This page is in English · Italiano
AI video

90-Day AI Creator Roadmap: What I'd Do From Zero (2026)

My 90-day AI creator roadmap: what to build in days 1-30, 31-60 and 61-90, and why I will not promise you a follower count at the end.

Some links in this article may be affiliate links. If you sign up through them, I may earn a small commission at no cost to you. I only recommend tools I personally use to make my AI videos.

TL;DR

  • I rewrote this 90-day AI creator roadmap without a follower target. No platform publishes a repeatable path to a specific follower count, so I am not going to sell you one.
  • Days 1-30: pick one niche and one format, freeze your tool stack, and ship on a schedule you can hold on a bad week.
  • Days 31-60: stop guessing. Test with non-followers using trial reels, then read watch-through and shares instead of likes.
  • Days 61-90: build the destination. A link, a capture, an offer. Attention with nowhere to go is not a business.
  • What should exist on day 90: a repeatable production system, one format that consistently outperforms your others, and one place people can go next.

Why I deleted the follower number from this roadmap

The working title for this article promised zero to 100K followers in 90 days. I cut it, and the reason is simple: I do not have a method that reliably produces 100K followers in 90 days, and I have not seen one that survives contact with a second account.

Here is the number that made me rewrite the whole thing. My most-viewed reel sits at 1.4M views. My follower count is nowhere near 1.4M. Views and followers are not the same currency, and a roadmap that treats them as one is lying to you before you even start.

There is a structural reason for that gap. Instagram’s own ranking documentation describes recommendations as driven by predicted actions on the individual piece of content, not by the size of the account behind it. That is good news if you are starting from zero, because a single reel can travel without an audience. It is bad news for follower-count roadmaps, because nothing about that system promises the viewer will tap follow.

So this roadmap targets what you can actually control in 90 days: a production system, a proven format, and a destination. Followers are a lagging output of those three. Treat them as the scoreboard, not the plan.

Days 1-30: build the machine, not the audience

Pick one niche and one format, then stop negotiating with yourself

Month one is not about reach. It is about removing decisions. Every time you sit down and ask “what should I post today,” you burn the energy that should have gone into shipping. Pick one niche and one repeatable format on day one and hold both for the full 30 days.

If you have no idea where to start, I keep a running list of AI reel niches that are working and why. Pick from it, do not invent a category nobody searches for.

Format matters as much as niche. A cinematic reel and a tutorial reel are different products with different jobs, and mixing them randomly for 30 days gives you 30 data points that tell you nothing. I wrote up how I decide between the two before I open any tool.

Freeze the stack on day 3

The most expensive mistake I see in my DMs is tool shopping disguised as research. Three days is enough to choose. After that, the stack is frozen for the month, even if something shinier launches on day 12.

My working stack for this phase is deliberately small. I use Higgsfield for cinematic motion and preset-driven camera work, and Runway (code BYMICHYDEV25) when I need finer control over a specific shot. That is two video tools, one editor, one voice tool. Anything beyond that in month one is procrastination with a subscription attached.

I catalogued the sequencing errors that kill month one in 10 creator AI stack mistakes. Most of them are not about picking the wrong tool. They are about picking the right tool at the wrong point in the pipeline.

Ship on a schedule you can hold on a bad week

Pick a cadence you can hit when you are sick, travelling, or out of ideas. Three reels a week that actually ship beat seven that exist only in your notes app. I run a fixed production schedule rather than a daily creative decision, and I laid out the exact version I use in my 30-day AI reel schedule.

By day 30 you should not be measuring growth. You should be able to answer one question: can I produce this format again next week without a heroic effort? If the answer is no, fix the process before you touch the strategy.

Days 31-60: read signals, not vanity metrics

Let trial reels do the guessing for you

This is the phase where most people start posting into the void and calling it consistency. Instagram gives you a way out of that. Trial reels let you share a reel with people who do not follow you first, and roughly 24 hours later you can see views, likes, comments and shares, plus how that trial compares to your previous trials, according to Instagram’s own documentation.

Instagram also notes you can push a trial to your followers manually if it performs, or have it shared automatically based on views in the first 72 hours, and that a professional account is required to use the feature (Instagram for Creators). For a month-two creator, that is a free A/B testing rig. Use it before you burn a good idea on a bad hook.

The three numbers I actually read

Likes are the least useful number on the screen. In month two I read three things instead: how far people get before dropping, how many people send it to someone else, and how many save it. Watch-through tells me whether the hook and the pacing hold. Shares and saves tell me the content did something for the viewer beyond entertaining them for four seconds.

One caveat worth knowing before you optimise blindly: Instagram publishes recommendation eligibility guidelines that describe content types it will not push to non-followers. Reading that page once is cheaper than spending month two wondering why nothing travels.

The day-45 cut

On day 45 I go back through everything and cut the formats that are not earning their production time. Not the ones with the fewest likes, the ones with the worst ratio of effort to shares. A reel that takes me four hours and gets the same traction as one that takes 40 minutes is a format I stop making, no matter how proud of it I am.

