In short
Building an “AI chief of staff” using Nebula agent platform to automate executive work: calendar/email/Slack triage, daily briefings, project status tracking, vision/offsite follow-ups, and lead sourcing.
Guest
Imran (Nebula user/hosted agent builder; teaches use cases and how to create/copy agents on Nebula).
Key claims
AI chief of staff can replace much of what a human chief of staff does (manage calendar, prioritize, filter messages, track projects) so executives can “do more with less time.” Nebula reduces setup complexity versus more technical agent stacks (e.g., Hermes/OpenClaw) and supports easy integrations (Gmail, Slack, Calendar, Jira/Linear, etc.).
Notable examples
“Blockage Radar” agent that scans Slack/email to identify team members blocked; “Project Status” agent that summarizes shipped/due/at-risk items from Linear/Jira plus Gmail/Slack; “Vision Tracker” using Granola offsite notes with weekly progress updates and motivational quotes; “Agenda” agent that emails/DMs morning meeting summaries and top priorities; “Late Checkout Prospector” agent that finds 10 ICP leads daily via web/LinkedIn search (optionally using connection/affinity signals like shared city/schools).
Written by AI. May contain mistakes. Listen to the episode to check what was said.
Chapters
Tap a time to open that second in VOUnderstanding AI Chief of Staff
0:57 to 2:30
Exploring the role and tasks of an AI chief of staff.
“Imran, by the end of this podcast, what are people going to learn?”
Introduction to Nebula Platform
2:30 to 4:32
Overview of Nebula as an agent creation and deployment tool.
“So Nebula is essentially an agent creation, deployment, and interaction platform.”
Creating Custom Agents in Nebula
4:32 to 7:10
Step-by-step guide to building custom agents for task management.
“Is it just more technical on those platforms and this is just less technical?”
Integration and Functionality of Agents
7:10 to 9:35
Discussion on how to integrate various tools with agents for efficient task handling.
“By the way, I'll just say it because you're going to get flamed in the comment section.”
Developing a Blockage Radar Agent
9:35 to 12:00
Creating an agent that identifies team members who are blocked and need assistance.
“Basically someone that, you know, the job of an executive is to make decisions.”
Enhancing Productivity with AI
12:00 to 14:01
Exploring how AI tools can enhance productivity and streamline communication.
“And while that's being pulled up, I noticed that the model you're using is Quen 3.6+.”
Connecting Tools with Nebula
14:01 to 14:15
Learn how to connect Nebula with various tools using different methods.
Benefits of Nebula's Integration
14:15 to 14:37
Discover the advantages of using Nebula for tool integration.
“And there's even two ways to connect it.”
Utilizing WAP with Nebula
14:37 to 15:30
Understand how to use Nebula in conjunction with WAP for business operations.
“it's a little bit more difficult to set up because it's a little bit more of an enterprise product.”
Setting Up Messaging Agents
15:30 to 16:04
Learn how to set up SendBlue messaging agents to automate customer interactions.
“an FAQ video, it'll redirect them to book a call with me and automate that whole process.”
Show all 25 chapters
Creating a Project Status Agent
16:04 to 17:10
Explore how to build a project status agent using Nebula.
“You can, but SendBlue comes in as an iMessage.”
Importance of Keeping Track of Goals
17:10 to 18:35
Understand the significance of tracking team goals after strategy sessions.
“where it's helping me figure out what I should work on, helping me get unblocked mentally.”
Developing a Vision Tracker Agent
18:35 to 19:32
Learn about the creation of a vision tracker agent to monitor team progress.
“all that information from those off-sites and that keeps people accountable and that reminds them to do things and stuff like that.”
Incorporating Motivational Quotes
19:32 to 20:15
Discover how to motivate teams by integrating quotes into the tracking process.
“Progress on their work towards what we discuss in the offsite will either be stored in Gmail, Slack, or on Linear.”
The Role of Project Status Dashboards
20:15 to 20:52
Explore the function of project status dashboards in enhancing visibility.
“All right, we have our project status agent.”
Creating Mini Apps within Nebula
20:52 to 22:48
Learn how to develop mini apps that enhance productivity through integrations.
“We're going to see if we can get it to show us an anonymized sample so that we can see how this comes out.”
Future of AI Agents in Workflow
22:48 to 25:09
Discuss the potential future impact of AI agents on workplace efficiency.
“Yeah, let's see something around agenda and focus management.”
Designing an Agenda Agent
25:09 to 27:39
Understand how to design an agenda agent to manage meetings and priorities.
“like it's a coworker, like it's Slack, but then it actually goes out and does real work for you.”
Leveraging Personal Software for Efficiency
27:39 to 28:01
Explore how personal software can enhance daily productivity and workflows.
