AI에게 인스타그램 운영을 시키면 팔로워를 몇 명 모을까? — Transcript
Full transcript
- 0:00How many followers can an AI-operated
- 0:01Instagram account gain? I’m going to
- 0:03try creating automated Instagram card
- 0:05news using AI myself. Starting today,
- 0:07I’ll be running an Instagram account
- 0:09for exactly 10 days using card news
- 0:11made by AI. Everything from selecting
- 0:12the topic to producing the card news
- 0:14will be handled by the AI itself. I
- 0:15will only be responsible for uploading.
- 0:17When you go on Instagram these days,
- 0:18the most viewed content is Reels first,
- 0:20followed by card news. Card news is
- 0:22often so informative and interesting
- 0:24that my fingers just naturally tap on
- 0:26it. You can’t help but click, and
- 0:28it’s really convenient to just swipe
- 0:30through, so I wanted to try making it
- 0:32myself this time. But card news is,
- 0:35after all, content. Since individual
- 0:38ideas are so important for content, it
- 0:39might be hard to gain thousands or tens
- 0:41of thousands of followers during this
- 0:43two-week test period. So, I want to
- 0:45show that this is possible. I’ll
- 0:48demonstrate the process of running an
- 0:50Instagram account with AI, so watch it,
- 0:52and if you think it looks good, try it
- 0:53yourself with your own content and
- 0:55ideas. After a brief planning phase, I
- 0:58entrusted the creation of the card news
- 1:00system to Claude Code. I only
- 1:01communicated my initial intent and
- 1:03discussed the implementation details
- 1:04with Claude Code as we built it.
- 1:06However, I hit a major obstacle right
- 1:08from the start. Actually, making the
- 1:10card news itself isn't that difficult.
- 1:12If I ask Claude Code to do it, it gets
- 1:14it done. But the AI can't decide what
- 1:16kind of account I should run, right? So
- 1:18I didn't automate this part and planned
- 1:20it myself. But honestly, this process
- 1:22was extremely difficult. To tell you
- 1:24the truth, this is already my fourth
- 1:26time filming this opening. The original
- 1:28start date for this test was last
- 1:30Monday, seven days ago. But the process
- 1:32of coming up with a topic and
- 1:33translating that into a system was so
- 1:35hard that I had to scrap it a few times
- 1:36. I had to find something that the AI
- 1:39could easily produce but that people
- 1:41would actually want to see, and after
- 1:43struggling here and there, it finally
- 1:45stabilized today. So, to give you the
- 1:47conclusion, the topic I chose is an
- 1:49international issues magazine. So, to
- 1:52give you the conclusion, the topic I
- 1:54chose is...introducing strange or
- 1:57interesting news and events from abroad
- 1:59. You've seen a lot of that on
- 2:00Instagram lately, right? I chose this
- 2:02topic because, while it varies from
- 2:04person to person, most people on
- 2:06Instagram seem to prefer entertainment
- 2:08or light content over learning or
- 2:10educational material. So, I thought it
- 2:13would be great to create snackable
- 2:14content, and since the world is vast,
- 2:16there’s no shortage of material,
- 2:18right? How many bizarre things happen
- 2:20around the world every single day?
- 2:21That’s why I decided on this,
- 2:23thinking it would be easier for AI to
- 2:25create. Based on this content plan, an
- 2:27AI card news automation bot was created
- 2:29. The system I built is as follows.
- 2:31First, this bot retrieves news daily
- 2:34from various overseas issue sites via
- 2:36RSS. RSS is a service that allows you
- 2:38to access content without visiting the
- 2:40site itself. Then, it filters the news
- 2:43to pick good candidates and sends them
- 2:45to me via Telegram. I review them and
- 2:47select the content I want to post. Once
- 2:50I press the confirm button in Telegram,
- 2:52the AI starts producing the card news.
- 2:54I created the design templates when
- 2:55building the bot, so it just needs to
- 2:57slot the content into that structure.
