“Needs Work” Usually Means Narrower Than You
Eleven cards, one job each, three tools stacked on every card. A red badge at the bottom that says needs work, an amber one in the middle for solid choice, a green one at the top for top pick. It is a clean format and it answers the question people actually have, which is not “what exists” but “which one”.
Now read what the card itself writes underneath each bottom-rung tool. “Team UI design.” “Self-hosted workflows.” “Blog post to video.” “Talking photos.” “AI inside your editor.” “Comment-to-DM classic.”
None of those is a description of something that does the job badly. Every one of them is a description of something that does a narrower job, very deliberately. The red badge is not measuring quality. On most of these eleven rows it is measuring breadth of fit — and breadth is only a virtue if your job is the broad one.
Which produces a simple correction to how you read a ladder like this: read the one-line description, not the badge. If the line names what you are actually doing this week, the colour above it is irrelevant.
What the three badges actually encode
| TOP PICK | SOLID CHOICE | NEEDS WORK |
|---|---|---|
| Widest fit | Good, with one clear specialism | Usually the narrowest, not the worst |
| Covers the most of the job for the most people | Excellent at one slice, thinner elsewhere | Built for a specific input, team or workflow |
| Right when your job is the ordinary version of the job | Right when your job is that slice | Right when the description names exactly what you do |
| It is also the most crowded and most expensive rung | Often the best value of the three | If it matches you, the badge is simply wrong for you |
All eleven cards, exactly as they run
Every word in this table is the card’s own — the job name, the three tools and the line it gives each one.
| The job | Needs work | Solid choice | Top pick |
|---|---|---|---|
| 01 · Smart Writing | ChatGPT — fast first drafts | Gemini — long docs and research | Claude — best long-form writing |
| 02 · Visual Images | Leonardo AI — stylised art presets | Nano Banana — fast edits inside Gemini | ChatGPT — best prompt understanding |
| 03 · Dynamic Videos | Pika AI — short fun clips | Kling AI — smoother camera motion | RecCloud — all-in-one video suite |
| 04 · Creative Design | Figma — team UI design | Canva — templates for everything | Photoshop — full pixel control |
| 05 · Digital Avatars | D-ID — talking photos | Synthesia — studio avatars, 140+ languages | HeyGen — best lip-sync and cloning |
| 06 · Modern Websites | Durable — a site in 30 seconds | Lovable — chat your web app into shape | Manus — agent builds and ships it |
| 07 · Powerful Apps | Cursor — AI inside your editor | Replit — code and host in the browser | Anything — describe it, get an app |
| 08 · Smart Automation | n8n — self-hosted workflows | Claude Code — agent that runs your tools | Zapier — 8,000+ apps connected |
| 09 · Creator UGC | InVideo — script to video fast | MakeUGC — UGC actors on demand | Creatify — best converting UGC ads |
| 10 · Faceless Videos | Pictory — blog post to video | Faceless.video — auto-posts your shorts | Syllaby — ideas, script and upload |
| 11 · Social Automation | ManyChat — comment-to-DM classic | BooSend — Instagram DM funnels | LinkDM — keyword links on autopilot |
WATCH THE ORIGINAL — 45 SECONDS
All eleven cards in sequence, each showing the weak pick, the solid pick and the top pick with its one-line reason.
▶ AI tools by usage, part 23 — the best tool for each job — BANI Academy on Facebook
Six rows where the bottom rung is the right answer
Here is the exercise that makes a ladder useful. Go down the red column and ask, for each one, is there a person for whom this is obviously correct? On this card the answer is yes six times.
Team UI design is not a weaker version of pixel editing. It is a different profession. If you are drawing screens that a developer will build, the top pick on that row will not help you and the bottom one is the industry standard.
Self-hosted workflows is the one row where the bottom entry has something neither of the others does: no usage meter at all. You run it, so nobody counts your tasks. The price is that you are now the administrator, which is a real cost — and the broader point about free plans being counted in units that do not convert is in 100 Tasks Or 1,000 Operations? That Isn’t A Comparison.
AI inside your editor is bottom of the apps row because the card’s job is “describe it, get an app”. For somebody who already writes code, an assistant inside the editor they already use is not a downgrade — it is the only one of the three that fits into an existing project.