“and letting me know what calls I should be having that aren't already on my calendar.”
Exploring the Project Status Dashboard
28:01 to 29:19
Learn about a dashboard created with Nebula that displays project data visually.
“Okay, so this is just a quick little dashboard that Nebula made for me.”
The Rise of Personal Software
29:20 to 30:12
Discuss the shift towards personalized software solutions over off-the-shelf options.
“At the same time, if you look at the fundamentals of those businesses, they're highly defensible because they have so much distribution.”
Daily Agenda Agents and Reducing Stress
30:13 to 32:43
How daily agenda agents can alleviate stress for founders by managing tasks efficiently.
“And especially if you want to build the one person,$1 billion startup, the negative side of that is you're the one dependency for everything, right?”
Using Agents for Lead Generation
32:44 to 34:06
Explore how agents can help in identifying and reaching out to ideal customers.
“We'll do one that's not related to the chief of staff one.”
Building Agents for Sales Success
34:07 to 36:20
Discussion on creating agents to automate tasks for sales teams and improve efficiency.
“that is within two degrees of separation away from them.”
Allocating Time for Self-Supervision
36:21 to 40:02
Learn the importance of self-supervision and automation in modern careers.
“I mean, I'm a people demon, meaning I love people and I try to understand people as much as I can.”
Transcript
Automatic transcript. May contain errors.0:00Can you create a chief of staff with AI? One that handles your priorities, that's doing research for you, that's doing a lot of the legwork that a traditional chief of staff assistant could do. Well, in this episode, I brought on Imran and he teaches us a bunch of different use cases for how you can create your own chief of staff with an agent platform he uses called Nebula. Now, I'm not saying you should use Nebula, use whatever platform you want. But what I am saying is by the end of this episode, I think you're going to look at this and you're going to be like, how don't I have an AI chief of staff?
0:33I mean, in this day and age, in 2026, it makes total sense to have AI automating a lot of these tasks to save you time and make you more money. By the end of this episode, I had my creative juices flowing around how I can be automating more using some skills and some agents. Learned a lot. And I want to thank Imran for coming on and showing us the whole, whole thing. Thank you. And I'll see you in there.
1:06Grateful to have Imran back on the pod. Imran, by the end of this podcast, what are people going to learn? By the end of this podcast, you're going to learn how to build your own AI chief of staff. And if you're too lazy to do it, you can just copy our setups. And when you say AI chief of staff, what do you mean by that? Yeah, so the whole idea is that everything that a chief of staff would do, things like manage your calendar, help you figure out the most important things to work on that day, filter out annoying messages, and basically just manage all the projects or things that you're working on.
1:39We're going to build agents that will help you do all of that so that you can do more with less time. And when you say agents, we've been around the block. We've seen people pitch AI chief of staff. Is this a real AI chief of staff? Be real with us, Imran. For all the boring stuff that a normal chief of staff wouldn't actually want to do or that you wouldn't want to do, we can build a really, really good solution. There is no reason why a human should be looking at your calendar, your email, your LinkedIn messages. We can build agents to do most of that. All right. Let's see. Let's do it. All right.
2:14So I know in the last video we were covering Hermes agent. It was extremely technical. There was a lot of setup, and it required that you have a little bit of technical knowledge. So for today's video, I wanted to show you guys how to do it using a tool called Nebula. So Nebula is essentially an agent creation, deployment, and interaction platform. So you have channels where you can talk to agents, you can create custom agents, and now there's a new feature where you can actually share agents. So at the end of this video, we'll essentially have agents in the description that you can copy and bring into your Nebula account in case you don't want to build them yourself.
2:54So the way I like to do this in the beginning is I always like to first do some research. So in the beginning, we'll say things like, tell me what are three to five things and a chief of staff does. The Nebula agent itself is kind of like the master agent. So this agent is the one that can actually create other agents inside of Nebula. It's really easy to think of a custom agent or an agent inside of Nebula as just a specific kind of AI tool that has a set of instructions. So basically we call that a system prompt. It has goals. You can define what the goals are. And it has integrations. So if it needs to use Gmail, Calendar, Jira, whatever it is, you can specify and you can organize each agent based on the tools that it needs to use.
3:49And it's basically the goal for what you want it to do. I find this really important because for a lot of us, we work in environments where we work with teams where everyone has a specific role and they have tools that they use. So when trying to automate a lot of the boring stuff, the easiest paradigm is to use the one that we're already using, which is there are certain people who have certain goals inside of a company and in order to achieve those goals they use a certain set of tools. So just like that, when we think about agents, the easiest way to start using these is to think about them the same way that you would think about a human, with just maybe smaller tasks.