- 2:58Once the card news is made, all I have
- 3:00to do is post it on Instagram. Actually
- 3:03, automatic uploading is possible as
- 3:04well. I found out that if you go to the
- 3:07official Meta for Developers site and
- 3:09get an Instagram API key, you can do it
- 3:11. So, in the early version I built last
- 3:14week, I actually made it so it would
- 3:16upload automatically. But doing it that
- 3:18way meant the posts didn't have any
- 3:20music. These days, there are almost no
- 3:22Instagram posts without music, right?
- 3:24So, I decided to handle the actual
- 3:26posting manually myself. After
- 3:27confirming that the system was working
- 3:29properly, I immediately received the
- 3:30first draft of the card news. How is it
- 3:32? Doesn't it look decent? This design.
- 3:34I mean, the layout and the composition
- 3:36of the elements. That was my
- 3:37instruction, and the AI was the one
- 3:39that polished it to look nice. Creating
- 3:41this was truly difficult. The stocks
- 3:44and coins I’ve dealt with until now
- 3:45are in the realm of mechanical judgment
- 3:47, so building those wasn't that hard.
- 3:49But this is in the realm of creativity,
- 3:51and I have to make it something that
- 3:53people will actually like. Ah, that
- 3:55wasn't easy at all. Next, I created an
- 3:57Instagram account to post this content.
- 3:59This is the account I made. I named it
- 4:02"Beshu Magazine.""Be" from Young-bae, "
- 4:04shu" from issue—Beshu Magazine. But
- 4:06honestly, I don't think the name is
- 4:08that important when running an
- 4:10Instagram account. You can tell just by
- 4:12looking at my YouTube channel name,
- 4:13right? It took me quite a while just to
- 4:14come up with the account topic, so if I
- 4:16spent time worrying about the name, it
- 4:18would have taken too long, so I just
- 4:19came up with it quickly. Since I made
- 4:21the account and prepared the content, I
- 4:23started uploading from the very next
- 4:24morning. At 9 AM, I received card news
- 4:27candidates via Telegram and created all
- 4:29the ones I liked. After that, I just
- 4:31uploaded them to Instagram at the times
- 4:33I wanted. They say there’s an ice
- 4:34cream in Japan that doesn't melt. Isn't
- 4:36that fascinating? I posted a few of
- 4:37those and checked the next day to see
- 4:39if the views or followers had increased
- 4:41. Zero. Zero views, zero followers—
- 4:43not a single increase. Of course, since
- 4:46it had only been a day, I could accept
- 4:48having no views or followers. But I
- 4:50couldn't just sit still, could I?
- 4:52Today's conclusion: If you have AI run
- 4:54an Instagram card news account, you get
- 4:55zero followers in two weeks. Good job.
- 4:58I can't just leave it at that, can I?
- 5:00I’ve never managed an Instagram
- 5:01account professionally, so I did some
- 5:03research on how to get posts seen. As
- 5:05it turns out, there was a reason why
- 5:07the card news views were at zero.
- 5:08Instagram content types have different
- 5:10ways of being exposed, and Reels are
- 5:12usually shown to non-followers. On the
- 5:15other hand, card news is mainly shown
- 5:16to existing followers. So, no matter
- 5:18how many card news posts I upload,
- 5:19they’re being shown to zero people,
- 5:21which means the account can't grow.
