253 - I Asked AI to Run My Business for a Week. Here's What Happened with Andrew Pignanelli

18 Nov 2025 · 41 min · 23 chapters

Ask about this episode

Ask anything about it. ChatGPT or Claude reads this page and answers with the times it was said.

Connect VO and ask about every podcast you hear, including the moments you saved. Add to ChatGPT · Add to Claude

In short

Andrew Pignanelli (General Intelligence Company of New York) explains Co-Founder, an AI “chief of staff” that helps entrepreneurs run “one person billion dollar companies” by automating the “million paper cuts” (scheduling, emails, research, recruiting, monitoring). He demonstrates human-in-the-loop approvals, natural-language automations, integrations via Model Context Protocol (MCP), and a persistent memory system.

Key claims

Co-Founder queues actions to avoid mistakes (he previously had AI send ~20 calendar invites to people in spam, requiring apology emails and blocking). It can draft/send emails and calendar invites with approval, monitor up to ~100 URLs for changes, and create automations without Zapier.

Notable examples

fundraising—Co-Founder reads investor intro email chains and adds “hold” events to his calendar; competitor monitoring by tracking websites/blogs/press releases; exporting calendar/to-do items into Google Sheets.

Guests

Andrew Pignanelli, co-founder of General Intelligence Company of New York; podcast host Nate (NICconomics).

Written by AI. May contain mistakes. Listen to the episode to check what was said.

Chapters

Tap a time to open that second in VO

Automation in Daily Tasks

0:00 to 0:40

Learn about the benefits and challenges of automating daily tasks with AI.

“I can say, do this every day and send BABM on Slack.”

The Concept of One-Person Billion Dollar Companies

1:16 to 2:18

Discover the idea behind one-person billion dollar companies and its implications for entrepreneurs.

“problem which is how do i as a founder extend myself without having to do it the traditional way, which is hiring a team, onboarding a team, and growing that way.”

Data Management and AI Integration

2:18 to 3:27

Understand the challenges of managing data and how AI can help synthesize information.

“And so that's really what we're working towards.”

Understanding the Model Context Protocol (MCP)

3:27 to 4:52

Learn about MCP and its role in connecting AI agents to various tools.

“We actually built these integrations ourselves because if you use someone else's standard, you tend to lose the ability to control the experience.”

Addressing the Million Paper Cuts Problem

4:52 to 5:45

Explore the small yet significant daily tasks that hinder entrepreneurs and how to alleviate them.

“where entrepreneurs are trying to fix, right?”

The Role of Co-Founder as an AI Assistant

5:45 to 6:40

Discover how Co-Founder functions as an AI assistant to help entrepreneurs manage their tasks.

“There's a lot of effectively either context building that you have to do as an entrepreneur or handing off context to another individual so they can go and execute a task.”

Use Cases of Co-Founder

6:40 to 7:58

Learn about the various use cases for Co-Founder, including scheduling and research.

“It's like, it's just like a thing doer, what we call it.”

Human Oversight in AI Scheduling

7:58 to 9:35

Understand the importance of human approval in AI-driven scheduling and task management.

“Are you, do you like, hey, co-founder, schedule a meeting with Barry for next week, find any time, and then it goes and does it?”

Interaction with AI Agents

9:35 to 11:19

Explore how to effectively interact with AI agents for optimal results.

“I know that kind of like sounds weird to say, but if you had unlimited money and could hire an assistant to handle your schedule, what would you need to tell them in order for them to work well?”

Integrations and Workflow Optimization

11:19 to 12:12

Learn about the integration capabilities of Co-Founder and how it optimizes workflows.

“And this to-do list will like, it can be really long or it can be really short.”
Show all 23 chapters

Memory System in AI Applications

12:12 to 14:01

Discover the significance of the memory system in AI tools for personalized experiences.

“Right now this is using Claude 4.5 Haiku.”

Exploring CoFounder: Automations and Memory

14:01 to 18:26

Learn about CoFounder's unique memory system and automation features for businesses.

“And I don't think you can see in this tab, but I connect to my.”

The Podcast Social Contract

18:26 to 19:56

Understand the host's request for listener support and engagement.

“I think I made the joke to you when I saw your video that I was like, yeah, I want to buy my website someday.”

Onboarding and User Experience

19:56 to 22:46

Discover how CoFounder simplifies onboarding and adapts to user preferences.

“people and hear really cool insights we both get a ton of value but i just want to help you keep your word.”

User Demographics: Who's Using CoFounder?

22:46 to 24:23

Explore the diverse user base of CoFounder, from tech founders to SMB owners.

“I think that's our biggest sort of segment.”

Monitoring and Competitive Analysis

24:23 to 28:00

Learn how to use CoFounder for monitoring competitors and managing tasks.

“I have a draft and send emails, and I have a draft and send calendar invites so that I don't have to think about it.”

Building Scalable AI Solutions

28:00 to 29:10

Discussing the scalability of AI integrations and the focus on depth over breadth.

“But only if you have the API integration, right?”

The Future of Autonomous Companies

29:10 to 30:58

Exploring the vision for creating autonomous companies and the role of coordination.

“I mean, yeah, it's a highly competitive space, but our goal is not to make a good AI assistant because then it would be competing directly against Anthropic and ChatGBT, right?”

