February 3, 2026. I posted a link. No thread. No elaborate hook. Just a URL to an article documenting my complete AI storybook video workflow: every tool, every prompt template, every production decision from three months of trial and error.

Here's what happened:

31,929 impressions. 462 bookmarks. 2,334 engagements. 299 likes. 50 reposts. 38 direct follows. 7.31% engagement rate.

One post. One piece of documentation. And it generated 32.4% of every bookmark I earned across 1,789 posts in an entire year.

If you're an AI creator trying to grow, that number should change how you think about content. It changed how I think about everything.

Let me give you the full picture.

I'd been making AI storybook videos for three months using a stack of tools: Adobe Firefly for image generation, Nano Banana Pro for advanced prompting, Google Veo for video, ElevenLabs for narration, Suno for music. The process was complex and I'd documented every step as I went.

The article itself was comprehensive. Complete prompt templates. Tool-by-tool breakdown. Production decision trees. The kind of thing I wished existed when I started.

But the post on X was just a link. No image carousel. No hook thread. Just the URL with its preview card showing the article header.

I expected moderate engagement. Maybe some saves from the AI art community. Instead it became the highest-performing post of my entire year by almost every measure.

31,929 impressions made it my highest-reach original post ever. 462 bookmarks was 7.7x higher than my second-best post. 38 direct follows from a single piece of content was more than most weeks produced combined. And 1,756 detail expands meant people were stopping to look closely at the preview card before deciding to save it.

The engagement rate of 7.31% doesn't sound dramatic on its own. But at 31,929 impressions, that ER represents a massive volume of real interactions: 2,334 engagements from people who saw the post, stopped, and did something with it. Most posts at that impression level see their ER collapse because they're reaching cold audiences. This one held.

Here's where the data gets interesting for AI creators specifically.

Across my 101-day active period, I tracked the statistical correlation between every metric and daily follower growth. Bookmarks came in at r=0.698. That's a strong relationship. Stronger than likes (r=0.504). Stronger than reposts (r=0.473). Even stronger than profile visits (r=0.687).

When people bookmark your content, two things happen. First, the algorithm reads it as a deep value signal and serves your content to more people who tend to save things. Second, and this is the part most creators miss, bookmarks represent future intent. A like says "I appreciated this." A bookmark says "I'll need this again."

And that distinction is exactly why documentation content dominates bookmark counts for AI creators.

When someone scrolls past a beautiful AI image on X, they see the finished product. If it's good enough, they like it. Maybe they save it. But the experience is complete. The image answers its own question. There's no open loop.

When someone scrolls past a link to a documented AI workflow, they see the promise of knowledge. The preview card says: "This person figured out how to make AI storybook videos and wrote down the entire process." That's not a moment of appreciation. That's a future resource. That's something they might need someday when they try it themselves.

Every AI creator has a version of this article sitting undocumented in their workflow. The prompt sequences you've refined over months. The tool combinations nobody else uses. The workarounds you discovered by accident. The failures that taught you which settings actually matter.

That documentation is worth more than your best image. My data proves it.

The numbers are almost absurd when you zoom out.

Over my entire year of tracking, I published 1,789 posts. Original content, replies, threads, shoutouts, everything. Across all 1,789 of those posts, the total bookmark count was 1,426.

The article accounted for 462 of them. One post, 32.4% of all bookmarks for the year.

For follows, the article generated 38 direct follows from the post itself. That's 22.5% of all post-attributed follows across 1,789 posts. One piece of documentation drove nearly a quarter of my total follow attribution.

But the documentation story gets bigger when you include the Stor-AI Time franchise (the storybook videos the article documented how to make). Across 49 Stor-AI posts, I earned 191 bookmarks. Add the article and its companion posts and you get 653 bookmarks from documentation-related content.

45.8% of every bookmark I earned in a year came from either the documentation itself or the creative output that documentation described.

If you're an AI creator posting finished images every day and wondering why your account isn't growing, this is the answer. The images are the product. The documentation is the growth engine. People follow the person who can teach them something, not the person who shows them something pretty.

I know this because I have the counter-example in my own data.

August 28, 2025. I posted a single image that hit 35,561 impressions. Higher raw reach than the article. It generated 29 bookmarks and +23 net followers.

Then nothing. No sustained growth. No momentum. No cascade of saves and shares over the following days. By the next week, my account was back to its dormant average of 0.9 follows per day.

Same account, roughly similar impressions, completely different outcomes. The August spike was a beautiful image that came and went. The February article was a documented process that compounded.

The bookmark count tells the story. The August image: 29 bookmarks from 35,561 impressions. The February article: 462 bookmarks from 31,929 impressions. That's 0.8 bookmarks per 1,000 impressions versus 14.5 bookmarks per 1,000 impressions. The documentation generated bookmarks at 18x the rate of a standalone image with comparable reach.

For AI creators specifically, this is the fundamental distinction. Your generated images prove you're talented. Your documented process proves you're useful. Talent earns likes. Usefulness earns bookmarks. And bookmarks predict growth at r=0.698.

The article posted February 3. On February 4, Adobe officially accepted me into the Firefly Ambassador program.

I didn't plan the timing. But the combination was explosive.

February 4, 2026. All-time records across every metric:

107 new followers in a single day. Previous best before that: 52.

But look at the bookmark multiple. 25.9x my active daily average. Impressions only went up 4.2x. Follows went up 6.3x. But bookmarks went up nearly 26x.