- 5:23Therefore, starting today, I decided to
- 5:24try making them as Reels to aim for
- 5:26better exposure. I’m taking the card
- 5:27news I already created, resizing it for
- 5:29Reels, and turning it into a video that
- 5:31slides through the content. So now,
- 5:33when I press the "create card news"
- 5:34button, it reports back to me with both
- 5:36the card news and a Reel version. I
- 5:38tried uploading the card news as a Reel
- 5:40as well. It definitely performed better
- 5:42in terms of views when uploaded as a
- 5:43Reel compared to a regular post. It’s
- 5:45Day 4. I’ve been uploading about
- 5:47three to four card news posts a day,
- 5:49and since I started adding Reels, the
- 5:51views have increased a little. But I
- 5:53still have zero followers. Why isn't
- 5:56anyone following me? Is it boring? So I
- 5:58told Claude Code about our account's
- 6:00current situation and asked it to
- 6:02search the internet for ways to improve
- 6:04it. It suggested a few things, like
- 6:06changing the profile bio, shortening
- 6:07the first screen of the Reels, and
- 6:09encouraging shares and saves. I applied
- 6:11those changes right away. But after
- 6:13running the account for a while, I felt
- 6:14like I wanted to try a different topic
- 6:16as well. So, starting today, I’m
- 6:17planning to create an additional
- 6:19account. Since most of the work is
- 6:21already automated, it didn't seem like
- 6:22managing one more would be too much of
- 6:24a burden. The topic I chose this time
- 6:26is an account that introduces today’s
- 6:28news. While the Bashu Magazine used to
- 6:30cover fun international issues and news
- 6:32, this project has a slightly more
- 6:34serious tone. I don't usually get many
- 6:36opportunities to come across news, so I
- 6:38think this will be helpful for me too.
- 6:40So, I need to decide on an account name
- 6:42, but what should I name it? Give me
- 6:44lots of news. Since it's "a lot," how
- 6:46about "Infinite Refill"? Then, in the
- 6:49sense that there is a lot of news here,
- 6:51I’ll name it "News Here" so people
- 6:53will check it out. Pretty clever, right
- 6:55? After finishing the simple planning,
- 6:56I sent the content to Claude Code and
- 6:58was able to build another card news
- 7:00system right away. And this time, to
- 7:01automate the workflow a bit more, I set
- 7:03it up so that it reports the content to
- 7:05be uploaded via Telegram, and then
- 7:06automatically uploads it once I press
- 7:08the approval button. I also became
- 7:10curious about whether it makes a big
- 7:11difference when there is no music. I
- 7:12received the card news in a new format,
- 7:14and the design was better than I
- 7:16expected. It was truly fascinating that
- 7:18this level of result could be created
- 7:19so quickly. That made me suddenly
- 7:21wonder how long it would take to create
- 7:23this manually, and what the difference
- 7:24in aesthetic sense is compared to a
- 7:26human designer. So, I visited an
- 7:28acquaintance who works as a designer
- 7:29and had them evaluate the design
- 7:31elements of the card news I created.
- 7:33I mean, it doesn't really stand out as
- 7:34obvious.
- 7:35Oh, can you tell if a person or AI made
- 7:37it? How about the composition and
- 7:38layout?
- 7:39It seems really neat and well-made.
- 7:42Then, if you were to make something
- 7:43like that in Photoshop, how long would
- 7:44it take to make about three or four
- 7:46slides?
- 7:46Well, this isn't actually that
- 7:48difficult. I guess you could say
- 7:50there's no variation in the design?
- 7:52Well.
- 7:53I mean, you just change the text and
- 7:56images, but I think thinking of the
- 7:58ideas is the hardest part. Thinking of
- 8:00the topic.
- 8:00Ah, thinking of the topic. Then, if the
- 8:02topic and content are already decided,
- 8:04how long would it take to start making
- 8:06this once those are set?
- 8:07Without a template, you could probably
- 8:09finish this in about 30 minutes, right?
- 8:12That takes less time than I thought.
- 8:13Since there’s nothing too difficult
- 8:15and no flashy designs. Then, how long
- 8:17does it take for AI to make this?
- 8:18I just click once to start it, and it
- 8:20actually finishes it in about 2 minutes
- 8:22.
- 8:23It's definitely much faster. I think
- 8:25designers need to know how to use AI.
- 8:26Designers these days.
- 8:28Oh, designers these days use AI a lot
- 8:29too.
- 8:30Ask them. Can everyone use it? Do they
- 8:32have the skills? Things like that.
- 8:33Oh. Then, in a way, it did create
- 8:35something aesthetically pleasing.
- 8:37Right? Exactly.
- 8:38Yeah. Oh, but the design is actually
- 8:39better than I thought? Ah, there’s so
- 8:41much of this stuff lately, so it turns
- 8:43out this is all done by AI.
