Auto-generated description: Six ways to use AI are illustrated, focusing on automating tasks, generating ideas, analyzing data, creating content, discovering insights, and developing tools.

It’s not often I link directly to a LinkedIn post. However, the author of this, Ben Cohen, doesn’t seem to have posted it elsewhere, so needs must. Cohen also doesn’t cite the original source of the analysis he references, but it looks like it comes from an OpenAI report entitled Identifying and scaling AI use cases in which the “six use case primitives” are:

  1. Content creation
  2. Research
  3. Coding
  4. Data analysis
  5. Ideation/strategy
  6. Automation

Cohen has renamed these into much snappier “stuff” language, along with a graphic (see above) which looks like the OpenAI logo. I like this framing, it resonates.

I do wonder how many of these six use cases that vehemently anti-AI critics have actually used. I can confidently say I’ve used them all, and probably hit four of the six categories most working days at the moment.

Turns out there are only six ways to use AI well.

OpenAI looked at 600+ of the most successful GenAI use cases.

Every single one fell into just 6 categories (which I’ve taken the liberty to rename):

Create stuff → Content, policies, presentations, images, emails, contracts

Find stuff → Insights, research, competitor analysis, trends

Build stuff → Tools, websites, apps

Make sense of stuff → Data analysis, dashboards, performance reports

Think stuff through → Idea generation, strategies, decision-making

Do stuff automatically → Workflows, email automation, customer chatbots

That’s it.

Source & image: Ben Cohen | LinkedIn