Skip to content

Generative AI & AI Workflows

Use generative AI well: prompting, tooling and workflows that hold up in real work.

  • Generative AI
  • Prompting
  • Productivity
Build my learning path

No account needed to see your path.

Nearly everyone has now used a language model. Far fewer can get consistent, trustworthy work out of one. The gap is not prompt tricks — it is knowing what these models are reliably good at, where they fail quietly, and how to build a workflow that catches the failures before anyone else sees them.

What you'll be able to do

  • Write prompts that produce the same quality of output on the tenth run as the first
  • Recognise the failure modes — confident invention, lost instructions, drift over long context
  • Give a model the right context instead of hoping it already knows
  • Chain a model into a real workflow with tools, files and checks around it
  • Judge when a task should not go to a language model at all

What happens when you pick this

  1. 1 Upskili checks what you can already do
  2. 2 It maps the gap between that and the skill
  3. 3 It sequences what is left, easiest dependency first
  4. 4 You practise it, and the path adjusts to how you do

Where this skill is used

Roles that ask for it. Name one when you state your goal and the path is built for that job rather than for the subject in general.

Common questions

Is prompt engineering still a useful skill?

The narrow version — memorising phrases that unlock better answers — has largely been absorbed by better models. The durable version has not: deciding what context a model needs, structuring a task so its output can be checked, and knowing which parts of a job should never be delegated to one. That is closer to systems design than to wording.

What is the difference between generative AI and machine learning?

Machine learning is the broader discipline of building systems that learn patterns from data. Generative AI is one application of it: models that produce new text, images, audio or code. Using generative AI well is mostly a skill of applying existing models; machine learning as a field also covers building and training them.

Can generative AI be trusted for professional work?

For drafting, summarising, transforming and exploring, yes, provided a person reviews the output. For anything where a confident wrong answer carries real cost — legal, medical, financial, or factual claims attributed to you — the model should be treated as a fast first draft that a competent human still has to verify.

Related skills

Band photograph by Steve A Johnson on Pexels.