- 8:44I don't know if it's actually done with
- 8:45AI. Even if I did it with AI, I did it.
- 8:47So it's not obvious, right?
- 8:49If you hadn't told me, I would have
- 8:51thought it was run by a reporter or a
- 8:52news outlet, at least a little bit.
- 8:55You don't have to do anything.
- 8:56I don't do anything. So that was the
- 8:59first channel. Aesthetically, what do
- 9:01you think of this channel? The second
- 9:02channel is—
- 9:03Oh, this is set up so well, like it's
- 9:05run by a real news organization. Oh, it
- 9:08makes me want to pause like this.
- 9:10This one. So, if you were to make this
- 9:12from scratch in Photoshop, assuming you
- 9:14had no templates, how long would it
- 9:16take?
- 9:16Well, since it's just pasting text onto
- 9:18an image, it doesn't take long. If I
- 9:20decide on what to include and what the
- 9:22captions will be, about 5 minutes.
- 9:24Oh.
- 9:24Since it's one page.
- 9:255 minutes. I don't think it would even
- 9:27take 5 minutes.
- 9:27That.
- 9:28This "News Here" one.
- 9:31I think you're pretty good at making
- 9:32things like this. It's kind of funny to
- 9:34me. But this definitely looks much more
- 9:36professional. Hmm, is that why it feels
- 9:37like a more trustworthy channel, or one
- 9:39you'd want to watch more?
- 9:41I mean, the previous one felt like "is
- 9:43this real or fake," but with this one,
- 9:45you're convinced that "oh, this must be
- 9:46real."
- 9:48Do you pull the photos from somewhere?
- 9:49Or does it make them itself?
- 9:50I guess it makes them itself.
- 9:51The photos are AI-generated images.
- 9:54Oh, but it makes them really well.
- 9:55Image generation technology has gotten
- 9:57so good these days.
- 9:58I mean, if we had to find these
- 9:59materials ourselves, it would all cost
- 10:01money. You’d either have to pay for
- 10:03them or make them yourself, but it's
- 10:05great that this just pops up with a
- 10:07click. Anyway, to give an overall
- 10:09review, the AI-made card news is
- 10:11well-crafted.
- 10:12Ah, well-made. No problems at all.
- 10:14Then, a day later, I checked the status
- 10:16of the "News Here" channel, and
- 10:17something amazing happened. One of the
- 10:19reels I made with card news hit the
- 10:21algorithm and recorded about 25,000
- 10:23views in just one day. Thanks to that,
- 10:25the account started getting exposure
- 10:26and the followers increased little by
- 10:28little. On the other hand, for the "
- 10:29Baeshu Magazine," it's still stuck at
- 10:31an average of 100 views and 0 followers
- 10:33. It was clear that there's a
- 10:34difference in account growth speed
- 10:36depending on the content topic and card
- 10:38news design. It’s finally been a week
- 10:40since I started the test. The result of
- 10:42doing it myself: Baeshu Magazine has 0
- 10:43followers and around 100 average views.
- 10:45News Here reached about 10 followers in
- 10:474 days, with an 80,000-view reel that
- 10:49caught the algorithm, and it frequently
- 10:51got over 1,000 views. News Here is
- 10:53definitely doing really well. You can
- 10:56tell because spam comments have finally
- 10:57started appearing on my account. You've
- 10:59seen them a lot on Instagram, right?
- 11:00For me, things like "Song-ari" or "Meal
- 11:02Tank" are popping up. Of course, I’m
- 11:03not interested in "Milk Tank Girl" Song
- 11:05A-ri, and I’ll delete all these
- 11:06comments, but it felt a bit rewarding
- 11:08in a different way, knowing that my
- 11:09account is getting some exposure.