Blog post to video is bottom of the faceless row because it starts from an input most people do not have. If you run a blog, it starts from an input you have a hundred of.
Talking photos and comment-to-DM classic are the same story: a specific capability and the most established option, respectively. Neither is a failure. Both are narrow.
The test, as a table
| If this is you | On this row | Pick | Why |
|---|---|---|---|
| You hand screens to a developer | Design | The bottom one — team UI design | Pixel editing is a different job |
| You will not pay a monthly bill, ever | Automation | The bottom one — self-hosted | You become the administrator |
| You already have a codebase | Apps | The bottom one — inside your editor | The top pick starts from nothing |
| You publish articles already | Faceless videos | The bottom one — blog post to video | It needs an input you have |
| You want one talking photo, not an avatar | Avatars | The bottom one — talking photos | Narrow capability, done |
| You run comment-to-DM and it works | Social automation | Stay where you are | Switching costs you a working funnel |
| None of the above applies to you | Any row | The top pick | That is exactly what the badge is for |
Four honest caveats
The card tells you what it is, and it is not a verdict. Every one of the eleven carries the same footer: these are how each tool performed for this job, when this was posted. That is a snapshot of a test, not a permanent ruling on a product, and the gap between those two things is where most tool-chart arguments live.
Two rows generate people, and that needs saying out loud. The avatar row and the UGC row both produce something that looks like a human being recommending something. Making them is fine. Running them as advertising without disclosing that the person is synthetic is a different matter — advertising rules in most markets require that endorsements are not misleading, and “that reviewer does not exist” is about as misleading as it gets. Label it, and you keep the tool and the trust.
“Describe it, get an app” still produces something you own and maintain. Three rows on this card — apps, websites, automation — end with a thing running in the world. Whether you can change it in a year, move it, or recover it when the vendor changes course depends on what you are actually left holding, which we worked through in Describe The Site, Get A Live Page. Then What Do You Own?
Eleven jobs is not eleven subscriptions. The post’s own opening line is the right one — stop opening five tabs for one task — but a card per job makes it very easy to end up with a tool per card. Before adding any of these, write the job as a sentence and check nothing you already pay for covers it, which is the method in It Promises No Overlap. There Are Three Overlaps In It.
Questions people ask
So are the rankings wrong? No. For the job as the card defines it, the ordering is reasonable and the reasons given are honest one-liners. The mistake is in the reading, not the chart: a single axis cannot hold a decision that has two.
How do I know whether I am the broad case or the narrow one? Write down what you did last week, not what you plan to do. Narrow cases announce themselves in your actual history — a blog you already write, a codebase you already have, a funnel that already converts.
Should I switch if I am already on a “needs work” tool and it is fine? No. A tool that is working is not a problem, and a swap costs you the week you spend relearning. Rank any switch by how much friction it costs rather than by the badge on somebody’s chart.
Eleven cards is a lot. Where do I start? With the job you most dread, not the row with the most exciting logo — and probably with one tool, not eleven. We made that case on a different sheet in Nine Of These Twenty-Four Are Tools You Use Once.
Who decides whether the output is any good? You do, and that has to be deliberate — the thing that produced the draft, the clip or the avatar cannot also be the thing that approves it. The cheapest fix is a rule written before you look: The Thing That Makes It Can Never Be The Thing That Grades It.
Read the line, not the badge
A three-rung ladder is a useful shape. It collapses a crowded category into one recommendation, and for most readers on most rows that recommendation is correct — that is what the green badge is for and it earns its place.
But the colours are measuring how well each tool fits the average version of a job, and nobody has an average job. So before you follow the arrow to the top of the card, spend ten seconds on the line underneath the red badge. Somewhere on these eleven cards, the thing labelled “needs work” is the only one built for exactly what you do — and the chart has quietly told you so, in four words, under a colour that says otherwise.
Learning a lot does not make anyone good. Doing a lot does.
If you want a deadline rather than a twelfth saved comparison, we run a free challenge — there is no barrier to entry, only a commitment fee that is refunded when you finish. It exists because a deadline uses the tool, and a saved card just owns it.
Which row on that card is a “needs work” tool you would actually defend? Tell us in the comments under the original reel.
More on building a channel for a global audience at mmoyoutube.com.
