How I Turned Claude Into My Personal Assistant (Complete System)

CChase AI
Computing/SoftwareAdvertising/MarketingSmall Business/StartupsInternet Technology

Transcript

00:00:00Claude should be running your life, or at least all the boring parts.
00:00:04And it's actually very easy to turn Claude into your own personal assistant,
00:00:07saving five to 10 hours every single week, just like I have.
00:00:10Now, we've talked about creating your own Claude OS on this channel before,
00:00:13whether that's a web-based setup like this one that includes voice
00:00:17or an Obsidian-based command center.
00:00:20But in all those videos, we tended to focus really on system design,
00:00:24how we should create our skill architecture,
00:00:26and how we should set up the actual Obsidian Vault.
00:00:29But today's video is going to be a little different.
00:00:31This one is going to look at my exact personal assistant setup,
00:00:35the skills and automations I use to run my business, my channel, my research, all of it.
00:00:41So you can get a good idea of what you need to do to recreate yourself.
00:00:45Now, when it comes to these custom fancy dashboards and command centers like you see here,
00:00:49we will get to that at the very end because there is some value in creating these
00:00:52when it comes to personal assistant setup.
00:00:54But the first thing we need to establish is,
00:00:56what the heck is this personal assistant even going to be doing?
00:00:59What is the boring, redundant work that we can automate and turn into skills?
00:01:04Well, for me, there's really three different buckets that I operate in that has a lot of like
00:01:08sort of drudgery, right?
00:01:10That takes up an inordinate amount of time, doesn't really require a lot of decision making,
00:01:14but it's just like grunt work that I offload to Claude Code.
00:01:18And those three buckets are productivity and sales.
00:01:21I kind of put that in the same category.
00:01:23Research and then content.
00:01:24So the productivity and sales stuff has a lot to do with my AI agency business.
00:01:28And then the research and content kind of fall in sort of the social media sphere.
00:01:32Makes sense.
00:01:33Now, the nice thing about this is a lot of what you're going to see here,
00:01:37I think does translate to almost everyone outside of the content stuff.
00:01:42I know a lot of people aren't creating content, but chances are you are working in some sort of
00:01:46sales or marketing environment where you have information coming in from clients or customers
00:01:51and you need to respond to them and you need to be able to do some sort of daily research
00:01:55to stay on top of whatever it is you're staying on top of.
00:01:57So a lot of what you see here, while it's specific to me, is something that can translate to you.
00:02:02Now, what we will look at here is probably what's going to give you the most immediate value.
00:02:05And this is sort of the productivity section.
00:02:06And it all spawns from this, the email triage automation.
00:02:11Now, the email triage skill does the following.
00:02:13Every morning at 8 a.m., it takes a look at my Gmail inbox.
00:02:16It uses the Claude connector to do so.
00:02:20It takes a look at everything that's come in over the last 24 hours and it puts it into a series
00:02:24of buckets, either leads, urgent, warm, sponsors, meetings, or noise.
00:02:30Once it has all the emails broken out into those specific buckets, it then does different
00:02:34things depending on the bucket, specifically the leads bucket and the sponsors bucket.
00:02:41Now, if it's a sponsor, Claude then automatically drafts up a reply with a boilerplate response
00:02:46and a link to my media kit saying, hey, here's what the prices are X, Y, and Z.
00:02:50However, if it's a potential lead, we then begin a process that is a little bit more in-depth.
00:02:54So on my website, every lead fills out a form saying essentially their budget, what they're
00:02:57looking for, their timeline, all the standard stuff.
00:03:00All this information is picked up by Claude Code in this inbox triage.
00:03:04From there, what it does, if it thinks it's a real query, so the email makes sense, they
00:03:09actually spent time on their response, the budget aligns, and it wasn't someone just putting in
00:03:13random nonsense, Claude then kicks off a web search.
00:03:17So just some background on who their company is, takes a look at their email, basically tries to
00:03:22get me just some preliminary information so I know what I'm doing if I walk into a potential
00:03:27discovery call with them.
00:03:28From there, Claude then gives me a recommendation saying, here's what I found out, here's what
00:03:32they want to do, should you move forward, should you not?
00:03:34I'm still the arbiter here of whether we go forward or not, obviously.
00:03:37If I say, yeah, let's do it, it then creates a draft email with a response as well as a link