Solving Organizational Problems

30:58 to 32:40

Identifying the most solvable problems within organizations and how to tackle them.

“Because that's like, you have to start somewhere and that's one of the most solvable things right now.”

Educating Users and Market Creation

32:40 to 34:12

Discussing how to educate users about tech products and improve new user experiences.

“I mean, new user experience is really important for this.”

Onboarding and User Preferences

34:12 to 36:28

Delving into the onboarding process and how user preferences are integrated.

“Are you going through that process and asking them a bunch of questions or are you just figuring it out as you go?”

Concerns About AI's Future Impact

36:28 to 38:15

Addressing concerns regarding job displacement due to AI advancements and future implications.

“But we want to make it much bigger than it is and much like really know you.”

Finding and Testing Co-Founder

38:15 to 39:23

Discussing how listeners can find and test the Co-Founder platform.

“They say, oh, well, yeah, I mean, you know, we're going to lose a bunch of jobs, and then AI is going to automate everything, and it's going to be great.”
Hear the part that matters, and keep it.Open this episode in VO. Double tap your headphones to save a moment as you listen.
Get VO free

Transcript

Automatic transcript. May contain errors.

0:00I can say, do this every day and send BABM on Slack. And co-founder is going to go ahead and make that an automation. Wow. That's all I have to say. And it sets up everything for you. I actually learned that one the hard way. It sent like 20 different calendar invites to people in my spam folder. I actually learned that one the hard way. It sent like 20 different calendar invites to people in my spam folder. And I literally had to send apology emails to everyone and then block the people that needed to block. One of the coolest things that I use it for while fundraising, the co-founder, read all of these and put them in my calendar as holds.

0:32And it did that and it saved me a ton of time.

0:39All right, Andrew Pignanelli, welcome to NICconomics. I am really excited to have you. The reason I'm excited to have you is because you founded a company called the General Intelligence Company of New York. And your first product is this product called Co-Founder. I saw this video that was released and there was a tagline that you had in there that really caught my ear, which was we want to facilitate one person billion dollar companies which sounds outlandish but you guys have raised a good amount of money you launched in on september 8th of this year today is october 27th and you have already 6 000 customers you've got in the mid hundreds of daily active users the company is growing really really fast and i think you're tackling this really cool unique problem which is how do i as a founder extend myself without having to do it the traditional way, which is hiring a team, onboarding a team, and growing that way.

1:28And I think for people who are listening who are entrepreneurs, that's what they want to figure out is like, how do I grow my business without necessarily having to scale the people side of things? So my first question to you is, what does it mean to you to help companies become one person billion dollar companies? Yeah, well, thanks for having me on. This concept is really important for us. We like to think about entrepreneurs all the time and the challenges they face every day. And going after this one person billion dollar company is like a long term goal on just solving those problems for them, you know, in their day to day lives.

2:00We wanted to make it really, really easy for people to start companies and to run companies. And that's what we started was with that goal of making it so fast, so easy that you could have a business idea at night on Friday and then by Monday morning have an actual real legitimate business. And so that's really what we're working towards. We've seen it resonate with people too on like the early signups of these like startup founders and entrepreneurs and things that are just trying to use it to give themselves more time in the day. Yeah, it's really interesting. I've been thinking about this a lot myself because there is so much data out there that we just individually create, right?

2:38I've got my Gmail. I've got Google Docs. I've got text messages. We've got Slack. So you've got your CRM where you track leads, where you document things, et cetera. and like i'm constantly thinking how do i create an end-of-the-day context document so that tomorrow i can just be off and running and have ai help me with a number of these processes and as i've been thinking about it you know the hard part is getting all those to tie in but also making sure there's enough capacity for ai to look through it synthesize it and then you know give whatever analysis and context it needs how do you do it like how are you connecting to all these disparate data points.

3:18Is it magic? Don't lie to me. It's magic, isn't it? It's definitely not magic. And I sometimes wish it was. We plug into all these different systems and we off into them. We actually built these integrations ourselves because if you use someone else's standard, you tend to lose the ability to control the experience. So I am a novice. Are you technical? Are you like an engineering background? Okay. So I'm not. I am what the kids call a vibe coder. Cloud code is my tool of choice. But you said a couple things in there. So MCP, I just want to define that for people. So that's the model context protocol.

3:58Can you explain exactly what that is? Yeah, so MCP came out in, I think it was February, where this is the Anthropics take on how do we get a bunch of agents connected to a bunch of tools. And it's this new standard where you take like an API, which you can plug into some service and you give it a series of instructions attached to the tools. And so you can essentially just give this thing to an agent and it can read the instructions and use the server to run the tool. So if we're talking about emails, there is a Gmail MCP where it says, oh, if you want to send an email, use the send email tool.

4:37Now, this is different than an API, which is essentially just a list of all the tools that you can use. And so I think that's a big difference. Let's back up actually even just a little more because I want to talk about each one of these things. What problem is this solving that you created where entrepreneurs are trying to fix, right? Like what problem is it actually attacking? Yeah, so we call this the million paper cuts problem, which is less of like one single thing and more of like there's a million things you do every day that are small and they take a bit of time and they take a bit of brand capacity.

