Teaching a tool that changes every month is brutal: the workflow you nailed last week is old by the time you write it up. A post structure is the repeatable shape of a whole post, the ordered beats that carry a reader from your first line to the thing you want them to save. These six structures are the shapes that earned the most reach across 531 posts from five creators who teach AI for a living. Steal the skeleton, pour your own workflow into it, and you turn a hard-won process into a post in 5 minutes.
What a post structure actually is
A post structure is a repeatable pattern for an entire post, not just the opening line, but the full sequence of beats that turns one idea into something worth saving. A hook wins the tap; a structure earns the save, and for AI content the save is the whole game, because a workflow people bookmark is a workflow they credit you for. This page is one spoke in our wider library of LinkedIn templates.
We clustered the highest-reach AI posts in our corpus into six structures that recur across the feeds of Allie K. Miller, Ruben Hassid, Charlie Hills, Nate Herkelman, and Maja Voje. Each one is a skeleton plus a real, verbatim post that used it, linked to the source. CaptureFlow is an AI content agent that turns your expertise into weeks of on-brand content for every platform. See every format it produces on the features page. If you are teaching AI under your own name, see how CaptureFlow fits your flow as a creator, or borrow the structures tuned for personal brands.
A credibility line, then the exact guides you wish you'd had.
N prompts, hacks, or moves people can run on the next prompt.
Tried a hundred tools, kept the few that survive real work.
Zero to working, one concrete instruction per line.
Name the outdated habit, then hand over what works now.
Let the tool do the job, then show the real before and after.
The 6 post structures
Each structure is a full-post skeleton, one real post that used it in the wild, and a couple more openers to prove the pattern. Swap the brackets for your own tool, workflow, and numbers.
1. The Resource Drop
Use this when you have done the reading so your audience does not have to: the guides, courses, or links that actually taught you a tool. It is the highest-reach shape in AI content because a fast-moving tool creates a fire hose of resources, and a hand-curated set reads as a shortcut through the noise. The credibility line up top is what earns the list underneath.
- 11[The price you paid to learn [tool] the hard way: hours logged, credits burned, dead ends hit].
- 22Here are the [number] resources I wish someone had handed me on day one:
- 33[Resource name]: [what it unlocks in one line] + [link]
- 44Repeat the line for every resource, hardest-won first, so the list scans like a curriculum.
- 55Tell them to save the post, because a link drop is only useful the day they actually need it.
Why it works: The credibility line ('1000 hours the hard way') is what earns the list under it. For a tool that changes every week, a hand-curated set of links reads as a rescue, and the whole post is built to be saved and reopened when the reader finally needs it.
| Opening line | Creator | Reactions |
|---|---|---|
| People think learning Claude takes days. It doesn't. | Ruben Hassid | 3,584 |
| My entire Claude skills library is now free. | Charlie Hills | 2,329 |
2. The Numbered Playbook
Use this when you can break a tool down into a countable set of moves: N prompts, N hacks, N ways to use it. Teaching AI is really teaching behavior change, and a numbered list turns a new habit into steps someone can run on the very next prompt. The number in the first line sets the contract, and each item has to earn its place by being usable mid-scroll.
- 11[A sharp callout of the wrong way people use [tool] right now].
- 22Here are [N] [prompts / hacks / moves] that fix it:
- 33[Number]. [The move, named in 2 to 4 words]: [the exact thing to type or toggle].
- 44Keep every item one screen-tap short, because a playbook people can act on is a playbook they save.
- 55Close on the payoff: [what changes once they run all [N]].
Why it works: One sharp callout ('stop typing long prompts'), then a numbered promise the reader can act on immediately. Framing 100 hacks as a saveable image turns a wall of tips into a single asset people bookmark and forward to their team.
| Opening line | Creator | Reactions |
|---|---|---|
| How to use Claude in 25 different ways: | Ruben Hassid | 2,756 |
| You are burning through your Claude credits by lunch. | Charlie Hills | 2,650 |
3. The Curated Roundup
Use this when the value is your judgement, not the news: you tried a hundred tools and kept the few that survive real work, or you are showing the exact stack you run. Among a hundred AI explainers all covering the same launches, curation is the differentiator. Group the picks by the job they do, not the hype they make, and the list reads as a point of view instead of a directory.
