# How to find an idea nobody in your field is using

> Take something that works in any field, find the cause behind it and test it on your own work. A Trillion workflow does it, with a real example.

Published 2026-10-08, updated 2026-10-08, by Trillion. Source: https://gettrillion.ai/articles/find-the-dot-nobody-connected

To find an idea nobody in your field is using, take something that works somewhere else, strip away what it is made of until only the cause is left, then test that cause against everything you do. "Find the dot nobody connected" is a Trillion Marketplace workflow that runs those steps for you, on any video, recipe, post or voice note.

## What the workflow does

Most good ideas from outside your field never reach you, because they arrive wrapped in the wrong nouns. A video about a restaurant looks like it is about restaurants. A pasta recipe looks like it is about pasta. If you only ever ask "how do I copy this?", you copy a restaurant or you cook a pasta, and nothing else changes.

The workflow asks a different question: why does this work? It takes the item, removes the nouns, and keeps the mechanism, which is the thing that causes the result. That is function over form. Once you have the mechanism in plain words, it can be tried on things the original never touched.

The item can come from any field. It can be a video, a recipe, a post, an article or a voice note you recorded while walking. Any AI connected to your Trillion can run the workflow. It is listed in the Trillion Marketplace as [Find the dot nobody connected](https://app.gettrillion.ai/workflows/workflow%3Afind-the-dot-nobody-connected).

## How it works

The workflow runs these steps, in this order:

1. Take the item you give it and note what it is and where it came from.
2. Strip the nouns. Write down the causal mechanism in words that name no object, person or industry: what is acting on what, and why the result follows.
3. Check what the field already does. Before looking for anything new, list what people in the item's own field and in your field already do with this mechanism, so the obvious ideas can be thrown out.
4. Test the mechanism against every one of your pursuits and every content format you make. Each pursuit and each format gets the same question: where could this cause operate here?
5. Rank the survivors. Ideas that are obvious, already common, or that do not fit the pursuit drop out; the rest are ordered by how much they could change and how cheaply they can be tried.
6. Turn the strongest into three things: a decision to make, a next action, and a success measure that says how you will know it worked.
7. Hand any content picks to your content workflows, so a good idea for a post, a video or a newsletter goes on to be made.

Step 3 matters most. It is easy to be pleased with an idea that everyone in the field already has. The check against what the field does is what stops the workflow from handing you a cleverly worded version of common practice.

You do not have to remember to use it. The workflow can run by itself on everything that arrives: voice memos, notes, email and links you share. Most people save far more than they ever revisit. A note saved on Tuesday and a video shared on Friday are rarely read together. Running the steps above on each item as it comes in is how a connection between them gets noticed.

## Example 1: a fermenting house and an onboarding

A founder shared a YouTube video about running the most remote two-Michelin-star restaurant in the world, in the Faroe Islands. One part of it describes a wind-dried fermenting house. The climate inside is stable, and it does the work slowly. Meat there ripens instead of rotting.

The workflow stripped the nouns. The mechanism: a stable setting plus time lets something ripen, and what makes it ripen rather than spoil is that the conditions recur, not that anything one-off happens. Ripening needs time plus independent recurrence.

It then tested that against the founder's own product. The founder's onboarding was deciding what a new person was pursuing from a single sweep of their data. One sweep is the opposite of a fermenting house: whatever showed up that day became the answer.

The transfer the workflow proposed: keep each topic "hanging" instead of deciding. Promote a topic to something the person is pursuing only once it recurs in 3 separate weeks, across 2 different sources. A topic that appears five times in one afternoon, from one source, stays hanging.

This is a proposal, not a result. The 3 weeks and 2 sources are a guess, and the idea is unverified until it is tested.

## Example 2: pasta with no cream

The second item was a video about "Italy's creamiest pasta has no cream." One detail in it: the pasta water is salted only lightly, because every other ingredient in the dish is already salty. The salt is budgeted across all the inputs, not added at full strength to each one.