5:14So think of it like you have a couple people trying to schedule a meeting and you need to get them together. You need to send a calendar invite. You need to send someone an email with info on your company. You need to get a document from one email, put it in another and send it across to someone. You need to forecast a budget, but it's relatively simple. Or you just need to research something. All of this stuff is relatively small on the surface, but if you're an entrepreneur and you're a business owner you have to do a million of these things every day and so our solution is with co-founder takes a lot of that off her plate and it's able to just do it for you we're pitching it as this sort of chief of staff or like self-completing to-do list that you have every day yeah because for a lot of entrepreneurs especially as you're scaling it's like oh it'd be nice to have somebody to schedule this meeting or to give me a prep document for this person I'm about to meet with or to be up to speed so that I can understand where I even pull sales data from a CRM.

6:14There's a lot of effectively either context building that you have to do as an entrepreneur or handing off context to another individual so they can go and execute a task. So like you're saying, it's like a million of these little cuts kind of throughout the day. This would allow across all of those data sources or I don't even know modes of communication, right? Slack, email, text, et cetera, for co-founder to help you as a chief of staff, right? It's like to kind of be an extension of you. Yeah, exactly. It's like, it's just like a thing doer, what we call it. And it pulls, and the reason it's good at that is because it pulls in the context from how you actually work.

6:52So like, instead of, instead of just doing stuff, it does stuff while knowing you. So instead of just scheduling a meeting, it knows, oh, you're not going to be there if you have to go across town, you're not going to be there in time. So you can't schedule at that restaurant. Or instead of just writing some random reply to an email, it can write a reply knowing the context you need to write that email, which is a really big thing that you see as an executive, is just having to have this knowledge to deal with these things. What are the biggest use cases you're seeing people use CoFounder for right now?

7:23We are seeing a lot of scheduling and emails. A big one is also research. One of the things GoFounder can do is it can monitor a set of like websites and give you updates on those websites. So if you give like 20 websites, it can give you an update everything every time something changes. So we're seeing people use this for competitor monitoring, for content creation, to let's say monitor a bunch of blogs and pull in content from them and then write your own blog piece. We're seeing people use it for recruiting. They use it for recruiting like automatically by setting up sequences. We see people using it for keeping themselves up to date every day.

7:57I think I'm the most curious about what's it doing in the scheduling? Is it making decisions? Are you, do you like, hey, co-founder, schedule a meeting with Barry for next week, find any time, and then it goes and does it? Or like, how does that interaction work? Depends on what you ask, right? So if you say find any time, it would use the, look at your calendar and find free slots on the calendar. if you say next week it's going to make a decision about what that means yeah maybe if you know you say next week it might be like monday and then it will look at the free slots in your calendar and then it will figure out who the person you were talking about is by searching them in your email and then it will actually send a calendar invite now we've built all these guardrails so that it's not just sending these things willy-nilly you know i actually learned that one the hard way by not having was very rails and plugging my email and calendar.

8:49And it sent like 20 different calendar invites to people in my spam folder, actually. Yeah, I literally had to send apology emails to everyone, and then block the people that needed to be blocked. The way that works now is like, it queues it up. And it's like, hey, this is the calendar invite this next week, you want to send it? Yes. And so the biggest This change from your workflow goes from like having to go and do that to just approving it. And now you're seeing all these things and you can just approve them and you can approve the emails that are written and then you can just hit all of these things at the same time.

9:23Can I see what that looks like? Yeah. And while you're pulling that up, I'm curious on your thoughts here. I've been thinking a lot about AI and agents in particular as if you want them to work well, you've got to treat them as humans. I know that kind of like sounds weird to say, but if you had unlimited money and could hire an assistant to handle your schedule, what would you need to tell them in order for them to work well? You'd have to be very specific. You don't just say, hey, schedule a meeting with Barry. It's like, oh, schedule a meeting with Barry at this place next week during these times.

9:53It's the same thing. It's just you're happening to be using AI, which feels a little bit weird. And so with what you're building, do you feel that philosophy? Am I oversimplifying it? Or is that like kind of a good way to think about how to interact with these agents? It is changing the way that you're thinking about this stuff now. Like we lean into stuff and companies where they're also down to lean into AI. if you're able to, you can lean into it really hard. And it does change the way that you work, right? So like instead of opening your email the first thing in the morning, maybe you'll open CoFounder and tell it to read your email.

10:29Or like instead of starting the chat session, you might make an automation that starts this chat session. What did you type in here? I'm looking at this right now. So CoFounder has a chat window that you go to and you're interacting with. So I asked it, what's my calendar look like on Thursday? And so this is to give you an idea of like how it works, right? So it used its memory system to see past calendar events, but future calendar events are not in the memory yet. And so it did a search and it found what I'm doing on Thursday, which this is all accurate. Like this is the things there. And then it's giving me a response, right?