00:03:43to my calendar so they can set up a meeting for me.
00:03:46If not, it just skips them.
00:03:47For all the emails in other buckets, if it's urgent, it's going to give me a potential drafted
00:03:51reply.
00:03:52But for everything else, it's just kind of marking them and what it also does is it gives
00:03:56me a full report, a written markdown file that goes into my obsidian vault that acts as a
00:04:00summary that I can read in a glance.
00:04:02The report looks something like this and then also lists all the noise that I don't really
00:04:05care about.
00:04:05And so with the simple setup, I save a ton of time every single day because I don't have
00:04:08to read all my emails.
00:04:09I don't have to manually filter every single lead.
00:04:12And if it thinks the lead makes sense, it's already done some preliminary research and drafted
00:04:16the response for me.
00:04:17And this is something you can very easily adapt for your own purposes.
00:04:20What sort of emails do you normally get?
00:04:22How can you put them in different buckets?
00:04:24And then more importantly, what do you want Cloud Code to do with the information it has
00:04:29once it's sorted everything?
00:04:30Now, the other half of the equation that saves me a ton of time is proposals and follow-ups.
00:04:33And this is what I do with clients after we've already had the discovery call.
00:04:37So I use Calendly.
00:04:38So post-discovery call, what I get in my inbox is a recap.
00:04:43Calendly automatically gives me a summary that includes all the key points of everything we've
00:04:47discussed.
00:04:48That recap then gets sent to Claude Code.
00:04:52Claude Code takes that recap and generates a branded PDF breaking down.
00:04:57Here's essentially what we're going to create.
00:05:00Here's the scope of work.
00:05:01Here's the pricing.
00:05:02Here's the form for you to sign.
00:05:04Let's go ahead and move forward.
00:05:05Now, obviously, I'm going to take a look at all that, but that saves a ton of time versus
00:05:09having to manually create some sort of proposal.
00:05:12From there, it generates a drive link that is sent to the client and the draft that includes
00:05:16the link itself.
00:05:17Now, this is what the proposal looks like.
00:05:19And this is just a dummy version.
00:05:20So I don't show any like client information on here, but breaks down the engagement, shows
00:05:24the timeline, and then shows the investment.
00:05:27And then at the very bottom, it just has some room for signatures.
00:05:30Now, lastly, in regards to follow-ups, that's a little more specific to some of like the AI
00:05:34audits I run.
00:05:35So I'm not going to go deep into that here since it's kind of technical and really won't
00:05:38apply to 99.9% of you.
00:05:40But that, in a nutshell, is kind of the productivity sales side of my Claude code personal assistant.
00:05:47Again, it doesn't have to touch every single thing you do.
00:05:49But like you saw here, a lot of this is just boring stuff that otherwise I would be spending
00:05:54hours a day on.
00:05:55So what are you spending hours a day on or hours a week on that we can offload to Claude?
00:06:00You should be doing it.
00:06:01Now, before we move into the research portion of my personal assistant, a quick word from
00:06:07today's sponsor, me.
00:06:09So I just released my Claude code masterclass, which is the number one way to go from zero
00:06:12to AI dev, especially if you don't come from a technical background.
00:06:15We focus on real use cases like the personal assistant stuff we're talking about today.
00:06:20And I go over other AI tools.
00:06:23I also have a codex masterclass in this community as well.
00:06:26So if you want to get serious about learning AI, this is the place to be.
00:06:30I'll put a link to it down below.
00:06:32And everything you see here, my exact setups, the Java system, the Obsidian Command Center,
00:06:36can also be found inside this community.
00:06:39So now let's talk about research.
00:06:40Now, obviously, mine is going to be all AI focused and yours may or may not be.
00:06:44So what you need to keep in mind is the idea that all of this spawns from knowing where your
00:06:50information originates.
00:06:52In the AI space, that's really a handful of places.
00:06:55It's like Twitter, GitHub, and then sometimes things will merge on YouTube first.
00:07:00But that's kind of it.
00:07:01Those are kind of the big three.
00:07:02So that's where we're going to go for information.
00:07:04Where you need to find your information for your niche, you got to figure it out.
00:07:08I can't tell you that.
00:07:09So you got to figure that out.
00:07:11But once you know that sort of source of knowledge, the setup is exactly the same for how I do
00:07:15it.
00:07:15So first things first, my main automation is my daily brief, right?
00:07:21My daily brief goes out to those wells, those sources of knowledge and figures out what the