What happened: the Ambassador announcement gave people a reason to explore my entire account. They found the article (already gaining traction from the day before). They found the Stor-AI Time series. They found the prompt engineering threads. And they saved everything.

This only worked because the body of work existed. The Ambassador title was the credibility signal. The documentation was the substance behind it. Without the article, without the documented workflows and educational threads, the announcement would have been a one-day spike and nothing more.

The viral moment wasn't one day. It was five.

186 followers in five days. 80,001 impressions. 612 bookmarks. That's 11.9% of my entire active period growth (1,565 total) concentrated into 4.7% of the days.

Notice what kept the momentum alive after the Feb 4 spike. Feb 6, I launched "The Mighty Monster Afang," a Welsh folktale adapted as an AI storybook video. It hit 10,426 impressions with 60 bookmarks and 9 direct follows. The article had documented how I make these videos. The video proved I wasn't just talking about process. I was actively producing with it.

For AI creators, this is the flywheel: document your process, publish the documentation, then release work that demonstrates the process in action. The documentation makes the creative work more meaningful, and the creative work validates the documentation.

Most AI creators I follow post images. Good images. Sometimes extraordinary images. Photorealistic scenes, surrealist compositions, hyperdetailed portraits. The prompting skill is clearly there.

But the images are self-contained experiences. You see it, you appreciate it, you scroll on. Nothing pulls you toward the profile. Nothing makes you think "I need to follow this person because they'll teach me something."

My data says the single most valuable thing an AI creator can do is document and publish their process. Not as a quick caption under an image. As a standalone, comprehensive, reference-quality resource.

Here's why the math works:

Bookmarks are the strongest growth signal. Across my 101-day active period, bookmarks correlated with daily follower growth at r=0.698. That's a strong statistical relationship. When people save your content, the algorithm reads it as deep value and serves you to more people who save things. AI creators who generate bookmarks grow faster than creators who generate likes.

Documentation generates bookmarks at 18x the rate of images. My article earned 14.5 bookmarks per 1,000 impressions. A standalone image with comparable reach earned 0.8. Nothing else comes close. And this wasn't an anomaly. My top 10 bookmarked posts of the year were dominated by process documentation and Stor-AI launches, not standalone images or clever threads.

The engagement rate holds at scale. Most posts see their ER collapse as impressions increase because they're reaching cold audiences. My article maintained 7.31% at 31,929 impressions. For comparison, posts in the 10K+ impression tier averaged just 6.1% ER across my dataset. Documentation content keeps people engaged even when it reaches far beyond your existing audience because the value proposition is clear and universal.

Process posts have a shelf life that image posts don't. A beautiful AI image gets engagement for maybe 48 hours. A documented workflow gets saved and resurfaces every time someone considers trying that technique. My article was still generating profile visits and bookmarks days after posting because people were finding it through others' shares.

Here's something AI creators don't appreciate enough: you already have the raw material for documentation that other creators don't.

Traditional artists work in physical media where process documentation requires separate photography, time-lapse recording, or detailed journaling. The process is fundamentally separate from the output.

AI creators work in text. Your process is your prompts. Your iterations are saved in your generation history. Your tool settings are quantifiable. Your failures are screenshots. Every single step of your creative process already exists in a format that can be documented and shared.

You're sitting on a library of reference material and posting only the cover art.

The creator who documents "Here's my complete prompt sequence for generating photorealistic surrealist scenes, including the 14 variations I tested and why 3 of them work" will outgrow the creator who posts "Made in Firefly" under a beautiful image. Every time. My year of data proves it.

If you create AI art and want to grow, here's what I'd build based on a year of tracking every number:

Pick your best technique and document it completely. Not a tip. Not a thread. A comprehensive article or guide. Include the prompts (real ones, not sanitized versions), the tool settings, the failures, the iterations. Make it something someone could follow step by step and get a result. Publish it on a platform that generates a link preview card with a strong visual.

Post the link, not the thread. My article was just a URL. The preview card did the work. Don't try to recreate the article as a thread. The article format signals "this is a reference resource." A thread signals "this is a conversation." Resources get bookmarked at 18x the rate. Conversations get liked.

Follow the documentation with the output. The day after my article, the Ambassador announcement hit. Three days after, the Stor-AI video launched. Each follow-up post reinforced the article's credibility. Don't let your documentation sit alone. Show it in action.

Track bookmarks, not likes. If your post gets 500 likes and 2 bookmarks, you entertained people. If your post gets 50 likes and 50 bookmarks, you created a resource. Bookmarks correlate with growth at r=0.698. Likes correlate at r=0.504. The gap is real and it compounds over time.

462 bookmarks from a single post. 31,929 impressions at 7.31% engagement. 32.4% of my entire year's bookmark total. 22.5% of all post-attributed follows.

The work was three months of documenting my AI storybook production process. Every prompt. Every tool. Every decision. The kind of documentation that most AI creators do privately and never share.

I shared it. And one post did what 1,788 other posts couldn't do alone.

If you're an AI creator, your workflow documentation is the most valuable content you're not making. The prompts in your notes app. The settings you've memorized. The techniques you've refined through hundreds of generations. That knowledge is what people will save, reference, and follow you for.

The beautiful images get you noticed. The documented process gets you remembered.

Glenn is an Adobe Firefly Ambassador and AI creator documenting the craft of prompt engineering and creative process at @GlennHasABeard. He publishes The Render newsletter and creates the Stor-AI Time series adapting world folktales through AI-generated video.

This article is part of a series analyzing one full year of X analytics: 365 days of account data and 1,789 individual post records. Every number is from official X analytics exports.