11:04So morning, I do this. During the day, I'm speaking at that time. And then in the evening, I have cocktail hour and then a party, right? So that's how it kind of uses the tools. And it's got this to-do list. And this to-do list will like, it can be really long or it can be really short. So I can say... Oh, is that native in CoFounder? Yeah. So this is one of the ways we trip the agent into running for a long time. So I can say something like, it's true. And I can say something like, okay, put this in a Google sheet with each event in a row. and so now this is going to use the spreadsheet tool to go and put this in a google sheet and you'll see the to-do list like light up and uh it will be able to show you that i'm assuming you've given it access to your google drive and your gmail and like all these things have already been ported in yeah i can show you that in a minute let's just uh let's watch co-founder run on the sheet for a sec and then i'll go and show you the integrations yeah i mean it plugs into like everything you use in your workflow.

12:11What are you running this on top of? What model is this using? Right now this is using Claude 4.5 Haiku. We can use 4.5 Sonnet or we can use ChatGPT type models. So you can see there it made this to-do list where it has been to add them. And it's going to cross items off on this to-do list. As it does it? Does it? Yeah. And the models have been a little bit slow today. So it might be a little slower than normal, but it will go in like, you know, do this across different integrations and things. And you can make these to do this really long. Oh, yeah, there we go. So this shows you the human in the loop aspect a little bit.

12:51It can feed a Google Sheet and if I click show more, it should be able to run that. But it needs you to actually say yes, run. Correct. Yeah. So let me try this. Is it to the point yet where it's, it doesn't need you or like, why does it need the human in the loop. We've set this up so that you can go and do it like that because we don't want the agent just doing stuff for you. Right. So let's say this is a very long, complicated spreadsheet that you made. You don't want it just writing to that spreadsheet without your approval. Now, what you can do, though, is you can set this to auto-approve, which will just keep running.

13:30Just do it. And it will basically say yes to everything. And so if you're doing something it's a really long time you're going to want it to do that so if i click this i might have to log in with my other google account but not yet see there it is and really quickly i want to show you what connecting stuff looks like if you go to the integrations page we've got all these integrations connected and so like google drive intercom loops notion and i can show you how this works if i if i click disconnect here it will just disconnect it it just takes like three seconds to connect to Google Sheets and you just hit continue.

14:03And I don't think you can see in this tab, but I connect to my. No, I can't. I'm just looking at the account connection in progress. Yep. And then I just click that and then now it should be good. It was just like two buttons on the other side. I clicked the Google login screen. So, you know, we plug into all this different stuff. How is this different than like if I'm in Claude and I've connected my Google Drive and my Gmail and I want to query it. Like, is this different? Or like, how is it functionally different? Yeah, so you can do pretty like simple operations in a very similar way to Quad.

14:41But one of the biggest differences is the memory system that we've added. So this like knows things about me, as well as, so I can say something like what's going on in engineering and it will pull stuff from its native memory. And it will tell me what, you know, the things that have happened from my memory and in my company. And I think that's the biggest difference is it knows me and it can do this stuff. Yeah, because if I was in Claude, I'd either need to have, if it's a project, I'd either have to add new context documents to the project or have some way of getting it up to speed on what's going on.

15:14It's not persistent. And there's one thing that I haven't shown you yet, which is the way you make automations. Now, CoFounder is the easiest way to make automations in the world. so I can say okay do this every day and send be a bm on slack and co-founder is going to go ahead and make that an automation and this and that's it that's all I have to say and it sets up the web hooks and everything and it sets up the the context and the tools and everything for you and so if I click this it will go to the flow and this is like this is the automation now it runs every time. So if I want to like try it, I can click test and it's going to test it.

15:55It's going to run it right now for me. And like, you don't have to go through Zapier. Like you don't have to go through Zapier. It runs the automations for you. Exactly. Runs it for me right now. So it's, it's testing this right now. It's like pulling up the flow. And so now it's initiating, checking GitHub and sending a Slack message. That's really cool. Yeah. I mean, it's all natural language based. Like if I want to change the plan, if I go to the flow gear and I want to change the brand, I can just click it and I can edit it and change it right there. If I want to change the integrations, they happen automatically.

16:28I can't even choose these because the agent just knows what I want. So like if I was a business owner, I'd be like, you know what? I want marketing report end of day, every day. I want engineering to report every day, end of day. I want finance once a week. And I could come in here and say, give me the marketing end of day report. This is how I want it to look every single day sent to my email. And then it would just create that automation. Exactly. Yep. And you can have it asked you for permission to do whatever you want, or you can have it run automatically. So, you know, if I said to run automatically, it won't ask me permission to send it in.

16:59It will just send me the report. And like, it's all with natural language. So you can just tell the agent you don't like how something is, and the agent will change it for you. And so this is like a big part of our thesis for the company, which is like agents all the way down. Everything is an agent. Yeah. And if you do that, then you can do everything with natural language and we can like things just kind of work together and it can operate together. And that's how - Can it respond to emails for me? Like let's say that I wanted to, would I create an agent to do that? Or if I was like, hey, I just want you to go through all the emails that I get every day with questions that you know how to answer and answer those questions and then star all the things that I need to look at.

17:38Because as an entrepreneur, Or there are things that I don't need to look at, but they come to my email because whatever. Payroll processed. Here's the update that it actually processed. Or there's a customer question, but it was handled by my marketing team. I don't necessarily need to see that. But how would I set it up as an assistant to basically handle my emails? Yeah. So it's just as easy. Like this one, for example, is someone wanted to take emails from investor updates and forward them to whoever this is. And co-founder just has access to your email, right? I don't get investor updates because I'm not a VC, but it was just as easy, right?