4:27And how should people think about, why would I do this on Nebula versus setting up an OpenClaw or stuff like that? Is it just more technical on those platforms and this is just less technical? Yeah, that's a really good question. The last time I was on the podcast, we talked about Hermes, we talked about OpenClaw a little bit. Actually, we called it Hermes. We called it Hermes. I called it Hermes. I'm sorry, I got a lot of flack on Twitter for that as well. But it is pronounced Hermes. We talked about it last time. you still need to have some technical know-how in the sense that you have to open up a terminal you still have to manually update it you have to go ahead and configure the models and if you're someone that's generally interested in tinkering with these tools, it can be a really fun experience you can actually save a lot of money too but I think Nebula has abstracted away a lot of that complexity if you're someone who's extremely busy, you already have a business or you already have a really busy job and you just want to get things automated if you just want AI working for you, this is the fastest way to get started.
5:32And so that's why I elected to use it for this one. Because typically speaking, if you want a chief of staff, you're probably a busy person. And if you're a busy person, you probably don't want to spend time setting up your little private AI server. There are, of course, companies where people, like the actual organization, will require that you run AI locally. But this is kind of built around that. and I know that the Nebula team has some stuff coming for that as well. So yeah. And it's also a very familiar interface. It looks like a chat. It looks like Slack. You have different channels. And then instead of teammates, you have different agents, right?
6:07So it's very familiar and it just kind of works. All right. So Nebula here tells us a chief of staff is a force multiplier for an executive. They focus on strategic planning and execution, agenda and focus management, cross-functional alignment, communication and stakeholder management, and special projects. So the first one that kind of like jumps out to me here is the strategic planning and execution. So you can imagine if you're an executive or even if you're just like running a business, you get a lot of messages every single day. And there is a certain level of like you needing to execute, but you also needing to unblock the people that are like the people that are actually executing on your behalf, right?
6:52And just getting work done. So the first type of agent that I want to create with Nebula is one that shows me, based on my Slack and my email, if anyone on my team is blocked. So let's say, let's go ahead and make an agent that shows me who on my team is blocked. By the way, I'll just say it because you're going to get flamed in the comment section. Why are you not using a Whisper Flow or something like that and typing with your God-given hands? I do have Super Whisper installed. Let's do it with Super Whisper, actually. Good catch, by the way. Let's make an agent that goes through my emails in Slack and lets me know exactly who on my team is blocked and waiting on me for something.
7:50And this is for the people. First of all, I'm just trying to protect you. You're my homie. No, I appreciate that. Also, to the people in the comment section and the people listening, I read every single comment. And I respond to most of them, or a lot of them. And so some of them hit me hard, you know? Some of them hit me hard. It's true, though, right? Because with Super Whisper or Whisper Flow, you can speak at like 150 words per minute. I don't know anyone that can type at 150 words per minute. You want to talk about real productivity? I think that's probably the biggest lift. Yeah, that's true.
8:33Okay, so Nebula, you can see Nebula here is doing a bunch of stuff in the background. It's really just searching through tools and toolkits. It sees that I already have Gmail, Slack, and agent management available. If that wasn't set up, I would just jump into the settings. and I would go over to integrations. And then here I can scroll down to where the tools are, and I would search for the tools I need. So maybe I wanted calendar, I could connect calendar. If I wanted Jira, I could connect Jira and things like that. So pretty straightforward. And you can see here on the left side, Nebula has already created a blockage radar agent.
9:10That's cool. Yeah, and it's going through. By the way, I'm one of those guys who, I want a chief of staff. And I don't have one. I think you need one too. When I was reading the description of chief of staff, because when I hear chief of staff, I'm kind of like, what does that even mean? But then when I saw the description, I'm like, oh, I want that. Yeah, yeah. Basically someone that, you know, the job of an executive is to make decisions. So everything around that is what a chief of staff's job is, it's optimizing the executive to make decisions. Cool, okay, so I see we have a blockage radar set up now, and it's on the left side as an agent.
9:52We're going to click on that. This is a separate agent here. And I think it's still running right now. You can see here also on this panel, so if I click on this, you can see the agent details. So here I have the name of the agent, I can specify which model it's using. so let's say I want to use the Nebula model I can use that one because it's cheaper and then there's a description and then there's goals so this is like a lot of what we talked about in the last video that we did together was about how tools like OpenClaw struggle with memory by segmenting out all of your tasks that you need to do into sub-agents you can actually fix a lot of the memory problem here as well Hermes solved the memory problem by having a self-learning loop if you just use sub-agents or agents inside of Nebula you can solve the memory problem by specifying the goals and that just gets tagged into the system prompt every time you can see exactly which tools are here that are set up already, you can see which accounts it needs and then the coolest thing that they just launched was that you can actually change this to public, so I can make this public it'll anonymize all the connections, so you won't be connected to my Gmail, and I can just click this button right here share this tab instead and now you can see that I have a blockage radar agent and there's a URL right here so maybe we can throw that URL in the description and you can just take this and clone this agent so if you like how my blocker agent works you can literally just bring it right into your Nebula setup and try it out for free that's really cool I think that was one of the things I was missing people were sharing skill files people were sharing their tools people were sharing CLIs but I want to be able to share something that I've built for me.