00:07:27heck is going on.
00:07:27And so for me, like I said, that's really X, that's YouTube, and that's GitHub.
00:07:32Now, the GitHub setup is simple enough for me.
00:07:35Here's one from last week.
00:07:36And what it does is it uses the GitHub API to find the top trending AI GitHub.
00:07:41And it breaks it out in a few categories.
00:07:42I look at like the top 10 trending for the week.
00:07:44So what was created this week?
00:07:46I look at the top five trending from those that were created within the last 30 days.
00:07:50And then I look at the fastest growing over the last 24 hours and the last 30 days.
00:07:57So what's new and growing and then what's kind of been around and it's kind of like shooting
00:08:00up the charts.
00:08:01This gives you like a good idea of like, okay, what are people actually playing around with
00:08:04these days in the AI space?
00:08:05Because this changes literally daily and it's impossible to like know about it unless you
00:08:10actually looking at GitHub.
00:08:11And then we have Twitter, YouTube, and also a general web search.
00:08:14So I'm able to pull the big headlines.
00:08:16I can see what's trending on YouTube.
00:08:18Importantly, it shows me the creator and the views and their subs.
00:08:21So that kind of gives you a good idea.
00:08:22Like, I don't care if someone with a million subs gets 10,000 views.
00:08:24But if some dude with 2,000 subs has 10,000 views, then clearly he's talking about something
00:08:29that people care about.
00:08:30And then also we look at Twitter.
00:08:33And beyond that, it does some basic analysis on like, hey, here's some content opportunities.
00:08:38But to be totally honest, when it comes to figuring out, hey, here's what the research shows and
00:08:43here's what we should actually talk about, AI is pretty hit or miss with that.
00:08:46What we really care about here is sort of just the raw data it's able to bring.
00:08:49And again, what's the big sell here?
00:08:50The big sell is I don't have to do all that myself.
00:08:52Am I still manually going to go on Twitter and YouTube and go through my subscriptions for,
00:08:57you know, five, 10 minutes a day?
00:08:58Sure, of course.
00:08:59But this stops that from being like a 30-minute process.
00:09:03It saves me time to time, especially on the GitHub side.
00:09:05To do that manually would be terrible.
00:09:07End of note, this research daily brief and this email triage brief we spent a bunch of
00:09:12time on, these are actually combined into a single automation that I just call like my
00:09:16morning automation.
00:09:18And this is set up as a routine inside of Claude code.
00:09:22So it's a local routine, runs on my computer as long as it's open and just fires off those
00:09:27automations automatically.
00:09:28Now, as for the rest of the research side, I have what I call my xPulse or this gets kind
00:09:32of a pulse on Twitter, as well as my YouTube dives, which are on demand and deep research,
00:09:36which is on demand.
00:09:37Now, the Twitter setup is something that is actually an app that I built that lives on
00:09:43Railway.
00:09:44So this lives on Railway.
00:09:46It's working 24-7.
00:09:47And what it does is it sends me messages on Telegram every hour or so, showing me the top
00:09:54tweet in the AI space that's trending.
00:09:56And this is also tied to a lot of the major creators, especially ones who are on like the
00:10:00Claude dev team and the OpenAI team.
00:10:02So if something is popping off, something big has opened up in the AI space, I know about
00:10:07it immediately.
00:10:08I don't have to sit there on Twitter refreshing all the time.
00:10:11So the first half, the daily brief and the xPulse are, you know, those are automated.
00:10:17These two are automated.
00:10:18I don't even have to think about them.
00:10:19These two, YouTube dives and deep research, these are specific on-demand skills.
00:10:23I can use whenever I find some topic that I care about and I want to learn more about.
00:10:28Now, deep research is super basic.
00:10:30And by basic, I mean, it's already a command set up inside of everybody's WOD code.
00:10:35So if you do forward slash deep research, that is a essentially preloaded dynamic workflow.
00:10:40It will spawn, if you just let it go nuts, up to like 100 sub-agents that do like adversarial,
00:10:46essentially, information gathering.
00:10:48Like they'll go on the web, they'll find information, then they'll test it against each other to see
00:10:51if it's right and do some synthesis versus your standard web search where you just send Claude
00:10:55code.
00:10:55If you say, hey, Claude, go find me information about, you know, Fable 5, it's just going to
00:10:59send a couple of sub-agents to browse the web.
00:11:01This is that on steroids, but it eats up a ton of tokens, but it's something like a lot of
00:11:05people don't actually use.
00:11:06I also have what I call my YT pipeline skill.
