OpenAI AI Foundations

A prompt is a brief. You learn to state the task, the audience, the constraints and the shape of the answer, then to improve a weak first result by changing the instruction rather than starting over.

• Task, audience, constraints and format
• Examples that show the style you want
• Asking for steps, options or a critique
• Fixing a weak answer by editing the prompt
• Splitting a big request into smaller ones
• Writing instructions you can reuse

You build A small library of tested prompts for five tasks you repeat.

A model cannot guess your situation. This module covers what background to supply, what to leave out, and how to hand over documents and data so answers fit your audience, your terms and your real constraints.

• What background changes an answer
• Pasting, attaching and summarizing source material
• Stating audience, tone and purpose
• Keeping a long conversation on track
• What not to share: sensitive and personal data
• Building a reusable context brief

You build A reusable context brief for one recurring piece of your work.

Output you have not checked is a draft, not a result. You learn to test claims, numbers and sources, to spot the signs of an invented answer, and to decide how much checking a task deserves.

• Spotting invented facts, quotes and sources
• Verifying a claim against a primary source
• Checking numbers, dates and calculations
• Asking the model to critique its own answer
• Matching the amount of checking to the stakes
• Recording what you verified

You build A review checklist, applied to three real outputs from your own work.

AI use touches other people: their data, their work and the decisions made about them. This module sets out privacy, confidentiality, bias, attribution and accountability, and how to follow your workplace rules while using it.

• Personal and confidential data
• Bias and unfair outcomes in generated content
• Crediting AI help and owning the result
• Reading and following your organization’s rules
• When a person must make the decision
• Telling others how AI was used

You build A short personal code of practice for using AI at work.

AI can explain, quiz and give feedback on your own attempts, which is a fast way to build the skills above. You learn to use it that way while doing the thinking yourself, so the skill is really yours.

• Asking for explanations at your level
• Having it quiz you and mark your answers
• Getting feedback on your own draft
• Practicing a skill in short repeated rounds
• Telling real understanding from a good-sounding answer
• Keeping a record of what you have learned

You build A weekly practice routine that uses AI for explanation, quizzes and feedback.

One good chat is useful; a process you can run again is worth more. You break recurring work into steps, put review points between them, and meet the idea of agents, which is where further study in AI leads next.

• Breaking recurring work into steps
• Defining what goes into and out of each step
• Placing a human check between steps
• What an agent is and what it needs from you
• Where further study in AI usually leads
• Choosing your next area of depth

You build A written workflow for one recurring task, with review points marked.

Bring every skill into one finished piece of work. You choose a real task, plan how AI will help, produce the result, check it, and explain which parts you trust and why.

• Choosing a task worth doing with AI
• Planning the prompts, context and steps
• Producing the work and iterating on it
• Checking and correcting the result
• Documenting how AI was used
• Reflecting on what to change next time

You build A finished project on a real task, with a record of what AI did and what you checked.