11:40And then you can come in and tweak it. So maybe you use Gmail, Slack, and Telegram. So you can come in and add Telegram and tweak it and then remix it and share it yourself. So I think that's super powerful. So let's see, the next run will be in six hours. I'm not going to run this because it's connected to my work Slack so I don't want to show you guys this stuff. But let's see, let's see a sample. Let's see a sample, no real data. And while that's being pulled up, I noticed that the model you're using is Quen 3.6+. What's the thinking there? One is the price. It's just much cheaper than Opus.
12:20But also, if I need something that's going to go through my email and Slack and tell me what needs my attention, I don't think that we need a state-of-the-art model for that. I feel like using Opus for that or Sonnet is probably a waste of money and resources because this is a very basic task if I was doing coding or if I was doing some tasks that required deep thinking yeah I would probably use the latest models but I would rate these tasks as pretty straightforward even an older LLM could tell you what's waiting on you Yeah. And then I had to spit out a sample briefing. So we've got emails waiting on me, Slack messages waiting on me, and then a summary.
13:03Pretty straightforward, right? But having this sent to you every morning like two years ago was like a job that you would hire someone to do. Right, like part of it. So cool, let's jump back into Nebula. Let's find the next one. There's the cross-functional alignment. This is a good one. So I'm thinking, let's make an agent that takes a look at Jira, Confluence, my email, and Slack, and lets me know about the status of every project that I'm working on. It should show me what work was completed yesterday, what's on the docket to be completed today, and let me know if any projects are at risk of falling behind schedule.
13:46That sounds like the job description of a project manager. and we're able to make an agent to do it. So let's see. Let's say I use linear instead of JIRA. Is that possible to connect into that? Yeah, totally. So you can specify that you want it to use linear and then again you can go into settings, integrations and you can search for or we can click add tool and then we search linear and then you can go ahead and connect it this way. And there's even two ways to connect it. you can do OAuth or you can do with the API key. OAuth uses this thing called Composio. And I've noticed that Nebula is really good at working with things like Gmail, Slack, and Google Calendar because they use Composio to manage the connections.
14:33Whereas tools like Hermes and OpenClaw, although you can use Composio, it's a little bit more difficult to set up because it's a little bit more of an enterprise product. So that's one of the benefits of using a tool like Nebula. You have basically the best-in-class bridges to connect to all the software that you use. I know this is a side tangent, but one of the tools is WAP. How would someone use Nebula with WAP? Yeah, so we run a line of our business on WAP. So I'm using it to keep track of revenue. I'm using it to basically give me more information about our customers. send out. We have a SendBlue messaging agent, so we talk to a lot of our customers or prospective customers via SendBlue.
15:23And I have the SendBlue agent actually, it has a goal. So if someone comes in and wants to learn more about our offer, it'll send them an FAQ video, it'll redirect them to book a call with me and automate that whole process. Cool. The one thing I will say is if you do set up a SendBlue messaging agent, make sure you follow their rules so your number doesn't get burned. With a lot of these automated outbound stuff, there's some nuance.
15:55You don't want to get your number marked as spam or you don't want to get read limited. Just make sure you read the docs. Can you mask your number? Aren't there tools to mask your number? You can, but SendBlue comes in as an iMessage. It's a little tricky because I believe it's like peer-to-peer. So they have their own internal limits on how many messages you can send a day. Okay, cool. Oh, this is a cool thing too, right? So we're jumping back into the Nebula agent. So you can see if it has a doubt about anything about building your agent, it'll just ask you a question. We can say we'll use GitHub issues or maybe we connect JIRA only.
16:39Let's say you want it to use linear. We can say, oh, actually we'll use linear. What time should the briefing arrive? Let's say 8 a.m. Pacific. How should I deliver? Let's say Slack DM. That's a nice touch. Yeah. It'll come in and actually ask you exactly how you want it. This is, I think, super powerful. I think in the last video we covered a lot about Hermes being a really good personal agent. I was even using it for some very elementary version of therapy where it's helping me figure out what I should work on, helping me get unblocked mentally. But if you actually just want to lock in and get work done, this is the way to do it.