00:11:11And so what that does is it then searches YouTube to find videos that are relevant to whatever
00:11:19I'm asking about.
00:11:19It then sends all those URLs to Notebook LM.
00:11:23Notebook LM itself then takes all the transcripts, does all the synthesis and sends that to me.
00:11:29Now I do that using a specific CLI, which is the Notebook LM Pi CLI, which actually gives
00:11:34Claude Code essentially an unofficial API to Notebook LM.
00:11:37Again, being able to do that means I get all the power of Notebook LM, but it doesn't cost
00:11:41me tokens to have Notebook LM and Google servers do all that synthesis on YouTube transcripts,
00:11:46which again, what are we doing here?
00:11:48We're trying to save time.
00:11:49I don't want to watch 10 YouTube videos about a specific topic.
00:11:52Just give me the summary.
00:11:53Do the synthesis.
00:11:54Tell me what's the throughput between all 10 videos.
00:11:57Tell me where they aren't aligned.
00:11:58And then tell me why I should care.
00:12:00And you put all that together and essentially what you get is a research setup that saves
00:12:06me undoubtedly five to 10 hours by itself minimum from this.
00:12:11If I try to do all this research on my own, I'd be dying.
00:12:13You know, I would be seriously struggling.
00:12:15And I think for a lot of people, again, whether you're doing AI-based searches or not, I think
00:12:20the productivity side and this research side should really go hand in hand.
00:12:24There's a ton of places I imagine in your domain where the sort of emails you respond to, the
00:12:30sort of things you're creating are based on certain research.
00:12:33So why aren't we sort of combining these things?
00:12:35Again, especially if these are like lower level, you don't really have to make decisions
00:12:38type tasks and just hand it over to a personal system of cloud code and have it do that for
00:12:42you automatically.
00:12:44Now, the last bucket here is content.
00:12:45And I'll spend the least amount of time here because I think this is the least relevant
00:12:48for most people.
00:12:49But just for those of you who might be doing content and want to know how I do it, for me
00:12:54specifically, when it comes to scripting hooks and outlines, this is simply a single on-demand
00:13:00skill.
00:13:01Now, I really like Callaway, super good creator on YouTube.
00:13:05So I essentially trained Claude on a bunch of his videos to create those skills in terms
00:13:10of here's how you should do a hook, here's how you should think about transitions and that
00:13:13sort of thing.
00:13:13But to be honest, I'm not someone who really does scripts.
00:13:16So this is just more of a brainstorming tool for me, but it is useful.
00:13:20I don't rely on AI to come up with the words that I say, but oftentimes it's nice to have
00:13:25a back and forth where you have some semblance of an idea, you bring it to AI, it's been trained
00:13:30on certain skills, so it's not just giving you generic garbage.
00:13:33And through that back and forth, some ideas, you know, kind of coalesce for you, the actual
00:13:37human being.
00:13:38So in terms of hooks and outlines, that's how I approach it.
00:13:41I know some people have everything scripted, just not how I work.
00:13:44Then we have packaging, again, similar to my hooks and outlines.
00:13:47I don't have it just create the titles or create the thumbnails, the thumbnails, I pretty
00:13:50much do all on my own.
00:13:52And then in terms of the titles, yes, it uses the same sort of like Callaway things where
00:13:56it's like, hey, here's the sort of words that you use, the sort of like psychological triggers,
00:13:59however you want to say it.
00:14:00But at the end of the day, it's more of a back and forth.
00:14:02I have it look at what titles I've done well and why.
00:14:06And it just, again, if it's just you and a piece of paper and you're trying to come up
00:14:09with everything 100%, I think you're going to struggle.
00:14:12And I think you also struggle if you just try to offload all creative purpose to AI because
00:14:16it's just not that good at these things that require quote unquote taste.
00:14:20But I think some sort of mixed approach helps a lot.
00:14:23But lastly, where AI really comes in handy for me is content repurposing.
00:14:28Now, this goes beyond just like dropping a video in a folder and it puts it across a bunch
00:14:32of different platforms.
00:14:33What this does is every day, twice a day at noon and 8 p.m., it looks to see if a new
00:14:38YouTube video has been posted by me.
00:14:40From there, it runs this content cascade skill where it then fetches the transcript from that
00:14:45video.
00:14:46And then it rewrites a blog, a LinkedIn post, and a tweet based on that transcript.
00:14:55Now, that skill took some time to get together because it's all about getting your voice correct.
00:15:00And your voice is your voice.
00:15:02It's unique.