- 11[The filter you applied: tried [big number] tools, kept only the ones that hold up in real work].
- 22[The [N] that made the cut, or the exact stack I run every week]:
- 33[Number]/ [Tool]: [the one job it does in your workflow].
- 44Group them by the job they do, not the noise they make, so the list reads as judgement.
- 55[The through-line: why these [N] beat the [big number] you left out].
Why it works: A tool stack framed as 'hires I never made' gives every app a job title. The curation is the value: he is not listing tools, he is showing the reader the exact team he trusts to run his work, which is far more saveable than a neutral roundup.
| Opening line | Creator | Reactions |
|---|---|---|
| Go-To-Market Strategist covers 135 frameworks. | Maja Voje | 631 |
| 42 tools. 7 layers. Most GTM teams obsess over only one. | Maja Voje | 344 |
4. Step-by-Step Framework
Use this when you can take someone from zero to a working result with a clear sequence of steps. Put the outcome and the constraint in the first line, then walk it one concrete instruction at a time, no theory. A framework people can actually finish is a framework people save, and a line like 'most people stop here' quietly pushes the reader to the next step instead of closing the tab.
- 11[Outcome plus constraint: how to go from [zero] to [result] in [one weekend / N steps]].
- 22Step 1: [the first action, small enough to finish before they lose momentum].
- 33Step 2: [the next action, and why it compounds on step 1].
- 44Keep going to the last step, one concrete instruction per line, no theory.
- 55[The line that says: most people stop at step 1, here is what finishing unlocks].
Why it works: Outcome plus a time box ('all 4 layers in one weekend'), then leveled steps anyone can follow from a standing start. 'Most people stay here forever' turns a walkthrough into a challenge, nudging the reader to keep climbing instead of stopping at level one.
| Opening line | Creator | Reactions |
|---|---|---|
| Give me one minute, and I'll improve your Claude Code experience immediately. | Allie K. Miller | 2,316 |
| Claude Code does more than write code. | Charlie Hills | 1,752 |
5. Contrarian Teardown
Use this when a habit your niche treats as best practice is already outdated, which in AI happens roughly every quarter. Name the old way as if it were settled, then dismantle it and hand over the corrected model. In a field that resets every few months, being the person who says 'that is how it worked last year' is how you get known for a point of view instead of blending into the explainer feed.
[The way everyone still does [thing] with AI], stated like it is settled. But that is [the outdated / the expensive] move now. Here is what actually works: [your corrected model, in one line].
Why it works: Name the outdated habit, then replace it with a cleaner model of how the work has changed. Dating each era ('2024', '2025') makes the teardown feel rigorous rather than opinionated, and it positions him as the person tracking where the field is actually going.
| Opening line | Creator | Reactions |
|---|---|---|
| If your barrier to building an app was code, that barrier is basically gone. | Allie K. Miller | 2,341 |
| Some companies will use AI for efficiency, then think about sustainability. | Allie K. Miller | 2,017 |
6. The I-Tried-It
Use this when you let a tool do a real job end to end and can show the result. The claim is the hook; the honest mechanics under it are the proof. Show the before and after, name the exact stack, and admit the one seam it still got wrong, because a demo people believe is far more powerful than a demo that sounds too clean. This is how you turn a workflow you ran once into content nobody else can copy.
- 11[The claim: I let [tool] do [task] end to end, and here is what came back].
- 22Before: [the old way, in hours or headcount]. After: [the new way, in minutes].
- 33The exact stack: [tool] then [tool] then [tool], each doing one job.