17:25Cool, and now we have our project status agent. So it's going to chug along and do its thing. Let's see, scan linear, Gmail, and Slack and build a daily briefing. so greg i have a question for you yeah now that you're seeing this what would be the first agent that you would go back and create after this if you go back to the uh definition of a chief of staff yeah my like you know as uh as a founder i'm kind of thinking about what are the highest value activities that I could be doing? So when I see the strategic planning and executives, so translating the CEO's vision into actionable parts, tracks key initiatives, make sure priorities actually move forward.
18:15So we have offsites. And in our offsites, we have these big goals and these big visions and we leave these offsites feeling really good. And then what happens is we kind of lose some of those docs and we lose track a little bit because things get busy. In an ideal world, I have an agent that has all that information from those off-sites and that keeps people accountable and that reminds them to do things and stuff like that. How often are these off-sites? Once a quarter? Once every six months. They're kind of like big vision off-sites. Okay. Let's make it. Let's do it this way. Let's make another agent called vision tracker.
19:08It should take notes from Granola where we record our notes for our off sites. And every week for each person who was inside of our or who was at our off site, it should report on their status towards the goal or the vision that we discussed in the off site. We do our offsites twice a year and the notes are stored in Granola. Progress on their work towards what we discuss in the offsite will either be stored in Gmail, Slack, or on Linear. Can I add one more small thing to that? Let's do it. So I think it would be really helpful to include in this DM or however we get it one quote that is from a famous person, call it, that just is related somehow to motivating people to actually complete the goal.
20:10Okay. Let's do it. Awesome. and we'll be able to share this to you we're open sourcing you greg there we go
20:25yeah very interesting so what do you think that would unlock for your team you think that they would just get more work done towards the ultimate vision i think that it's always you know i don't blame the team at all and sometimes i'm this way too i just you lose track of the big vision as you start working in the business, not on the business. That makes a lot of sense. All right, we have our project status agent. We're going to see if we can get it to show us an anonymized sample so that we can see how this comes out. And then again, same thing. I'm going to jump in the right side. I'm going to click this side panel here.
21:08I'm going to change the model to the Nebula model because I am price sensitive and I'm going to go public on this. Again, I don't think you need Opus 4.6 to be able to give me a briefing on the status of a project. And then you can see, if you go to this link and we can put it in the description as well, you can see now that we have the project status agent, it's got linear, Gmail, Slack bot, and then you can just literally come in and clone it. All right, let's see what the sample looks like. So you can see here it would have what shipped yesterday, the top three cards, or whatever shipped yesterday, what's due today, and then things that are at risk.
21:48Very straightforward. I think this is pretty useful. Totally. You know what it gets me thinking of is how many companies don't have agents, are not running with agents to 99.9%. Another thing I wanted to show you too is there's also this notion of mini apps. So we can say, can you create a mini app for this where every day it has the number of tickets that were completed the previous day at the top, how many tickets are due today, and then the at-risk items and be able to deep link into each linear card. obviously we haven't connected linear yet so you can just use samples there so there's actually basically like a replet inside of Nebula where you can spin up and host these mini apps that connect to your agents this isn't something I've really seen anywhere where it's kind of like a new paradigm essentially you can make a web app that when you click a button can actually make a call back to an agent and then the agent could update the web app in real time I haven't seen this yet this is something kind of new but I think this is going to be pretty sick so I'll let this run in the background and let's see where our vision tracker is out okay looks good we just have to connect granola and let's do the same thing give me an anonymized sample of what the weekly vision tracker would look like just so that we can see the format let's see if it'll insert the quotes in here too if it doesn't it's fired you don't need HR for agent number one value add I'm trying to think of a more fun there's got to be a more fun one we can do something a little maybe off the cuff we did cross functional alignment that's kind of the project management stuff agenda and focus management That could be good.
23:57Yeah, let's see something around agenda and focus management. So one thing I have set up, and I think I showed you in the last video, is that I have my own kind of daily briefing, but it's more personal. So what if we can actually make daily briefings for everyone on the team? Yeah. And then show the top three priorities for the whole team. That would be cool. Let's make sure this one's done. Oh, one really interesting thing about using Nebula is that when you actually fire off, when you tell the main Nebula agent to create another agent, it doesn't like clog this main thread. So you can see like the vision tracker is still working.
24:32The project status agent is still working, but I can still jump into the main Nebula agent and talk to it. That was one thing that I felt like was a drawback of Hermes is that like I would go spin up my own web UI, like my own like webpage to manage it, or I would have like different Telegram channels, but I felt like multitasking in the same kind of workspace was like kind of tedious, right? I have to keep waiting for a query to be done. So that's another thing that I like about tools like Nebula and whatever else is going to come out. I think this is going to become, if we look at the industry in general, I think this type of tool is going to become the standard where you're talking to agents like it's a friend, like it's a coworker, like it's Slack, but then it actually goes out and does real work for you.