00:15:03No one can tell you how to create a skill that's going to like work best for that.
00:15:06My suggestion, if you want to do something like this and have it do these sort of things,
00:15:10is you need to have examples of your own writing that you've done 100%.
00:15:13And what you do is you have a consistent back and forth.
00:15:15And it's kind of like a cycle, right?
00:15:17So I give Claude examples of my writing.
00:15:21I then say, turn that into a skill.
00:15:24Okay.
00:15:24It's then going to run that skill and it's going to give you its example.
00:15:29And what are you going to do?
00:15:30You're going to destroy their example.
00:15:32You're going to absolutely eviscerate it and say, this is wrong.
00:15:34This is wrong.
00:15:35This is wrong.
00:15:35This is right.
00:15:36Here's how we'll change this.
00:15:38So then you give them the updated example.
00:15:40It updates the skill.
00:15:41It gives you a new example.
00:15:43You destroy it.
00:15:44You do this over and over again for like 10 times.
00:15:47And then you continue to do it over and over again when you do it live.
00:15:50And over time, you will eventually get a skill that works for you.
00:15:55That's really the only way to do it.
00:15:56There's a huge human in the loop component if you're doing any sort of like humanizing or
00:16:01in your voice type writing.
00:16:03So don't listen to anybody that says, hey, here's just like this.
00:16:05Here's this one shot skill that's self-improving.
00:16:07No, no, no.
00:16:08You have to be in the loop.
00:16:09And that's the way to do it.
00:16:10And so in this case, I did it for a blog because it's a little different voice for a blog versus
00:16:15a LinkedIn post versus a tweet.
00:16:17And to be honest, I'm very lazy in terms of actually posting these tweets.
00:16:20But LinkedIn and the blog post, totally viable.
00:16:24And again, it comes from a single YouTube video.
00:16:26I don't have to think about it.
00:16:27It's automatically posting these things, putting it in my voice because otherwise I just wouldn't
00:16:31do it.
00:16:32So that is the meat and potatoes of my personal assistant setup with Claude, the automations
00:16:36and skills that I actually use.
00:16:38Now, what I hope you got from that is that this is unique.
00:16:42These are the things that I've identified that I don't really want to spend time on so I
00:16:46can spend my time doing things that I consider higher leverage.
00:16:49I don't want to do emails.
00:16:50I don't want to do a ton of research in the morning if I don't have to.
00:16:53And I don't want to repurpose my content.
00:16:55I don't want to have to manually write every proposal or filter every lead.
00:16:59Only you can identify what those sort of issues are for you.
00:17:02But the point is, once you identify them, you simply turn them into a skill.
00:17:05You make sure they work.
00:17:06You automate them.
00:17:08And you do that a few times.
00:17:09And all of a sudden, like you are saving five to 10 hours a week with that simple setup.
00:17:15Now, to turn that into a more advanced, coherent system, however, requires some sort of structure.
00:17:20Cool.
00:17:20We're creating all these briefs.
00:17:21It's doing stuff with emails.
00:17:22It's creating content.
00:17:23Well, like, how do I give it like a brain?
00:17:26How do I make sure it's able to reference things it's done in the past?
00:17:29How can I take all these briefs?
00:17:30How can I put them on a single dashboard?
00:17:32That's where we start to get into like the Cloud OS videos and the Cloud OS type lessons.
00:17:37And in this last part of the video, we'll touch on that briefly because that is something
00:17:41we have done plenty of content on in the past.
00:17:43And I'll link that here as well.
00:17:45Now, this is where you hear everyone talking about Obsidian and a second brain with Cloud Code.
00:17:50Why do we care about Obsidian?
00:17:51Is Obsidian in itself necessarily changing the way Cloud Code works?
00:17:55The answer is not really, right?
00:17:58Even something like this is just sort of just a fancy visual knowledge graph.
00:18:02But what Obsidian can do is allow us to organize everything that's created in our personal assistant
00:18:09setup, which over time creates a map that we hand to Cloud Code so we can quickly and correctly, truthfully answer our questions.
00:18:19So how do we set up Obsidian in a way to do that?
00:18:21Well, luckily, it's pretty simple.
00:18:23The Carpathi method is the most common one you will see.
00:18:26I want you to imagine all of these things simply as folders.
00:18:30Okay.
00:18:30These are all just folders.
00:18:31At the top, we have the vault.
00:18:33This is the folder we have designated as the Obsidian vault.
00:18:36Underneath the vault as subfolders, we have multiple sections.
00:18:40We have the raw section.
00:18:41We have the wiki section.
00:18:42And we have the output section.
00:18:44We just need places to put information that makes sense.
