16:22Alex Finn · Listicle
7 Tips That Turn ChatGPT 6 Astra Into AGI
One creator's seven-tip playbook for turning ChatGPT 6 Astra's computer-use and reasoning upgrades into real workflow gains.
September 7thThe argument in one line.
A company can run with almost no human employees if you rebuild its existing functions as specialized AI bots that report to a chief-of-staff bot, feed on real context, and talk to each other the way departments already do.
Read if. Skip if.
READ IF YOU ARE…
- A solo founder or small-team operator who wants to delegate specific functions, like recruiting, marketing, or ops, to AI instead of hiring for them.
- Someone already using a general chatbot but frustrated that it forgets context between sessions.
- A business owner curious what 'AI running the company' actually looks like in practice, not theory.
SKIP IF…
- You're looking for a step-by-step technical build guide — this is a philosophy and workflow overview, not a tutorial.
- You need security or compliance guidance for handing AI access to email, Slack, and financial accounts — none of that is addressed here.
The full version, fast.
Anthony Pompliano argues a company is the most productive unit of labor there is, so the right way to use AI isn't to reinvent business, it's to rebuild each existing function as a specialized bot. He set up a chief-of-staff bot in Grok, had it interview him for context on his businesses, people, and goals, then connected email, Slack, and calendar so it has real-time state. From there he spun up bots for recruiting, product design, and more, all reporting to the chief of staff and holding daily standups with each other. He also runs Sylvia for personal finance and Instinct for day-to-day tasks. His bet: as context and models compound, AI increasingly becomes the brain of the operation, freeing humans for the parts of life only humans can do.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Create a free account →Where the time goes.
00:00 – 00:48
01 · Why I'm building a company with almost no human employees
States the video's central bet up front and grounds it in why a company is the highest-leverage unit of labor to automate.
00:48 – 02:20
02 · The breakthrough: bots that finish long jobs, not just answer questions
Explains what actually changed with AI: bots can now complete persistent, multi-hour jobs, so companies get one specialized bot per department instead of a single generalist assistant.
02:20 – 04:28
03 · Turning Grok into an AI chief of staff
Reframes AI as the brain of the company and walks through signing up for Grok, creating a chief-of-staff bot, and letting it interview him for context on his businesses, people, and goals.
04:28 – 06:57
04 · The recruiter bot that found 10 candidates
Shows the specialized-bot model in action: a recruiting bot returns a stack-ranked candidate list with move-likelihood scoring, and a real candidate gets contacted and impressed within 24 hours.
06:57 – 08:41
05 · The product-designer bot audits the website, then the daily bot standup
A design bot critiques the live site and the source code together, and Pompliano describes running a literal daily standup meeting between his bots.
08:41 – 10:49
06 · Beyond Grok: Sylvia for money, Instinct for daily life
Introduces two more AI products: Sylvia for portfolio-specific personal finance, and Instinct, an iMessage assistant that books things and grabbed a canceled ticket reservation at 2:30am.
10:49 – 12:58
07 · Managing bots instead of people, then the ask
Closes on the shift from managing people to managing people and bots, predicts more automation ahead, and asks viewers to subscribe.
Lines worth screenshotting.
- A company is the most productive unit of labor in the economy, so the right AI strategy rebuilds its existing functions rather than reinventing what a company is.
- AI's real breakthrough isn't answering questions, it's finishing persistent, multi-hour jobs on its own without supervision.
- Treat AI as the brain of the company and the human's job as feeding it context, not as a tool a human team happens to use.
- The more context an AI has, the more valuable it becomes, so context-gathering is now a core management job, not a one-time setup step.
- Interviewing yourself through the AI produces better onboarding than writing one long prompt, because the AI asks for what it's missing.
- People outside your formal org chart still need to be documented for the AI if their context overlaps with your business.
- A specialized recruiting bot didn't just return names, it stack-ranked candidates and scored their likelihood of leaving their current job.
- Tenure length alone means nothing: a 5-year employee is a strong move-candidate, but a 5-year founder at the same company is not.
- A single bot that sees both the visual site and the underlying source code catches SEO and conversion issues a human review misses.
- Running a daily standup between your bots, same format as a human team, makes each bot noticeably more useful over time.
- A finance AI given access to every real account and private asset gives portfolio-specific answers, while a general chatbot only gives generic advice.