- 11:12I’ve heard that for an Instagram
- 11:13account to grow, you need to follow
- 11:15other accounts, leave comments, and put
- 11:17a lot of effort into promoting your
- 11:18channel, but I haven't done any of that
- 11:20for "News Here." I just worked hard on
- 11:23posting feed posts and Reels. But
- 11:24considering that, isn’t 1,000
- 11:26followers and 80,000 views a pretty
- 11:27decent result? On the other hand, "
- 11:29Baeshu Magazine" isn't performing well
- 11:31at all. The AI is starting to struggle
- 11:32with gathering content, and the view
- 11:34counts aren't great either. So, it’s
- 11:35unfortunate, but I think it’s time to
- 11:37say goodbye to "Baeshu Magazine."
- 11:38It’s a shame. Since I stopped running
- 11:40"Baeshu Magazine," I needed a new
- 11:42channel to fill that void. I decided to
- 11:44run an account that I’ve always
- 11:45wanted to try. The theme of the new
- 11:47channel will be AI news. I’m really
- 11:49just making this because I want to see
- 11:50it, and you know I have a lot of
- 11:52interest in trying out various things
- 11:53with AI, right? But if you look at
- 11:55Instagram, there are so many accounts
- 11:57that introduce the latest AI features
- 11:58or how to use them. So, I’m going to
- 12:00try doing that too. I’ll go through
- 12:02domestic and international sites to
- 12:04collect news, organize it into card
- 12:06news, and post it. What should I name
- 12:07the channel? Ah, this is honestly the
- 12:09hardest part. Since it's an AI account,
- 12:12"AI A-a-AI Unbelievable"? Should I go
- 12:15with "AI Unbelievable"? That’s not
- 12:16bad, is it? That’s how I succeeded in
- 12:18creating my third Instagram account.
- 12:20And this time, I tried something new as
- 12:21well. While watching my videos, haven't
- 12:23you ever wondered what it would be like
- 12:25to see the AI do everything fully
- 12:26automatically? "Baeshu Magazine," which
- 12:28I made first, was manual. Because the
- 12:31AI would create the content, but I had
- 12:32to manually upload both the Reels and
- 12:34the feed posts, and "News Here" was
- 12:36semi-automatic. The feed posts were
- 12:37uploaded automatically, but I had to
- 12:38upload the Reels myself. This time,
- 12:39I’m going to try making it fully
- 12:41automatic. When content arrives via
- 12:43Telegram, I just press the publish
- 12:44button, and both the feed posts and
- 12:45Reels are uploaded automatically. The
- 12:47reason I couldn't upload Reels
- 12:48automatically until now was because of
- 12:50the background music. To add music, you
- 12:52have to upload it manually. So this
- 12:54time, I secured copyright-free music in
- 12:56advance and set it up so the AI
- 12:57overlays that music onto the video and
- 12:59uploads it automatically. With that,
- 13:01the system for "AI Unbelievable" is
- 13:03complete, and I’ll now be running two
- 13:05channels: "News Here" and "AI
- 13:06Unbelievable." Running the newly
- 13:09created system the next day, I found
- 13:10the "AI Unbelievable" method extremely
- 13:12convenient. First, it collects news
- 13:14from overseas and sends me card news
- 13:16candidates, then I just press the
- 13:18publish button under the card news I
- 13:20want to create. Then it generates the
- 13:22final card news and reels, and
- 13:23automatically posts them to Instagram.
- 13:25It’s definitely much better than the
- 13:27existing system since it requires less
- 13:28manual work. So, I switched "News Here
- 13:30Is" to a fully automated system as well
- 13:32. Besides, I had a great start this
- 13:33time by gaining followers in just one
- 13:35day. It’s finally been 9 days since I
- 13:37started the test. On the 9th day, "News
- 13:39Here Is" has 14 followers. The highest
- 13:41view count is 102,000. I could
- 13:43occasionally see view counts of 3,000,
- 13:452,000, and 1,000 as well. And the "AI
- 13:46Nonsense" account has one follower and
- 13:49has been maintaining views around 100.