18:15You can read my emails. And so I'm going to try to see if I can do this without leaking all the emails in my inbox on this podcast. But we'll see. What was your email address? NIK. Email from NIK at cofounders.com. I think I made the joke to you when I saw your video that I was like, yeah, I want to buy my website someday. So it's going to do when it's loading here, it will do a memory search. And it was not able to find you in the memory search, but it will be able to find you in the long run agent search. And so now it's just going to search my Gmail and find it. So it does two searches. Is it like storing a, I don't know, static context document every day of kind of updating your history?

19:02And then if it doesn't find it there, it runs an actual agent search? Correct. Yeah. So that works more for like concepts and like people you know. Does it currently respond to emails for you? Yeah, if you want to. So I can say, let's see, send an email to Nick saying I'm excited to be on his podcast. And so when it will pull up this thing, it's not going to send it immediately. This is one of those things where I have to approve it. But yeah, I mean, we've got a little email out in our tier. And one of the things is it makes a to-do list every time. So very simple tasks take a minute. We're looking into making that faster.

19:41but hey i don't know if you remember this but when we started this podcast we entered into a social contract i would spend time energy and money producing this podcast interviewing these individuals and giving you insights into how to build buy start grow your business and you would like subscribe and leave me five star review now out of that we both get to talk to really cool people and hear really cool insights we both get a ton of value but i just want to help you keep your word. So would you do me a favor? Will you go leave a five-star review for me on Apple or Spotify? It would really help.

20:14And if you want, even share this with a friend. You could kind of just operate within co-founder and never leave that window. Like you don't need to go to Slack. You don't need to go to your email. I mean, I don't know, maybe I'm wrong. It just feels like you could get any updates here and not have to like context or not or windows switch, I guess. Is that, did you design it that way? We did, yeah. And we're actually launching this feature we call like to-do list or home where instead of mini to-do list, it has an email. I don't know. I guess it's like running issues with the Gmail tool. But you could do everything from within CoSounder and basically make this to-do list and then have it complete itself.

20:55Can you tell it, hey, these types of emails, you can always respond if you know X, Y, Z. like can you effectively automate the responses does that make sense because like i've handed off things to my executive assistant she just answers emails i don't i don't have to like check anything but there are certain things she can't she stars them for me and then i look at them later yeah so here's an example of like one that looks at at my email for spam and it's really inefficient but it does work where it will say oh you know if it if it's from here march the email as spam and like here's an example of some email from someone and it scanned all these emails and it said oh you know these are none of these are spam so it didn't actually need to mark them as spam but i could also have it respond to these emails so i like respond to the last one and then it will just like pull it up and write a response if i'm an entrepreneur if i wanted to use this how hard is the onboarding?

21:55Because if I just connect my Google Drive, there's a million things in there. Do I have to tell the AI what's what or does it go through and sort of comb it and self-identify what's important and what's not? It goes in and it finds all that stuff out. It structures the memory for you on its own. And actually people love our onboarding. Truth is sweet. I've become known as like one of the nicest onboardings in the industry. And it figures out your style, it figures out your calendar. And it just, it reads this for you. And we actually designed the onboarding so that you only have to type something in twice, which is your name.

22:30And it's actually already filled in for you. It's your name and your company's website. And both of these things are filled in for you if you use your company email. And so. What industries are you seeing people use this in right now? Is it just a lot of tech founders or are you getting people, you know, SMB owners all the way to corporate individuals? Like who's actually using this tool? We've seen a lot of tech founders. I think that's our biggest sort of segment. We also have people from like mid-sized startups, like thousand person companies, some of the companies you might have heard of. I don't know if I should name drop it or not right now.

23:05That's okay. Yeah, it's fine. Big companies. The biggest companies, many people. They're huge. They're huge. We do have a number of investors that use it too because they like the monitoring system a lot. Also, I've talked to a bunch of investors while fundraising, and I think it stuck with them. And then we've had some requests for like services businesses as well, like people doing agencies or like accountants and things like that that are just using it to save themselves time. What's the monitoring function? You can set it, you can give it like a list of URLs and it will monitor those websites for changes.

23:44And so what DCs will use it for is companies they're interested in or competitors of deals they're looking at. So you can give it like 20 URLs. These are all the competitors. And then say, oh, monitor these for any blog posts or press releases so you know when they get funded. How many could you monitor? Is it just like, what's the limit? I actually don't know. I think it's like 100 or something. We may have just set that because if you go higher, it gets really expensive because we're basically putting a little mini, very cheap mini agent on these websites. It's a check. Yeah, I mean, I think 100, I could look into more.

24:23Are you using this? Of course. I use it for managing my schedule. I use the monitoring for competitors. I know all my competitors. I know everything about them. What's your favorite thing? So what do you use it the most for? I like the scheduling. I have a draft and send emails, and I have a draft and send calendar invites so that I don't have to think about it. Okay, so one of the coolest things that I use it for while fundraising, and I think this is the most useful thing that I've used it for ever, was when you're fundraising, you get introduced to like 40 different people, you have to meet them all as quickly as possible.

