human actions:
How to Build It:
Mode:
Use Lindy’s “Virtual Cloud Computer” mode — it runs a full browser in the cloud and simulates clicks/typing.Trigger: Every 3 hours (scheduled run). Open Twitter → Scan mentions.
Logic: Identify spam by patterns:
Repetitive praise (e.g., “Flo is a genius 🙌🔥”)
Emojis + no context
Same phrase from multiple accounts
Action: Simulate clicks to block or mute those users.
No API required, the agent behaves like a human.
Use “confirm before block” for early runs. Train it gently, then go full-auto.
Agent Four - CRM Agent
We all know great conversations happen IRL or on Zoom… and then vanish.
You forget to log them in your CRM.
What if you had an agent that captures new contacts, updates CRM, and reminds you who to follow up with?
How to Build It:
Input Method: You simply message Lindy directly: “Met Sarah, she’s a PM in SF at Stripe.”
Agent Behavior:
Logs the contact (name, title, location)
Tags with relevant metadata
Adds to your CRM database
Proactive Sync: Every week, the agent scans your calendar + inbox. If it finds a contact you forgot to log, it pings you: “You met James on Thursday, add him?”
Smart Recommendations: If you’re traveling, it cross-references with your contacts: “You’re going to Paris next week. Want to meet Sarah (design lead at XYZ)?”
Agent Five - Recruiting Agent
Sourcing takes hours: search, filter, dedupe, email, track…
Instead, build an agent swarm that handles everything from candidate search to first outreach:
How to Build It:
Command Input: “Find 10 software engineers who work at Zapier.”
Agent Chain Reaction: Once that’s entered, Lindy spins up a group of mini-agents that:
Search for candidates on LinkedIn
Check if they were contacted before
Gather their email addresses
Draft personalized outreach messages
Log each contact attempt in a spreadsheet
Outreach Control: The agent respects Gmail limits, so it queues messages if needed. You can also review status:
“Which candidates haven’t responded yet?”
And pause/resume campaigns on command.
If you want to hear how Lindy compares to tools like n8n and ChatGPT…you’ll have to listen or watch the full episode on Apple, Spotify or YouTube.
Key Takeaways
Where to find Flo
If you prefer to only get newsletter emails, unsubscribe from podcast emails here.
If you want to advertise, email productgrowthppp at gmail.
Related
Podcasts:
Newsletters:
P.S. More than 85% of you aren't subscribed yet. If you can subscribe on YouTube, follow on Apple & Spotify, my commitment to you is that we'll continue making this content better.
Share this post