Why I post AI content to a new Instagram, not my main

If you have been posting for a while, your account is not a blank slate. It carries a history: the kind of thing you post, the kind of person who followed you for it, and a set of signals the platform has learned from all of that. That history is useful when you post more of the same, and it works against you the moment you post something different.

That was exactly my situation. I made a new Instagram profile for AI content, and the reason is simple. I did not want to fill my own timeline with AI stuff going out. I think everyone who follows me are getting sick of my episodes anyway, so adding a second type of content on top of that was never going to go well.

My profile is very much set up of an audience who are not Underway specific. They did not follow me for AI videos. So if I put AI videos in front of them, the most likely result is that they scroll past, and those weak signals then travel back into how the account gets treated. You end up doing worse with the new content and worse with the old content at the same time.

This is the bit people miss. The question is not whether the content is good. It is whether the people being shown it are the people who want it.

The obvious option is to post everything to the main account. It is where your followers are, so it feels like a head start. In practice you are asking an audience assembled for one thing to carry a second thing, and you are asking the algorithm to re-learn what your account is about using the worst possible data: half-interested scrolls from people who followed you for something else.

The second option is to post to the main account and then repost to a new one. This is the one that sounds sensible and is not. Post it on the big account first and the same video on the second account can be counted as copied content. You have spent the fresh account's best chance on a duplicate.

The third option is to split the content by format rather than by account: Stories here, Reels there, a different caption style. That tidies your grid but it does not change who is being shown the video, which is the actual problem.

The fourth is to not post the new content at all until the main account is ready for it. That usually means never. The main account is never ready, because an audience does not change shape on its own.

In my head I'm like, I create a new one, a fresh one, and it's okay, you start from zero. Zero sounds like a loss. It is not. Zero means no conflicting history, no audience built on something else, and nothing for the algorithm to argue with.

Posting fresh content to a fresh account means the algorithm will find that pocket of people who are interested in this content. It has nothing else to go on, so it tests the video with small groups, watches what they do, and keeps moving towards the people who watch it. On a mixed account that signal is buried under years of other behaviour.

And the early numbers are not nothing. 254 viewers yesterday, which is pretty decent. That is more than some of mine are getting on an account with a real following. A fresh account with the right content can beat an established account with the wrong content, which is the whole point.

You have to start somewhere. The new profile has one photo on it and that is fine. Nobody is judging the grid on day one; the algorithm is judging the video.

This is the practical rule that follows from everything above. Post to the new, specific account first. If you want it on your main account too, it goes there afterwards, and ideally with its own words rather than an identical copy.

The reason is order. You could say post to mine and then post here, but then it would be counted as copied content. The platform's view of which post is the original is largely a matter of which one appeared first. Give the original to the account that needs the distribution most, which is the new one.

I apply the same order across platforms generally: the new specific account gets the first post, and anywhere else gets the same clip with its own words written for that site. Same video, different wrapper, different order of priority.

One: decide whether the new content is genuinely off-theme for your existing audience. If your followers came for something unrelated, it is. If it is an extension of what you already post, you may not need a second account at all.

Two: make the new profile and accept that it starts with one photo and no followers. Do not spend a week on branding. The content is the thing being tested, not the bio.

Three: post the new content to the new account first, consistently, and let the algorithm run its tests. It needs repeated tries to find the pocket of people who want it.

Four: if you want it on your main account, post it there after, with different words, and treat it as a secondary placement rather than the main distribution.

Five: judge it on the right number. Not followers, which take months, but viewers per post. Two hundred and fifty-four people who chose to watch an AI video is a better signal than a few thousand of your existing followers scrolling past one.

The setup itself is minutes. A profile, a photo, a name. I had the new Instagram made and posting before I had anything else on it.

The ongoing cost is the real one: you are now feeding two accounts. That is only sustainable if the content for the new account is cheap to produce. In my case it is AI-made video, which is why the split works at all. If every post for the new account were a full shoot, I would not keep it up.

There is also a patience cost. A fresh account has no head start, so the first weeks are quiet by definition. You are trading immediate reach for the right reach, and that trade only pays off if you keep posting long enough for the algorithm to find the audience.

Let me put on record how good one of these is. I didn't even give it this idea. It just created it, and it's such a good idea. The video is about a man whose dad created a bakery. It's been running for thirty years. There's just no documentation around it. There's no recollection. There's nothing. So he's taking over, and he's gonna put that right today.

That captures the exact feeling that I have around it. And it applies across so many business sectors. A shop, a studio, a band, an app: the same thing happens everywhere. Nobody films the business before it works, and then the record is gone.

You can kind of tell that that's AI. But it's getting better. I envisage that kind of thing being done by real people at some point. That is the direction this goes: AI-made scenes now to make the point, real people telling their own story later.

This is also why I built Underway. It reads the tools you work in, asks you about your day out loud, and hands back a finished numbered episode, video or podcast, with no editing. The bakery with no documentation is the problem it exists to stop happening again.

A separate account suits you if your new content is clearly a different thing from what your audience followed you for, and if you can produce it often enough to feed a second feed. It also suits anyone testing a new direction who does not want to risk the account they already have.

It does not suit you if the content is really the same thing in a new coat. Splitting for the sake of tidiness just halves your posting and gives the algorithm two weak accounts instead of one strong one.

It also does not suit you if you cannot keep both going. A new account that goes quiet after two weeks has cost you the time and taught you nothing. Be honest about whether the new content is cheap enough for you to keep making it.