25:13Let's do the agenda one, okay. Let's make an agenda agent that every morning at 6 a.m. goes through my calendar and lets me know who I'm meeting with. It gives me two lines about each person I'm meeting with. It tells me my top three priorities for the day. And it has any relevant reminders from my email that maybe haven't been put in my calendar for that day. What else would you add, Greg?
25:46That's by looking at the calendar. but what are things that like that's by looking at book stuff but what are meetings that I should be booking that I'm not booking you can say maybe you guys have a knowledge base so I would say check my notion check my email of course check Slack and also cross reference it with the vision docs that we created earlier in the year Yeah, and I think the use case I was just thinking about is maybe in my notion I have a little note that says, you know, Imran, six months ago, let's just say for my agency or something, Imran told me six months ago that in six months he wanted to build an app and needed design help and engineering help.
26:42Most of the time you forget about that stuff, right? If it actually went into my second brain and was like, oh, by the way, I noticed that you haven't met with Imran. Six months ago, he said that you guys should meet. I should know about that. I should be reminded of that. We're human beings after all. Do you have a second brain setup right now in Notion? Is that what you're doing? Yeah, I use a mix of Notion and Google right now. I'm also using a little bit Obsidian. I haven't made the full move over to Obsidian yet, TBD. But yeah, right now it's between Notion and Google. Cool. If there's someone I should catch up with and that information is stored in Notion, please resurface it.
27:34Essentially, I want you to be proactive about following up with people and letting me know what calls I should be having that aren't already on my calendar. You can kind of just word vomit these things and it just works, as you can tell. That is the one downfall of using Super Whisper that just kind of gets you to just keep talking. But it'll clean it up. All right, let's see this project status dashboard. Okay, so this is just a quick little dashboard that Nebula made for me. So in case I didn't want to see this text, in case the text gets boring, it actually spun up an entire web app for me, which is publicly accessible as well.
28:17And so let me share this tab. So this is now a full dashboard that I have. Again, this is sample data, but essentially I could plug in my real data that the agent will resurface every morning. And then maybe I don't want to see everything as a daily digest inside of Slack. Maybe I want to see it visually. This is a way you could do it. And I have a few mini apps as well. I have one for trending GitHub repos. This is one I made where every morning it'll update it with the most trending GitHub repos and I can click into them, get more information and go see what's up. This is one I have running for myself.
28:51So yeah, that's mini apps. Yeah, mini apps to me feel like what they call personal software. Exactly. Yeah, which is this idea that it used to be in the future but now it's the present. In the present, instead of going to find off-the-shelf software for certain use cases, you'll just spin up software specific and personalized to you based on your use case. Exactly. I'm really confused about this when I think about the industry because on one hand, people are saying now that you can make software, companies like Duolingo and Salesforce have no moat because anyone can go make a CRM, anyone can go make a language learning app.
29:33At the same time, if you look at the fundamentals of those businesses, they're highly defensible because they have so much distribution. And even though for the last 36 months people have been able to create their own personal software, I haven't felt like it's replaced things at the enterprise level. And I think there is more that goes into enterprise-level software, whether that's compliance, uptime, and things like that. but I think personal software is like you said it's like a new category it's something we've never really had that's accessible so it's really cool that we can just do it inside of a tool like this for a couple bucks totally alright daily agenda agent it's going through again I didn't have to configure any of these tools I didn't have to tell it which skills to do and if I click in it already knows the goals and it already builds out a system prompt and then tells me exactly what it needs to connect to and if I don't have it connected it prompts me to connect it right here which is pretty cool so how many hours a week do you think this would save?
30:39me? yeah for you personally you know the truth is it wasn't the hours that was that's intriguing to me it's like the the anxiety and the stress that you get right like there's so much if you're a founder you are just dealing with a million fires at any given time and there's just so much going on. And especially if you want to build the one person,$1 billion startup, the negative side of that is you're the one dependency for everything, right? Which no one really talks about. But the positive side is you're a 1 % billion dollar company. Champagne problems, right? But I think what you need in both those cases, founder or if you want to build a one one person billion dollar company is you need to as you said in the beginning be making decisions and you can't be spending time and like in the back of your mind stressing about who am i meeting today or let me go and like check their linkedin recently or i know I should be reaching out to so and so but I'm not sure right so I think that like why I'm I love like tinkering with tools like this is is because I'm a hobbyist frankly I'm a tinkerer and I'm always looking for ways to feel superhuman you know yeah this is yeah this is it's also like a little it's on the same time it's anxiety inducing for me in the sense that like i know i can be doing so much more with agents that i'm not doing already like that's i think the biggest uh the biggest thing that i've been like wrestling with mentally it's like man like i can do everything now but i have to really like choose like zone in on the few things that will actually move the needle for me and i think that's going to become the new skill set is like actually being able to have judgment and like pick what to work on yeah and and frankly that's like why i wanted to have you on the show which is like giving people and myself ideas around what are the different types of agents, sub-agents that you should be building if you're building a business or you're working at a company and you're trying to save time and make more money.