24:53And so you immediately turn these conversations into scheduling. And the problem is someone has an assistant or someone needs to get another person on. And so you have all of these email chains where there's a date and time decided, but they're not in your calendar. And it is impossible to tell when you're free and when you're not. So I told CoFounder, read all of these and put them in my calendar as holds so I can reason about when I'm free. It did that and it saved me a ton of time. I mean, I would have had to have an actual assistant to be able to do something like that. I've had this pendant, it's called the Limitless AI, but there's others like Plod and I can't remember.

Read the full transcript

25:32Anyways, where it records a lot of conversations. Could there be a function within CoFounder where I'm like, hey, go take all the conversations I had today and just update my calendars on things that I said I would do. We literally have a limitless integration. Do you really? Yes. It's because Legion on our team has limitless and he likes to use it. So he'll like plug stuff in. So an example, we actually have always on microfilms in our office that plug into co-founder. It's a little crazy. It's creepy. I get it. Yeah. But then every day it just like assigns context and like tasks to people or just like tell someone if you're working on something like, hey, like this, here's something you were talking about with this guy, like about this feature or something.

26:17What's the automation with Limitless? Because I've found Limitless doesn't have a great ecosystem. So the fact you guys have an API integration with them, I'm like giddy as a school child right now. It can list all the life logs and it can get all the life logs. So it's a read-only thing. It can't like push to it. But if you have, I think life logs - But it can pull from it. Yeah, it can pull all the things from it. It can search for specific ones. It could do things like that. And so you can, if you pull stuff from Limitless, you can say, oh, you know, every day pull all my stuff from Limitless and then summarize it for me and give me a transcript or something.

26:53I'd use Limitless when I go to conferences and I'd meet people and I'd have it on and I'd be like, oh, Brian, nice to meet you. And then at the end of the day, I'd be like, who did I meet with? And it would list them all. And what do they do? And so I'd have like this CRM, if you will, of people that I've met with, companies that they're with and like things that we spoke about when I want to follow up with them etc but I had to do it all manually because I and there probably is better automations at this point this was like six months ago so someone's probably screaming at the screen right now I'm like you moron you can totally do this but for me it would just be nice to have something that at the end of the day goes and pulls all those logs and then creates here's to-dos here's a schedule here's who you should follow up with updates a CRM like like does it could it do those things could it could I tell it to do that and at the end of the day it's populating my crm and email and setting calendars etc yeah it would be really easy well really in fact in fact you could take you could tell it you could just tell it exactly what you told me co-founder could make that happen oh my gosh we like to say it's the simplest automation experience in the world and it's just it's just because it's all natural language right and if you're describing like if you can effectively describe what you want done, you can get it to happen.

28:02Wow. We're pushing it as far as we can. It's really interesting. But only if you have the API integration, right? Like if I'm trying to get it to tie to something you don't have the API for, then we got to figure something else out. Yeah. And like we're working towards how to make that scalable because there's a trade-off between how wide you go and how tall you go, right? So we could have like a million integrations with like no depth and that's one way to do it. or we could have like 30 with a lot of depth and that's the other way to do it. And we're picking the latter option because we want to build this system.

28:37Our goal is to build this system of a company, right? And not to build into everyone's system of a company. And I think there's a big difference there. But how worried are you about competition? Because like I was talking about Comet or Atlas, I created a workflow automation in Zapier last night using Atlas. And I was just like, hey, go to Zapier and create this for me. And it did. It wasn't like anything incredible, but it created the automation for me. Are you worried about competition or like how do you kind of view a changing landscape and trying to build a moat at the same time? It just like that would scare me.

29:11I'm not super worried. I mean, yeah, it's a highly competitive space, but our goal is not to make a good AI assistant because then it would be competing directly against Anthropic and ChatGBT, right? And they have way more money than we do. Our goal is to make autonomous companies. and that is really where we're going to unlock the value and it's really you know our value is going to come from enabling somebody to start a business and run a business immediately and in a way that is legitimate and not just slop you know allowing them to do that and co-founder is just our first step and so you know with our vision competition doesn't worry me very much how do you enable these companies so co-founder is essentially our first step and it's like this chief of staff assistant thing, right?

29:58But we're really focused on solving the coordination problem between departments. So when we talk about like how we're actually going to get to the one person billion dollar company, it has more to do with being able to coordinate things together as opposed to just making this assistant. And what that actually means is we want to plug into all these vertical agents that people are building. We want to wrap it up and make it really easy for an entrepreneur to use. So for example, you've got coding agents, you've got product analytics agents, and you've got customer support agents. If you can coordinate those together, they're all operating autonomously.

30:33You can work towards a high level goal of like increase my customer retention, but each of them individually is not going to do that. And so we're working on coordinating them together for the product department and the sales department and the marketing department and the ops department and building these agents that can work on goals that are much higher level and then eventually enable these people to own these companies and run these companies. So the first vertical that you tackled was chief of staff assistance. Basically. Yeah. Because that's like, you have to start somewhere and that's one of the most solvable things right now.