- 13:51The "AI Nonsense" one is really
- 13:52convenient and nice, but it doesn't
- 13:54really get that many views. Naturally,
- 13:57it's tough since the accounts used for
- 13:58this test don't post video reels and
- 14:00only compete with card news. But
- 14:03looking at how "News Here Is" is doing
- 14:05quite well, there clearly seems to be a
- 14:06big difference in content. So, I
- 14:08decided to find an expert today to
- 14:10diagnose the status of these channels.
- 14:12It's my friend who has about 60,000
- 14:14Instagram followers. I showed them "
- 14:16News Here Is" and "AI Nonsense" and
- 14:18asked for ways to make card news
- 14:19accounts successful. It’s been
- 14:22running for about a week, 35 posts, and
- 14:2414 followers.
- 14:25Oh, the view counts are actually coming
- 14:27in quite well. The views are high, but
- 14:28the conversion rate to followers seems
- 14:30too low compared to that. 35 posts in a
- 14:33week is actually an absurd number, and
- 14:35I'm even curious how you did it, but
- 14:38the first problem I see right when I
- 14:40enter is that the profile isn't set up
- 14:42at all. Looking at it now, it says "
- 14:45Meet the latest news daily." It says
- 14:46this, and if it's only like this, it's
- 14:48obvious people won't follow. You should
- 14:50write a total of four lines here; maybe
- 14:52the first line is the "News Here Is"
- 14:54slogan, and the second is about your
- 14:56track record, like, does "News Here Is"
- 14:57have any social proof? It probably has
- 15:00nothing since AI does it automatically.
- 15:02Oh, then put "domestic news" on the
- 15:04second line.
- 15:05Then on the third line, write about the
- 15:07value you can provide, and on the
- 15:10fourth line, maybe "Follow us to see
- 15:12five news pieces every day." If you set
- 15:14it up like that, the follower count
- 15:15would have grown much better. Honestly,
- 15:16if it hit 100,000 views, shouldn't you
- 15:18have gathered at least 1,000 followers?
- 15:20There's even 6,000, this makes no sense
- 15:21, really.
- 15:22But for one week, isn't this doing
- 15:23pretty well? It’s honestly absurd, it
- 15:25feels like this program has perfect
- 15:27timing.
- 15:28Wow, this could actually work. How many
- 15:30minutes a day did this take you?
- 15:32I spent about two or three hours
- 15:33setting it up initially, and now it
- 15:35uploads automatically with just a
- 15:36button click.
- 15:37Wow, then it's like I can just leave it
- 15:39running.
- 15:40So, that was the hopeful part. Now, we
- 15:42head to the hopeless part. The second
- 15:44channel is "AI Makes No Sense," which
- 15:45is an AI-specialized channel.
- 15:48Yeah. Oh, it’s already starting.
- 15:50I see there’s a lot more text here.
- 15:52The readability isn’t that great. So,
- 15:54the visibility is low.
- 15:54Yeah, a bit low. But I think it’s
- 15:57great how the images change in the
- 15:58background to keep the video flowing. "
- 16:00AI Makes No Sense." Also, looking at
- 16:03these, the topics aren't that
- 16:05interesting. But can you actually have
- 16:08AI handle all of this when you're
- 16:09setting it up?
- 16:10If you tell it what kind of news and
- 16:12what tone you want for the post, it
- 16:13filters everything exactly like that.
- 16:15Wow, that’s insane. I feel like it
- 16:17would be hard to make this even if I
- 16:18wanted to. Are you not selling this?
- 16:20I’d actually like to buy this. I
- 16:22don't sell them. Then, how do you
- 16:24actually grow an Instagram account?
- 16:26It’s tough with just a card news
- 16:28account.
- 16:28Oh, but it is possible with a card news
- 16:30account, too. It’s possible, but you
- 16:32have to post Reels along with them.
- 16:35Right now, it just looks like a Reel,
- 16:36but it’s actually card news. Vrew is
- 16:39really good these days, so you go in
- 16:41there, make it into a real video, add a
- 16:44voiceover, and create genuine content.
- 16:47If you don't just post card news, but
- 16:49also Reels, Stories, and set up your
- 16:51profile properly, wouldn't the account
- 16:52grow as a complete package? The
- 16:55mechanism itself seems really good,
- 16:57seeing how it uploads automatically.