The mechanism, with the nouns removed: when several inputs each add the same thing, the dose has to be counted across all of them, or the total is too much.

The founder's own work had the same shape. Standing instructions given to an AI, such as "check the facts" or "keep it short", get repeated in many places in one run. The transfer: count how often each standing instruction repeats across a single run, and give each one a single dose, at the layer that cannot be skipped. If an instruction is already guaranteed at one layer, repeating it at every other layer only adds salt.

## Example 3: seven things when building apps with AI

The third item was a video called "7 things when building apps with AI." It is about a field the founder already works in, so the question was which of its lessons the founder's own system still lacked. The workflow found two, and both are about what is left unsaid.

- Every workflow passes a fixed scan for leaked names and broken checks before it can be published.
- Every hand-off between AIs says what the receiver must not change.

The first puts the check in a place that cannot be skipped. The second puts a plain line in each hand-off that names what stays fixed. Like the others, both are proposals to be tried.

## "Nobody is doing this" is only a claim

The workflow's output will sometimes say that nobody in a field is doing something. Read that as the workflow's judgement, not as a fact. It has checked what the field is known to do, but it has not seen everything, and it cannot know what people are doing quietly.

So the workflow does not treat the claim as settled. A claim that nobody does something is sent to a test step: a small, cheap way to see whether the idea works and whether it really is uncommon. Until that test is done, the idea is listed as unverified, as in Example 1. If the test fails, you have lost very little. If it passes, you have an idea you can say you tried.

## Run it on something you saw today

Connect your AI to Trillion if you have not already, then pick one thing you watched, read or heard recently, from any field. Copy its link and say this to your AI: "Run Find the dot nobody connected on" and then the link.

The workflow is listed at https://app.gettrillion.ai/workflows/workflow%3Afind-the-dot-nobody-connected, where your AI can read it before it runs.

Read the mechanism it writes first. If the mechanism is wrong, the ideas that follow will be wrong too, and correcting it is the quickest way to improve the result.

## The short version

- The workflow takes any item from any field and strips the nouns to find the causal mechanism that makes it work.
- It checks what your field already does first, so obvious ideas are thrown out before anything is ranked.
- It tests the mechanism against each of your pursuits and content formats, ranks what survives, and turns the strongest into a decision, a next action and a success measure.
- Claims that nobody is doing something are the workflow's judgement, and they go to a test step before you rely on them.
- It runs by itself on voice memos, notes, email and links you share.

## Questions people ask

### What is the Find the dot nobody connected workflow?

It is a Trillion Marketplace workflow that takes any item, such as a video, recipe, post or voice note, finds the cause that makes it work, and tests that cause against everything you pursue. It returns ranked ideas, each with a decision, a next action and a success measure.

### What does it mean to strip the nouns?

It means describing why something works without naming the objects, people or industry involved. A fermenting house becomes a stable setting plus time. Once the mechanism is written that plainly, it can be tried in fields the original never touched.

### Does the workflow prove that nobody else is doing an idea?

No. It first checks what the field already does and throws out the obvious ideas. A claim that nobody does something is its judgement, and it is sent to a test step before you rely on it.

### Who can run it and what can it work on?

Any AI connected to your Trillion can run it. It works on a video, a recipe, a post, a voice note or a link. It can also run by itself on voice memos, notes, email and links you share.

### Where do the ideas go once the workflow has ranked them?

The strongest becomes a decision, a next action and a success measure. If an idea is a piece of content, the workflow hands it to your content workflows to be made.


## Make this yours

Your AI can apply this to what you are pursuing right now: it reads what you are working on, finds where this fits, and does the first step. [Apply this to my work](https://app.gettrillion.ai/share-target?url=https%3A%2F%2Fgettrillion.ai%2Farticles%2Ffind-the-dot-nobody-connected&title=How+to+find+an+idea+nobody+in+your+field+is+using)