31:08That's what we saw at least. But what do you think is the next easiest solvable, most solvable problem within an organization? I think the product department, but I don't think entirely. Like I think, you know, just enabling a higher level above value coding. So take, for example, like something like Lovable, where you can make an app full stack, right? We want to take that and extend it to, okay, you have your app. Now you can extend it to more business. You can plug in analytics and you can have an agent operate on the analytics and you can plug in a customer support agent. And now you've got this product department as opposed to just a product.

31:46And so that's what we're going to be doing next. how do you think about educating the market? Because for example, I worked in hospice. Do you know what hospice is? Yeah. Hospice is end of life care. And many people think hospice is, oh, now I'm dead. But the truth is many people live sometimes more than a year on hospice. Hospice is really just about managing symptoms. You have a terminal diagnosis and a prognosis of six months or less, and we help you die comfortably. I mean, that's kind of the easiest way to say it. But there's a lot of education that goes into that because many people just think, oh, I'm on hospice.

32:18I'm dying tomorrow. You're just going to give me morphine. So we have to go out, talk a lot. That's the first time you probably had your product compared to hospice. But your product is similar in the sense that there's an education component on how do you use these things or what do you need to have in order to even unlock the potential there? How do you think about educating or creating a market that really isn't educated? Yeah. I mean, new user experience is really important for this. So I think pushing content out. Getting people towards the product is one thing, but then you have to push content out within the product.

32:50So when someone opens it up and they see this empty box, it's like, what do I do? That's the biggest problem you have to solve when you're working on an app like this. And we're solving it by using this memory system that we've built up to see what you want to automate before you know it. And this is really, really key. And one of the things that we're putting a lot of work to right now is like, okay, I know that whenever you get an invoice from this email, you forward it to your accounts payable department, or I know every Monday you send an email to this person with this information or something like that.

33:25Pulling those things in and like making these automations for people is really, really big. And so this is like a one way of educating is like how this is how you do it. But then there's the content on like, okay, well, how do you build these automations into your life that we're pushing out on like content of like how do we use it how do i use it what is what does this person use it for you know like there's two problem sets one problem set is just getting them to make the decision to use it right like okay i've hook up google and hook up slack and hook up whatever all these other tools great now it's there then there's problem number two well how do i use it oh it's like the first time i hired a virtual assistant cool i have a virtual assistant now.

34:05How do I use them? I have to educate them. I have to tell them exactly what I want. And sometimes I don't even know what I want, right? So what does the onboarding look like for new customers? Are you going through that process and asking them a bunch of questions or are you just figuring it out as you go? Like how does it? Yeah. Our approach has to do with figuring it out from the data that they connect. So we require our users to connect their emails. and when you have their emails, you can see what integrations they use and their patterns in using these integrations. And so you can figure out how to help people based on that.

34:42And just internally, like if you just give the agent access to that and you say, automate this stuff for this person, you'd be surprised as to how well it actually works. And so that's kind of our approach and pulling in things like style are also really important, like style for emails. Your chief of staff has to write the queue and talk like you. And then if they're sending emails as you, it has to be like you. So there's another really big thing with co-founder is like this preferences memory as well. Does it prompt you? It's like, okay, cool. I went and I studied you and I know everything about you.

35:16Does it then tell you, here are five things I can do for you? Or do you have to be pushing it to say, go write this email, do my scheduling? Like there's gotta be like a push and pull, I'm assuming. How does it work? Yeah, so you come on and we get your preferences for your email. and your calendar by reading your email and reading your calendar. And we figure out what you do based on all this info available on you online. And then we use all of those things once you're done onboarding to make four automations for you and say, hey, based on this, I made this for you. Yeah, and so you get the first couple.

35:51After that, you've got templates that people can use, but we're trying to make it so that constantly you're getting this info of like, Like, hey, I made your life a little better. Here's how you do that. Or like, oh, you know, I saw you're spending time on this. Here's how you do that. And then are there lots of like proactive things like that where, hey, you haven't talked to me in a while. Let me just send you this message saying, I'm noticing you're talking a lot more in Slack. Would you like me to create daily stand downs for you? We're a bit early on proactivity. We have like, we have a system that gives you these automations and it can run these every week.

36:26So it can give you new automations every week. But we want to make it much bigger than it is and much like really know you. So it's not all the way there yet, but it's coming very soon. What are you personally the most excited about building for yourself? Like selfishly. Obviously, you solve your scheduling problem. But as you look at this, maybe a year out, two years out, three years out, what do you think are going to be some of the coolest features that AI, maybe it's not even within co-founder but what do you think of the coolest features that are coming down the pike with ai i think the coolest feature that we haven't built this yet and no one's really built this yet is the ability to like have this sort of dashboard where you spin up a bunch of agents and you control like an army of agents basically so you know one of our goals is to make it feel like you're playing a video game but it's a real business in the real world and i don't know all the details to what that will look like, but I do know we're going to build it.

37:24It's amazing. That's the thing I'm mostly excited about. I know it's a little abstract, but it's like, okay, well, what if we can make it like playing a video game? Does any of this stuff freak you out? Like, does this at all make you nervous about where AI is heading? I think there are serious concerns about what people are going to do for work in like five years. I believe that one of the best avenues will be starting a company and like having equity in that company. But aside from that, like, I can see this thing doing the jobs of multiple people. Like, these white-collar jobs, these knowledge-word jobs are not long for this world.