- A persistent, low-cost assistant that never gives up can grab a canceled reservation at 2:30am while its owner sleeps.
- The job of running a company is shifting from managing people to managing people and bots at the same time.
- The end goal of automating the business isn't working more hours, it's getting time back for family, hobbies, and being human.
Build a chief-of-staff bot, then specialize from there.
AI OPERATIONS
One AI acts as chief of staff routing work to specialized department bots, and the human's real job shifts from doing the work to feeding the system context and asking better questions.
01Why I'm building a company with almost no human employees
- Treat a company as the most productive unit of labor in the economy, then ask how AI can automate the functions inside it rather than reinventing what a company is.
- The leap forward isn't AI answering questions, it's AI completing long, multi-hour jobs on its own, which is what makes real automation possible.
02The breakthrough: bots that finish long jobs, not just answer questions
- Build one bot per business function, marketing, finance, ops, product, growth, engineering, instead of one generalist assistant.
- Have those specialized bots talk to each other, not just to you, so information moves between departments automatically.
- Reframe AI as the brain of the operation and yourself as the one who feeds it context, not as a human team that merely uses a tool.
03Turning Grok into an AI chief of staff
- Interview yourself through the AI: have it ask questions to extract your life, your businesses, and your team instead of writing one long prompt.
- Include people who sit outside any single business line if they create crossover context the bots will eventually need.
- Connect email, Slack, and calendar once the static context is loaded so the AI has real-time state, not just a one-time briefing.
04The recruiter bot that found 10 candidates
- List every role your company actually needs, then create one specialized bot per role instead of hiring or searching manually.
- Ask for more than a name: have the bot return a stack-ranked list with the match rationale and a likelihood-to-move estimate before you spend time reaching out.
- Tenure alone doesn't signal willingness to move — a 5-year veteran and a 5-year founder read the same on paper but need opposite outreach.
05The product-designer bot audits the website, then the daily bot standup
- Point a bot at your live site plus your source code and site map together, since visual critique and technical SEO findings compound when the same bot sees both.
- Feed every new mockup or screenshot back through the same critique bot before it ships, turning review into a habit rather than a one-off audit.
- Run a daily standup between your bots, what did you do, what's today's goal, what are you blocked on, the same ritual you'd run with a human team.
06Beyond Grok: Sylvia for money, Instinct for daily life
- Point a specialized finance bot at every account, bank, brokerage, crypto, cards, plus your private assets so its answers are specific to your actual portfolio, not the generic answer a general chatbot gives.
- A persistent, cheap, always-on assistant beats a smarter one that only responds when asked, since it can sit on a task for hours and grab the outcome the moment it's available.
07Managing bots instead of people, then the ask
- Expect the job to shift from managing people to managing people and bots at once, and treat learning to direct bots as the new core skill.
- The payoff of offloading execution to AI isn't more hours at the desk, it's getting time back for the parts of life a computer can't do for you.
Terms worth knowing.
- Grok / Grokbot
- The AI chatbot Pompliano built his 'chief of staff' and specialized department bots inside of, using it as the central hub that routes tasks between them.
- Chief of staff bot
- A single AI agent that holds the full context of the business and relays tasks to specialized bots for marketing, finance, recruiting, and other functions.
- Sylvia (CFO Silvia)
- Pompliano's AI product for personal finance, built by connecting every bank, brokerage, crypto, and private asset account so its answers are specific to the user's actual portfolio.
- Instinct
- An iMessage-based personal assistant AI that reads a user's email, prioritizes tasks, and takes action on their behalf, such as booking reservations or flights.
Things they pointed at.
Lines you could clip.
01:55
“Artificial intelligence products, they're the brain of the company. The human's job is to give the AI as much context as possible.”
08:23
“You actually can have all of the bots hold a meeting together every morning, have them do a daily standup, the same way you would do with a team of humans.”
10:49
“Sometime around 2.30 in the morning, something opened up. Somebody had canceled. And Instinct was able to grab the reservation before anybody else.”
11:16
“As a human, all of a sudden, I went from managing tons of people to now I manage tons of people and tons of bots.”
Word for word.
Read-along
Don't just watch it. Burn it in.
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
metaphoranalogy
AI has completely transformed the way that I run my companies, and I'm quite literally using artificial intelligence in a number of new products to not only save costs, drive more revenue, become more productive, but frankly, I think that I can eventually run these companies and have almost no human employees, just have AI running so much of these businesses.