- 16:59Oh, a follow-up question: how do you
- 17:01reach 60,000 followers on Instagram?
- 17:04You have to do those three things I
- 17:06mentioned every day—one Reel, one
- 17:08card news, and four Stories—and keep
- 17:11that routine going. And a popular tip
- 17:13these days is this. Take a format
- 17:15that’s already gone viral overseas
- 17:17and upload it while keeping that format
- 17:19but tailored to your own topic.
- 17:21Anyway, the goal of this test is to
- 17:22gather followers using only card news.
- 17:25Wow, this is surprisingly insane, and
- 17:27the fact that they created this fully
- 17:29automated system in under 10 minutes a
- 17:30day is so tempting; I really think this
- 17:32is worth trying, for example, if you're
- 17:34not starting from zero but already have
- 17:36a 60k follower account. If you equip
- 17:38this, it’s bound to grow, right? For
- 17:40those already doing it,
- 17:41I think people already doing this would
- 17:43need it—it automatically creates and
- 17:45publishes card news, so wouldn't it be
- 17:47pretty impressive, not just cheating
- 17:48the system? Oh, it’s finally the last
- 17:50day of the test. Now the system is
- 17:52fully automated, so I can just press a
- 17:54button to upload posts and short-form
- 17:56videos. Performance varied by content,
- 17:58but in the case of the news-type
- 17:59account, it reached 134,000 cumulative
- 18:01views, and post impressions and views
- 18:03were gradually increasing. So, I’ve
- 18:05left my Instagram card news account
- 18:07operation to AI for 10 days. I didn't
- 18:09get involved at all in the research or
- 18:11the production of the card news. I only
- 18:13performed a final review to prevent
- 18:15duplicate or inappropriate posts from
- 18:16going up. So, the final result for the
- 18:19news account was 42 posts and 15
- 18:21followers in one week. The total view
- 18:24count was 145,000. And the first
- 18:26account I operated, the issue magazine,
- 18:28ended up with 0 followers and an
- 18:30average of 110 to 120 views. And the
- 18:33last one I made, the AI-themed account,
- 18:35recorded one follower and about 1,900
- 18:37total views in four days. The average
- 18:39view count is around 100. As you can
- 18:41see while watching this video, I don't
- 18:43think it's easy to run an Instagram
- 18:45account using only card news. Since
- 18:47card news itself isn't easily exposed
- 18:49to non-followers, I think you need to
- 18:51make it in a Reel format like I do or
- 18:53create other additional content to see
- 18:55any growth. Nevertheless, there’s a
- 18:57reason why I gritted my teeth and bet
- 18:59everything on card news for this video.
- 19:02I’ll work hard to research how to
- 19:03automate short-form videos like Reels
- 19:05or Shorts and bring that to you as well
- 19:07. Anyway, if there’s one thing I felt
- 19:09after this test, it’s that automating
- 19:11card news is extremely convenient. And
- 19:13the results come out exactly as much as
- 19:14the effort you put in. If I had
- 19:15actually appeared in the videos or
- 19:17posts to explain things, spent more
- 19:19time researching, and made the card
- 19:21news look better, my Instagram account
- 19:23would have probably grown faster. As
- 19:25expected, nothing comes easy. But as I
- 19:28mentioned earlier, content is all about
- 19:30how you approach it, so if there’s
- 19:31content you’ve always wanted to try
- 19:33and it’s suitable for automation, I
- 19:34think it would be a good idea to try
- 19:36automating it like this. So, today's
- 19:38conclusion is that while it’s very
- 19:40convenient to let AI run your Instagram
- 19:42card news account, you only get as much
- 19:43quality as you put in effort.
About this transcript
This page contains the full transcript of AI에게 인스타그램 운영을 시키면 팔로워를 몇 명 모을까? by 이영배, generated from the public captions YouTube serves with the video. The transcript has 4,264 words across 648 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.
What you can do with it
Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.
Free YouTube transcript tool
YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.