Read the full transcript
33:02We'll do one that's not related to the chief of staff one. I know we talked a little bit about the SendBlue agents. If you're in sales and you want to set meetings or you just want to talk to your customers, you can do it via iMessage. Since our last podcast, I've gotten a lot of inbound of people who have been asking me questions about agents. Mainly the people who have been really interested are people working in e-commerce who are doing automated dashboards. They're enriching their customers with data from other sources. Maybe they'll go search someone up on LinkedIn or maybe they'll set up a flag for a famous person that they can send them more stuff.
33:44That was one use case I saw. Or with people in the sales teams, one thing that they're always looking for is just more leads. How do I actually figure out, okay, I know my ideal customer, how do I go search that? LinkedIn has a little bit of a monopoly on that. But I've had some people DM me and say that they actually set up agents that will go out and scrape and find their ICP that is within two degrees of separation away from them. So that type of stuff is pretty cool too. We could do one of those or any other ideas you have? Yeah, let's do that. Yeah, okay, cool, cool. So let's do it for late checkout.
34:22What's your ICP? Well, for our agency, we work with the biggest companies in the world like Salesforce, Nike, Dropbox, all these big companies. And so our ICP is actually the chief product officer or the CEO of those companies that are looking for how can I do AI transformation and also a new product suite that's AI first. It's very specific. There's only a few thousand people on the planet that meets this. Cool. Let's see if we can actually search potential ICPs. I know obviously most of your business is word of mouth anyway. and the top of funnel is this podcast that's part of it but let's try it let's try it let's build an agent that sources 10 people who are my potential icp for a company called late checkout every day it should search the web or linkedin and just find me 10 of these people that fit the ICP, bonus points if I'm connected with them on LinkedIn or I'm a second degree connect with them on LinkedIn.
35:37You know what's another little hack there before you hit it? I find that if people come from the same city, went to the same schools, if there's some amount of commonality, I find that there's a greater chance of it working out. I mean it happens to me all the time even like people reach out to me and like oh I went to McGill which is like the school I went to and I'm like more likely to be like oh yeah let's meet for coffee type thing or something like that. Okay. That's actually a good one. I didn't know we had a sales demon on the podcast right now. I mean, I'm a people demon, meaning I love people and I try to understand people as much as I can.
36:30And that's where that's coming from. Yeah, that's good. I mean, this stuff is really important because I know everyone wants to know. Of course, if you're operating at the executive level, there is a notion of your job is to make decisions every day. But there's also just a subset of people that just want to make money. like they're just really like directly or like what's the fastest way to get this thing to make money well it's if you're starting a business you obviously always start with an icp you talk to your icp you figure out what their problem is and then you try to package a solution for them whether it's a startup or agency info product whatever it is right um so this is like a really easy way for people to find leads right like i work at a fund every single day founders are coming up to us and saying like yo we need more leads it's like well this is a good way to do it so let's see it's building it right now.
37:16Late checkout prospector. So it's using a search. Yeah, it's using search, not LinkedIn. It's interesting. Yeah. Think this one we won't make public? This one you got to build yourself. We can make it. We can make it public. I got nothing. You know, I got nothing. Cool. Oh yeah, another thing about Nebula, you can just select different models to you just like you were able to with what's the nebula model i don't know how much i'm allowed to talk about it but i do know the specifics and i know it's really good okay yeah i don't i don't i don't even know i need to know the specifics i just need to know like when are we using nebula yeah like are you using the nebula model you know generally speaking the nebula model is probably one of the best bang for buck models okay um i don't know how much i'm about it publicly but it's it's like it's like two very good models it is it is a base model and a supervisor model and the supervisor comes from a very expensive model and the base model is a you know another really good model so basically it's a lexus you know it's a lexus bank for buck bank for buck yeah we don't need a ferrari all the time you know sometimes you I understand.