And before I explain... how I've set it up and what exact products I'm using, I think it's important to understand philosophically why I want to do this. Artificial intelligence is new.
It's constantly changing. The landscape is evolving on a day -to -day basis. New technology is coming out.
People are trying to figure out how to use it. But to me, a company is the single most productive unit of labor that we have in the economy. And so what I'm trying to figure out is how can I automate a lot of things?
Automation is not new. We've seen automation in blue collar work. We've seen automation in factories and warehouses.
We've even seen automation in white collar. jobs as well. But automation, when it comes to artificial intelligence, just got a massive leap forward because now you can have persistent, long jobs done by these AI bots.
Historically, people just used AI to just ask it questions. But now if I can tell a bot to go do something that could take hours to actually accomplish, that is a big breakthrough and I want to go and try to implement it. Now, my philosophy behind doing this is to not recreate the wheel.
I think there's a lot of people who simply say, I have AI now, so I need to go recreate what a company looks like. That's not how I approach this. Instead, what I do is I simply say, how do organizations operate today?
And then how can I use AI to do the exact same thing? I have specialized bots that take on different tasks. There's one for marketing.
one for finance, there's ones for operations, product, growth, engineering, and many others. And then what I do is I have these bots actually communicate with each other. I'll explain in a minute how I do that.
As I started to use AI, not only have I realized that a lot of these products have started to improve in terms of accuracy, but more importantly, is that they are getting jobs done with higher quality output than humans would do. I think a lot of people say, our company is made up of humans, and those humans have the tool of AI.
But I think that's the wrong way to think about it. The way I think about it is that artificial intelligence products, . The human's job is to give the AI as much context as possible.
So my job as the CEO of a company is to get as much context as I possibly can and keep feeding it into the AI. The more context that the AI has, the more helpful it's going to be to me. And then I look at the other half of my job is to simply keep asking the AI a bunch of questions.
Sometimes I need to brainstorm with the AI. Other times I need to ask it how other people have done things. Or I could even ask the AI to go and actually do research on something and come back given the context of our business and tell me what it thinks that I should do next.
Now, there's a lot of people that they're arrogant. They think that they're smarter than artificial intelligence. They think that AI is just a simple tool.
But the reason why I want our businesses to have AI as the brain is because it is super human intelligence. It should be smarter than us. And it's really good at dealing with complexity.
while humans are not. And so naturally, as the technology improves, and you keep beating it context, of course, it is going to become more valuable to you over time. And eventually, I actually think AI can autonomously run many products, services or companies.
And so if we're headed in that direction, I want to be at the forefront, I want to make sure that we're using AI as the central hub for all information, and that all of our employees are able to actually interact with the AI in order to get the same amount of information and make very high quality decisions as me. Now in terms of how do you get the context into the AI, this is a very important hack that you need to make sure you do.
I recently signed up for Grokbot. And when I signed up, I created a chief of staff. And then I said to the chief of staff, I want you to ask me questions or interview me so I can give you the context that you need.
I started with, here's my life. Here's my businesses. Here's what each business does.
Here's the business model. Here are the names of the people and their roles at each one of these companies. I then went and I said, by the way, there's a bunch of people in my life who are not actually working at any one of these businesses I've told you about, but they're important for you to know as well.
We have somebody named Isaac. He's the videographer. We have another person who works on a completely separate business that I hadn't told the AI about.
But because there's crossover, it was important for the AI to understand who they were. And so the more context I continued to give to the AI, the better positioned it was. It knew my companies.
It knew what they did, what their business model was, who the key people were, what our goals were, and what metrics I cared about. And then I supercharged the context. I gave it access to my email, to Slack, and I gave it access to my calendar.
Now, all of a sudden, The AI is armed, not only with institutional knowledge, and I essentially onboarded it like a new employee, but then I gave it the persistent context of what's going on in real time because of Slack, email, calendar, et cetera. And then once I did that, I began looking at What are all of the different roles or tasks inside of the business?
Well, obviously we need a product designer. We need a product growth person. We need an engineering team.
We need to have a marketing person. We need to have somebody focus on operations. We need to have a researcher and we need to have a ton of other roles.
And so I began creating specialized bots inside of Grokbot. for each one of those tasks. Take something like a recruiter.