By day 60 you should have one format that consistently beats your others. Not a viral hit. A repeatable winner.

Days 61-90: give the attention somewhere to go

The destination comes before the offer

Month three is where most creator roadmaps skip straight to selling something. I do the opposite: build the path first, decide what sits at the end of it second. A viewer who wants more from you needs somewhere obvious to go, and the bio link plus a working reply flow covers that long before you have anything to sell.

The mechanics matter more than the copy here, because the messaging rules are strict and easy to break by accident. I documented the setup I actually run, including the platform limits, in my Instagram DM funnel guide.

What the platforms actually require before they pay you

If your plan is platform payouts, it helps to know the real thresholds instead of the ones in a thumbnail. On YouTube, the expanded Partner Program requires 500 subscribers with 3 valid public uploads in the last 90 days, plus either 3,000 qualified watch hours in the last 12 months or 3 million qualified Shorts views in the last 90 days (YouTube Help). The full program requires 1,000 subscribers with either 4,000 qualified watch hours in 12 months or 10 million qualified Shorts views in 90 days (YouTube Help).

Instagram is deliberately less specific. Its help centre states that monetization tools each carry their own product eligibility requirements covering accounts, behaviours and payments, and that creators need an authentic, established presence rather than a single published follower number (Instagram Help Centre; see also Subscriptions eligibility). If someone quotes you a hard Instagram follower threshold for monetization, ask them where they got it.

Read those numbers again and notice what they do not say. None of them describe 90 days. That is the honest answer to why this roadmap does not end at a payout.

The first money is usually not platform money

In practice, the first revenue most AI video creators see comes from doing the work for someone else, not from a payout dashboard. One good client covers more than a year of Shorts revenue at small scale, and you can land one at day 90 with a portfolio of three strong reels. If that is the direction you want, I broke down how I structure and price that work in my agency pricing guide.

The 90-day AI creator roadmap at a glance

Phase The job What you measure What exists at the end
Days 1-30 Remove decisions. One niche, one format, frozen stack. Did it ship? Nothing else. A production process you can repeat without heroics.
Days 31-60 Test with non-followers, cut what does not earn its time. Watch-through, shares, saves. One format that consistently beats your others.
Days 61-90 Build the destination, then decide the offer. Clicks to destination, replies, enquiries. Somewhere for attention to go, and a reason to go there.

FAQ

Can I actually reach 100K followers in 90 days?

I have never done it and I would not plan around it. A single reel can travel far without an audience, because recommendations are driven by predicted actions on the content rather than account size, but nothing in that system promises the viewer follows you. Plan for the system, treat followers as the scoreboard.

How many posts per week does this roadmap assume?

Whatever you can hold for 90 straight days. I would rather see three a week for 13 weeks than daily for two weeks and then silence. The cadence is a constraint you set once, not a decision you make every morning.

Should I be on Instagram or YouTube for this?

Pick one for the 90 days. They reward different things and splitting your attention in month one is how you end up with two weak accounts. The YouTube thresholds above are public and concrete, which makes YouTube easier to plan against if payouts are your goal.

What if my day-45 cut leaves me with nothing that works?

Then the problem is usually the niche, not the format. Go back to a category people actually search for, keep the production process you already built, and run another 30-day block. The machine is the asset, and it survives a niche change.

Do I need paid tools to run this?

Not on day one. You need one video tool you know well and an editor. I add tools when a specific shot needs something my current stack cannot do, not on a schedule.

Show Your Work

# SHOW YOUR WORK - 90-Day AI Creator Roadmap

DAY 1     Write one line: "I make [format] about [niche] for [person]."
          If you cannot finish it, you are not ready to post.
DAY 3     Freeze the stack. Two video tools max, one editor, one voice tool.
DAY 4-30  Ship the cadence. Do not change niche. Do not change format.
DAY 31    Turn on trial reels. Every new reel ships as a trial first.
DAY 45    The cut: rank formats by shares-per-hour-of-production.
          Kill the bottom half. Yes, including the one you love.
DAY 60    Name your winner format in one sentence. That is your product.
DAY 61    Build the destination: bio link + reply flow. Offer comes after.
DAY 90    Checkpoint - do you have (1) a repeatable process,
          (2) one format that wins, (3) somewhere for attention to go?
          Three yes = the next 90 days are about volume, not discovery.

Tools: Higgsfield -> higgsfield.ai/?ref=michydev
       Runway -> runwayml.com/pricing (code BYMICHYDEV25)

Where to go from here

This roadmap is the sequencing layer. It tells you what to work on and in what order, but it does not hand you the production system for days 1-30. That is what I built the Blueprint for: the exact workflow I use to take an AI reel from idea to published, without the month of tool shopping.

You can grab it here for $27: The AI Reel Blueprint. It is the shortest path I know from “I want to do this” to “I shipped something I would show a client.”

Written by . Last updated August 26, 2026.

Want video like this for your business?

Tell me about your project. I usually reply in under 4 hours.

Get in touch