- 44Show one honest seam, the thing it still got wrong, so the win reads as real.
- 55[What this means for the reader who has not tried it yet].
Why it works: The claim is the hook ('made this entire video by itself'), and the honest mechanics under it ('I didn't see a single frame') are the proof. Naming the exact stack and the real seams is what separates a demo people believe from one they scroll past.
| Opening line | Creator | Reactions |
|---|---|---|
| Claude Just Changed Video Editing Forever | Nate Herkelman | 1,457 |
| Claude Code runs my entire GTM system. Not a chatbot. Not a copilot. | Maja Voje | 1,209 |
6 rules that make any AI structure work
- One post, one workflow. A structure teaches a single move; a second tool or tangent splits the reader and both get forgotten.
- Front-load the payoff. AI readers scan for the save, so put the tool, the number, or the outcome in the first line.
- Show the exact thing. The prompt, the toggle, the model name, the link: specifics are what make a post saveable instead of vague.
- White space is structure. One idea per line, because a wall of text about a tool nobody has touched yet never gets read.
- Date your claims. In AI, 'works today' beats 'works forever', so name the model and the month and let the post age honestly.
- Reuse the skeleton, never fake the result. Keep the beats, swap in the workflow you actually ran, and the credibility stays yours.
How to use these post structures
- 1
Start from the workflow or take you already have, then match it to the structure whose job it fits: a set of links to The Resource Drop, a bold demo to The I-Tried-It, a hot take to the Contrarian Teardown.
- 2
Copy the skeleton and fill every bracket with your real specifics, the tool, the model, the hours it saved, the exact prompt.
- 3
Read it back one line at a time. If a line does not help the reader run the move themselves, cut it, and make the last line point at what they get for trying.
- 4
Short on time, paste your notes into the free LinkedIn post generator to draft the whole post in your voice, sharpen the opener with the hook generator, or turn a recorded walkthrough into posts with transcript to posts.
The takeaways
- 01A post structure is the repeatable shape of a whole post, the ordered beats that carry a reader from your first line to the thing they save.
- 02The six structures that dominate AI content: The Resource Drop, The Numbered Playbook, The Curated Roundup, the Step-by-Step Framework, the Contrarian Teardown, and The I-Tried-It.
- 03Structure solves the hardest part of teaching AI: turning a workflow you just ran into a post someone can act on, without starting from a blank page.
- 04Specifics are the moat. The exact prompt, model, or number you plug in is what makes a post saveable and traceable back to you.
- 05One post, one workflow, front-loaded, dated to the model and month so it ages honestly.
- 06Reuse the skeleton, never fake the result. The workflow you actually ran is what makes any of these land.
Turn these into posts
Frequently asked questions
- What is a LinkedIn post structure?
- A LinkedIn post structure is the repeatable sequence of beats an entire post follows, from the opening line to the closing payoff. Unlike a hook, which is only the first line, a structure shapes the whole post, so you can reuse the same pattern for a new tool, prompt, or workflow each time.
- How do I turn an AI workflow into a post people actually save?
- Start from the exact process you ran, then pour it into a structure built for it: a set of links becomes The Resource Drop, a numbered set of tactics becomes The Numbered Playbook, a full demo becomes The I-Tried-It. Naming the tools, the model, and the real numbers is what turns a workflow into something worth bookmarking.
- How do I stand out when everyone is posting about the same AI tools?
- Lead with judgement, not news. The Curated Roundup and the Contrarian Teardown both reward a clear point of view: which few tools survive real work, or which common habit is already outdated. Your read on the tool, not the tool itself, is the part a hundred other explainers cannot copy.
- How do I keep posting about fast-moving AI without burning out?
- Work from a handful of structures instead of a blank page, and let a tool handle the repetition. CaptureFlow is an AI content agent that turns your expertise into weeks of on-brand content for every platform, trained on your past posts so the voice stays yours. You capture one idea in 5 minutes, and it drafts the fan-out for you to review.