We need to hire someone into a specific role. That person has to be senior and they need very specific qualifications. I could have easily went and got a human to go do the recruiting, or I could have gotten on LinkedIn myself.
It would have taken a lot of time and frankly been expensive if I was spending my time or if I had hired somebody else to do it. Instead, I simply told the chief of staff, tell the recruiter to go create a list of the top 10 people that they think I should reach out to for the specific role. All of a sudden, a couple of seconds later, I got back the list of 10 people.
But it wasn't just a list of 10 names. It also came with their LinkedIn profile. It came with a description of how their experience matched this specific job and the tasks that they're going to be responsible for.
And then more importantly, is it explicitly called out? Was it high or low likelihood of this person moving and why? So somebody just took a new job six months ago.
It's very unlikely that they're going to take a new job. So that's a low likelihood. But if somebody has been at their job for five years and they've done the exact thing we need them to do here.
then it's likely that it's high probability you're going to be able to convince them to move. But if that person has worked at the same company for five years, but they're the founder, it's very low likelihood that they're going to move. All of that nuance usually gets lost in these static AI products.
But with something like Grokbot, with a specialized bot focused on recruiting, it was able to give me an exact list of who the top 10 people were. It stack ranked them based on the priority. And then it also gave me the context that was necessary to know whether it was high or low probability of these people moving.
And so guess what? I started to reach out to people and then I started to set up meetings. Within 24 hours, I had calls or meetings with a number of people on the list.
One of those people asked me, how did you find me? Rather than lie or say, oh, somebody recommended you. I just told him what happened.
I set up the bot. I gave him a specific prompt. They came back with your name.
They told me there was a high probability of you wanting to move. And I reached out to you. And guess what that candidate said?
They were impressed. They're like, that's one of the coolest use cases of AI that I've seen. Makes me want to work at your company even more.
And so this is ultimately how I think about using AI. Is that not only am I building a second brain for the company, and that's the central hub. That is where all the information and context lies.
But then by creating these specialized bots, I can actually run the company by automating so many of these different tasks. I can go and have that recruiter working. while at the exact same time, I can have the product designer analyzing our website, which I recently did.
When I analyzed the website, guess what happened? Not only did it look at the actual visuals and go page by page, including screenshots and describing everything from different design decisions that could improve the product to the copy and how we could increase conversion on the website, but then it also was able to look at the actual source code, look at the site map, and explain to me a number of different things that could help in terms of SEO, or maybe even things that we actually had built.
We had hit it on the website and it suggested that we pull it forward for the actual user to be able to consume and interact with. That level of detail done in minutes for essentially no cost. that is impossible for a human to do especially at the quality of the information and so now we have an entire engineering team they continue to create new mock -ups new landing pages and new parts of the website and then they send me the screenshots or the pdf overviews and i feed it to the bot and i say analyze this critique it tell us how we can improve it so now the team is starting to reorient itself and use the artificial intelligence to be smarter than us that's the beauty That's what it was built for.
And I think we are just at the precipice now of the AI actually being helpful enough to run a company. Now, I've given you a couple of examples today, but Grokbot's not the only product I use. I think Grokbot is incredibly valuable.
I love the fact that I can have the chief of staff and then the specialized bots underneath it. I like that the bots can talk to each other. And maybe here's one more little pro tip for you.
You actually can have all of the bots. hold a meeting together every morning, have them do a daily standup, the same way you would do with a team of humans. And in that daily standup, you can tell the bots, I want to know what you did yesterday.
What are your goals for today? What are you blocked on? What information do you need from any of the other bots?
You would be shocked as to having the quote unquote executive meeting of your bots. how much more valuable each bot will become, and how much more productive your automated team all of a sudden is in the market. And so GrokBot is a key component to what I do on a day -to -day basis.
I think you should go check it out. But I also use a number of other products. For example, I use Sylvia.
Sylvia is something I use to manage my personal finances. That's the product that we're building. Sylvia is something that allows me to go and connect all of my financial accounts, bank, brokerage, crypto, credit cards, etc.
I also upload all of my private investments, real estate, cars, collectibles, et cetera. And then I begin talking to Sylvia, trying to get information as to what I should do for taxes, mortgages, estate planning, my personal portfolio, et cetera. But it all comes with the superhuman intelligence of AI, plus the context of my personal portfolio.