38:02So that makes me nervous. And I don't think any of the experts have given convincing answers as to how we're going to go from where we are now to, like, this post-economic, you know, AGI society. They say, oh, well, yeah, I mean, you know, we're going to lose a bunch of jobs, and then AI is going to automate everything, and it's going to be great. what's like, well, okay, what happens when everybody loses their jobs, though? Like, there's a gap. If you look at these clips of these people, these CEOs of these big AI companies, they don't give the best answers. And they don't give concrete answers either.

38:35They basically say, I don't know. And so that also concerns me a little bit. Yeah, I'd be worried if our ability to scale energy production and capacity was clear. but we don't even have enough energy to be able to power all the things that we'd want to do with AI. I think that will be a bottleneck anyways which will give us time. If the power was all there right now I probably would be a little more worried because it feels like that's the limiting factor at this point but I don't know. We're all just making it up. Yeah, I think like, I mean it doesn't really keep me up at night but it is something that I think about and it is something that I think the world's going to have to reckon with at some point.

39:14Well, where's the best place for people to come find you or if people want to look at co-founder is it is it live is it beta is it invite only like how would you even mess around live on co-founder.co so you can go check it out there you can connect with me on twitter uh just my name andrew pipimelli and um you can also follow a company on twitter but yeah i mean we're hiring we're hiring great engineers in new york you can go and test out co-founder for yourself we're gonna have sock two in like a week so if people are concerned about data, then we got all of that. By the time this comes out, they'll have SOC too.

39:48Don't even worry about it. So that's it. That's how you find me. Cool. Awesome, man. Well, this was fantastic and we'll talk soon. Thanks so much, Nate. All right. Hopefully you liked that episode. And if you've made it this far, you're either really committed or you're stuck doing yard work and you can't actually skip on your phone. So while I have you, the show is growing, but I have a favor to ask of you. Will you please help me grow the show? I want to reach more people. there's a couple things that you can do like and subscribe is the simplest thing obviously you want to get notifications for when the next episode is coming out but if you go the next step will you leave me a review five star on spotify or apple what that does is it tells the algorithm that oh hey this is a high value podcast because more people are leaving reviews for it and it then pushes it out to more people so that's why when people are like will you log and subscribe and put the five star rating it's not just to make themselves feel better it's actually to get more exposure for the show.

40:42So if you do that for me, I would greatly appreciate it. And I'll see you next time.

From the publisher

MY NEWSLETTER - https://nikolas-newsletter-241a64.beehiiv.com/subscribe


Join me, Nik (https://x.com/CoFoundersNik), as I interview Andrew Pignanelli (https://x.com/@ndrewpignanelli). I sat down with Andrew, founder of The General Intelligence Company of New York, to dive into his incredible product, CoFounder. 

The core idea behind CoFounder is facilitating one person billion dollar companies. As entrepreneurs, we constantly struggle with the million paper cuts problem, all those tiny daily tasks that take up huge chunks of our time and brain capacity. Andrew explains how CoFounder acts as a chief of staff or self completing to-do list, extending your capacity without needing to hire a huge team.

We discuss how these AI agents use a persistent memory system and deep API integrations to understand you and automate crucial tasks like scheduling and managing emails. 

You won't believe how easy Andrew has made building complex automations using simple natural language, you can skip the need for complicated tools like Zapier altogether! We also explore powerful use cases like automated competitor monitoring, how CoFounder uses guardrails to manage tasks like sending calendar invites, and even integrating with tools like Limitless. 

If you want to see the future of autonomous companies and learn how to scale your business without scaling people, you need to hear this.


Questions This Episode Answers:

1. How can I grow my business without having to scale the people side of things?

2. What is the million paper cuts problem and how can AI agents fix it?

3. How can I build complex automations using just natural language?

4. What is the difference between an API and the Model Context Protocol (MCP)?

5. What specific tasks like scheduling, email, and competitor monitoring are CoFounder users automating the most?

__________________________

Love it or hate it, I'd love your feedback.

Please fill out this brief survey with your opinion or email me at nik@cofounders.com with your thoughts.

__________________________

MY NEWSLETTER: https://nikolas-newsletter-241a64.beehiiv.com/subscribe

Spotify: https://tinyurl.com/5avyu98y

Apple: https://tinyurl.com/bdxbr284

YouTube: https://tinyurl.com/nikonomicsYT

__________________________

This week we covered:

00:00 – Intro: One-person billion-dollar companies

00:40 – What CoFounder is and why it’s exploding

03:05 – How AI connects all your tools (email, Slack, CRM)

05:28 – The “million paper cuts” problem CoFounder solves

07:44 – Top use cases: scheduling, email, research, monitoring

10:07 – Treating AI like a human assistant

12:33 – Live demo: reading calendar, creating tasks, using tools

15:00 – The easiest automation builder (natural language flows)

22:04 – Onboarding: how CoFounder learns your style and workflow

29:12 – Competition, vision, and the future of autonomous companies

More from Nikonomics - The Economics of Small Business

All 108 episodes
253 - I Asked AI to Run My Business for a Week. Here's What Happened with Andrew PignanelliNikonomics - The Economics of Small Business · 41 min
Listen in VO