00:18:47So raw is unstructured data.
00:18:49Like, hey, we had Cloud Code do a bunch of research on stuff.
00:18:52It hasn't done any synthesis.
00:18:53It's just dumping all that raw information here.
00:18:56Well, after it's, you know, gotten all this information, we probably want to turn it into some sort of report.
00:19:01Well, that goes into the wiki section.
00:19:04And, hey, we have that report.
00:19:05We want to turn this into a slide deck.
00:19:06Well, that goes into the output section.
00:19:09Simple enough.
00:19:09Now, underneath all those folders, you're going to have more subfolders.
00:19:13Because, hey, if I have a wiki folder, but I have a million different reports,
00:19:17how does Cloud Code know where it needs to go to find information?
00:19:21Well, underneath every folder, we're going to have an index document,
00:19:24what's essentially is like, you know, the table of contents for that subfolder that says,
00:19:28hey, here's where this is, here's where that is, et cetera, et cetera.
00:19:32So imagine we had, you know, essential reports about AI agents and RAG systems and content creation.
00:19:38And under those reports were even more detailed reports about specific things,
00:19:42like autonomous coding and tool use patterns.
00:19:44So in those subfolders, what would you have?
00:19:45You would have index.
00:19:46So you kind of get where I'm going with this.
00:19:49We have things broken out into different subfolders that make sense,
00:19:53depending on their use case.
00:19:54And in each subfolder, we have an index file that says,
00:19:59hey, here's everything in the subfolder.
00:20:00So that Cloud Code always knows where it's going.
00:20:03It can never get lost.
00:20:04Now, when I say get lost, is Cloud Code actually getting lost?
00:20:07No, but if it's not organized, what happens is we're going to increase the token cost, right?
00:20:14If we just rely on grep for everything, that isn't necessarily the most efficient way to do it.
00:20:18So if we don't have an efficient setup, it's going to increase the token cost.
00:20:22It's actually going to decrease the accuracy over time.
00:20:26As you get bigger and bigger file structures, and you have more and more things in there,
00:20:32you need some form of organization.
00:20:34This is essentially a filing cabinet for Cloud Code.
00:20:37And yeah, we get these cool knowledge graphs,
00:20:39but that sort of idea of the filing cabinet of the map is the real value play.
00:20:42And the second half of the equation is these sort of like command centers and dashboards.
00:20:45Like, cool, Cloud Code is my personal assistant.
00:20:47It's grabbing emails.
00:20:49It's doing research on contents, repurposing stuff.
00:20:51But like, where does that information go?
00:20:53Where does it live?
00:20:54Where can I see it?
00:20:55More importantly, can I see everything in one place?
00:20:58Yes, we can totally set that up.
00:21:00Now, can that place be the terminal?
00:21:03Not really.
00:21:05That's like not the purpose of the terminal.
00:21:07And it's not even really the purpose of the Cloud Code desktop app.
00:21:09So when we talk about these command centers,
00:21:12or we talk about some web app dashboard like this, that's the value, right?
00:21:15It's the one-stop shop that you really can't get anywhere else
00:21:18because it needs to be 100% customized.
00:21:20And you can really see that shine in the subsidiary base setup, right?
00:21:23I have all my metrics here, my token burn.
00:21:25I have all these different automations that we're talking about are here.
00:21:28I can essentially run them at a click of a button.
00:21:30The audience section here is essentially a different breakdown of that morning report.
00:21:35And the same sort of stuff lives here in the web app version as well.
00:21:38I got the metrics.
00:21:39I got the skills.
00:21:40Down here, I have the documents so I can pull up those reports.
00:21:43And this is the exact same thing I would find inside of my Obsidian Vault.
00:21:46Like it's all connected.
00:21:48So these dashboards give you the observability.
00:21:50Obsidian is what allows you to keep it all organized.
00:21:52And it's these sort of skills and automations that are actually saving you time
00:21:56and doing the work as a personal assistant, as an executive assistant.
00:22:00And so you put that all together and you have a way to make Cloud actually work for you.
00:22:04Essentially, totally hands-off.
00:22:06So that's where I'm going to leave you guys for today.
00:22:08I hope that was somewhat eye-opening in terms of how you should approach Cloud as a personal assistant.
00:22:13If you want to get your hands on the Cloud Code Masterclass or the skills I talked about today
00:22:16or any of these sort of dashboards, my exact setups can be found inside of Chase AI+.
00:22:20There is a link to that in the pinned comment.
00:22:22And as always, I'll see you around.