The same way that I use Grokbot to do specialized work. If I go and I try to use ChatGPT or Google and ask, how do I get my tax rate down? It's going to give me a very generic answer.
It has no context to my personal finances. With Sylvia, it will go asset by asset in my portfolio and will actually allow me to know what to do with each individual stock, each piece of real estate, or any other asset that I own. But I don't just use Grokbok and Sylvia.
Instead, I also use a product called Instinct. Instinct is a brand new personal assistant AI. You communicate with it through iMessage.
And what I found it to be very good at is reading through your emails. prioritizing certain things, and then taking action on your behalf. It can go and get you dinner reservations.
It can go and book flights. It can find hotels. It can even go and buy things for you on the internet.
The beauty is by using something like this in terms of AI, it is persistently available. It's very good at its job. It does not give up.
And it's also very cheap. And so anytime you combine those things, it's obvious why people are flocking to these products. I saw someone online talking about the fact they wanted to go see the Odyssey.
They needed two tickets, but it was sold out. And so Instinct sat there and it constantly kept checking every couple of minutes. Sometime around 2 .30 in the morning, something opened up.
Somebody had canceled. And Instinct was able to grab the reservation before anybody else. The person woke up and the bot had been successful.
They were able to go see the Odyssey that night. That's the power of artificial intelligence. And so whether you're using GrokBot, to go and look at your metrics, your site, recruit different people, or maybe even look at the finances.
It's super powerful. You now have the most intelligent type of technology. It's persistently available and it's low cost.
That is every business owner's dream. Silvia allows me to take that same idea and apply it to my personal portfolio. And Instinct allows me to go and use the exact same approach, but do it for my personal life, for things that are maybe non -financial, but I still need to get done.
It's weird. As a human, all of a sudden, I went from managing tons of people to now I manage tons of people and tons of bots. But I think that over time, bots are going to become way more valuable as people learn how to control them, how to manage them, and how to unleash their productivity.
I am just starting on my journey of figuring this out, but already I have Grokbot, Sylvia, and Instinct all working to better optimize my life. It's able to do things with my time, with my money, and with my output. And my guess is, as I get better at using these technologies, as I learn more, maybe even from some of you, not only am I going to have more bots in the future, but actually my companies are going to become more profitable.
They're going to be more productive. And my guess is that more and more people in the future are going to be able to figure out how to use this technology to make a living for themselves, capture financial security. and ultimately have more free time so they can spend it on the important things in life like spending time with their family and friends doing things that they enjoy like their hobbies and not just slaving away behind a computer that's what technology is for so let's get back to being human let's let the technology do the automated work and let us use our brains and simply spend our time enjoying life that's it for today please remember to subscribe to the channel.
It's my one ask for you. It's like saying thank you to me for this video. 90 % of people who watch videos on this channel are not subscribed, but we're going to change that.
So go hit the subscribe button and then let me know in the comments, what do you enjoy about these videos? What do you not? How are you using these technologies inside of your business?
Try to teach me something. I read every comment and it'd be really helpful for me if you told me something that I could actually implement in my business. I'll see you guys next time.
The Hook
The bait, then the rug-pull.
Anthony Pompliano opens with the boldest possible claim: he thinks he can eventually run his companies with almost no human employees. What follows isn't hype, it's the actual setup: a chief-of-staff bot, a roster of specialized department bots underneath it, and a growing list of AI products handling his money and his calendar.
Named ideas worth stealing.
02:20model
Chief of Staff + Specialized Bots
- One chief-of-staff bot holds full business context
- Specialized bots per function: marketing, finance, ops, product, growth, engineering, recruiting
- Bots communicate with each other, not just with the human
- Bots run a daily standup: what got done, today's goal, what's blocking them
Pompliano's org chart rebuilt as AI agents, mirroring how a real company already divides work.
Steal fora solo founder's ops stack
How they asked for the click.
VERBAL ASK
12:26subscribe
“Please remember to subscribe to the channel. It's my one ask for you... go hit the subscribe button and then let me know in the comments, what do you enjoy about these videos?”
Direct, low-pressure ask framed as reciprocity ('like saying thank you'), paired with a request for engagement (comment how you use AI) rather than a bare subscribe plea.
MENTIONED ON CAMERA
FROM THE DESCRIPTION
PRIMARY CTAWhere the creator wants you to go next.
OTHER LINKSAlso linked in the description.
Visual structure at a glance.