38:40We'll make this one public too. Ta-da. And this will be our last one. And at the end, you can leave us with one thing that we should be thinking about. Okay. Yeah. Cool. Oh, I see some people already. We're cooking. We're cooking. Dell, Expedia. Oh, wow. There we go. So this one actually just went ahead and just found the leads. It didn't even need my LinkedIn connection. This is interesting. Some of these people on this list, some of these companies on this list are in fact clients of ours already. There you go, man. You know your ICP well. Yeah. Yeah. All right, so if there is one thing I will leave you guys with after this video, I think no matter what role you're in, whether you're a founder, whether you're a PM, whether you sell solar panels, I think you have to figure out a new split of your time, which is like in business you had this idea that you would work in the business and then work on the business.
39:42I think all careers are going to become like that where you need to allocate a percentage of your time to work on your job. To actually take a step back and supervise yourself and be like, okay, what did I spend time on today? And then how do I automate parts of that using agents? A lot of people were already doing this before AI with tools like Zapier and NAN. They've been around for a long time. But now that we have the ability to actually reason on the data, all these automations are much more powerful. So I challenge everyone to almost watch themselves work over the next week and try to pick three to five things that they can automate out or build a tool for.
40:22Challenge accepted. I have done some of this, but I'm not fully there yet. and frankly it's like it's never ending that's how I feel about it right once you integrate this then you're like oh you know what I should do this or you know it would be cool if it adds this and it's always getting better and getting better so it isn't you know just for people listening it's not going to be this is not going to be something that you can do in a week and it's probably not something that you can do a year it's just going to be something that you're constantly it's a muscle that you're constantly going to be flexing flexing flexing getting bigger bigger bigger and thank you for Imran coming on, showing us this, showing us some use cases and I'll include links on where to follow Imran and connect with him in the show notes, in the description and I want to thank you again for coming on and I always enjoy chatting with you.
41:18Yeah, you as well, man. And we'll include links to play with some of these agents. Yeah. All right, take care. Later.
From the publisher
I'm joined again by Imran Muthuvappa to walk through how to build your own AI Chief of Staff using a tool called Nebula. Imran shows me how to spin up specialized agents that handle the work a real chief of staff would do — surfacing team blockers, tracking project status, holding people accountable to offsite vision goals, running daily agenda briefings, and prospecting ICP leads. We also get into mini apps, model selection for cost efficiency, and why personal software is becoming a real category. By the end, the takeaway is clear: every role now has a "work on the job" component where you supervise yourself and offload tasks to agents.
Links Mentioned:
Try Nebula: https://startup-ideas-pod.link/nebula
Precall Agent: https://startup-ideas-pod.link/precall-agent
Project Status Agent: https://startup-ideas-pod.link/project-status-agent
Lead Gen Agent: https://startup-ideas-pod.link/lead-gen-agent
Timestamps:
00:00 – Intro
02:13 – What is Nebula
02:54 – What an AI Chief of Staff Actually Does
4:26 – Nebula vs OpenClaw vs Hermes
06:09 – Building the Blockage Radar Agent
09:47 – Agent Features
12:04 – Choosing Cheaper Models for Simple Tasks
13:15 – Building Project Status Agent
13:53 – Connecting Tools to Agents
17:38 – Building Vision Tracker Agent
22:07 – Mini Apps and Personal Software as a New Paradigm
25:13 – Building a Daily Agenda Agent and Second Brain Integration
30:25 – Hours Saved vs. Anxiety Reduced for Founders
33:02 – Building the Lead Gen Prospector Agent
39:19 – Final Thoughts: Automate Three to Five Things
Key Points
An AI Chief of Staff handles the boring executive support work — calendar, email, LinkedIn, project status — so a human can focus on decisions.
Nebula lets you build, deploy, and share custom agents through a Slack-like interface, where each agent has its own goals, tools, and system prompt.
Voice input via SuperWhisper or WhisperFlow gets you to roughly 150 words per minute, which Imran calls the biggest productivity lift available right now.
Cheaper models like the Nebula model handle most chief-of-staff tasks well — reserve frontier models like Opus or Sonnet for deep coding or reasoning work.
Mini apps inside Nebula are spinnable web dashboards that connect back to your agents — a glimpse of personal software replacing off-the-shelf tools.
The new skill is judgment: picking which three to five things to automate out of your week.
The #1 tool to find startup ideas/trends - https://www.ideabrowser.com
LCA helps Fortune 500s and fast-growing startups build their future - from Warner Music to Fortnite to Dropbox. We turn 'what if' into reality with AI, apps, and next-gen products https://latecheckout.agency/
The Vibe Marketer - Resources for people into vibe marketing/marketing with AI: https://www.thevibemarketer.com/
FIND ME ON SOCIAL
X/Twitter: https://twitter.com/gregisenberg
Instagram: https://instagram.com/gregisenberg/
LinkedIn: https://www.linkedin.com/in/gisenberg/
FIND IMRAN ON SOCIAL
X/Twitter: https://x.com/imranye
Alif: https://alif.build/