Key Takeaway

Offloading redundant, low-decision tasks like email triage, research summaries, and content repurposing to a Claude Code-based assistant saves five to 10 hours weekly.

Highlights

  • Email triage automation runs daily at 8 a.m., categorizing messages into leads, urgent, warm, sponsors, meetings, or noise.

  • Potential sales leads trigger automated web searches for background information, followed by AI-recommended actions and drafted meeting requests.

  • Content repurposing automations run at noon and 8 p.m., automatically converting new YouTube transcripts into blog posts, LinkedIn updates, and tweets.

  • A GitHub monitoring routine uses the API to identify top trending projects, fastest-growing repositories over 24 hours, and new 30-day releases.

  • Obsidian vaults organized with index files act as a searchable filing cabinet, reducing token costs and increasing retrieval accuracy for Claude Code.

  • An AI-built application hosted on Railway pushes trending AI news to Telegram hourly to eliminate manual social media monitoring.

Timeline

Email Triage and Productivity Automations

  • Automated email triage sorts messages into specific business buckets every morning.
  • Sales lead processing includes preliminary background research and calendar link generation.
  • Calendly recaps trigger the automated creation of branded PDF proposals with scope and pricing.

The system processes the Gmail inbox daily at 8 a.m. using Claude Code. For potential leads, the AI determines validity, performs a web search on the company, and drafts a response with a calendar link if the lead is promising. Post-discovery calls are handled by syncing Calendly summaries to generate structured proposals, significantly reducing manual administrative labor.

Automated Research and News Briefings

  • GitHub API integration tracks daily and monthly trending AI repositories.
  • A dedicated Railway-hosted app provides hourly AI trend updates via Telegram.
  • Notebook LM synthesis summarizes YouTube transcripts to avoid manual video viewing.

Daily briefings pull data from X, YouTube, and GitHub to identify actionable trends without manual browsing. For deep dives, an on-demand YouTube pipeline sends transcripts to Notebook LM via an unofficial CLI, allowing for rapid synthesis of multiple videos. Adversarial sub-agents in the 'deep research' command perform information gathering and verification to ensure accuracy.

Content Repurposing and Creation

  • Content cascades automatically generate blog posts, LinkedIn updates, and tweets from video transcripts.
  • Iterative refinement of AI skills using personal writing examples is necessary for maintaining a unique voice.
  • The system runs twice daily to monitor for new YouTube uploads.

Content creation utilizes a back-and-forth process to refine hooks and titles based on historical performance data. For repurposing, the system monitors for new video uploads and uses a specialized skill to rewrite content for different platforms. High-quality output requires a human-in-the-loop process where the AI's drafts are repeatedly corrected against provided style examples.

Knowledge Management and Dashboarding

  • Obsidian folders categorized into raw, wiki, and output sections structure the AI's knowledge base.
  • Index documents within subfolders allow Claude Code to navigate and retrieve information efficiently.
  • Centralized dashboards provide observability over token usage and trigger on-demand automations.

Obsidian serves as the structural 'brain' for the system, preventing the AI from getting lost in unstructured data. By using index files, the system maintains a searchable filing cabinet that keeps token costs low and accuracy high. Custom web-based dashboards provide a single interface to manage metrics, execute skills, and access reports generated by the various automations.

Community Posts

View all posts