What Kinds Of Workplace Tasks Does Generative AI Handle Well?

Which workplace tasks suit generative AI, when human review matters, and how to pilot a low-risk use.

The takeaways

01Start with language-heavy tasks that allow more than one reasonable output.
02Drafting, summarising, transformation, and classification are common task shapes.
03Check accuracy, privacy, consequences, and accountability before using AI for a task.
04Pilot a low-risk task with human review, then measure quality, time, and consistency.

Read this comic as text

Reader Which of these tasks should I give to generative AI?

Robot Start with language-heavy work where several reasonable outputs are possible.

Takeaway: Start with language-heavy tasks that allow more than one reasonable output.
Reader What task shapes should I look for?

Robot Drafting, summarising, transforming content, and classifying varied inputs are common strengths.

Takeaway: Drafting, summarising, transformation, and classification are common task shapes.
Reader If the AI can produce an answer, is it suitable for the job?

Robot Not always. Accuracy, privacy, consequences, and accountability still determine whether it should be used.

Takeaway: Check accuracy, privacy, consequences, and accountability before using AI for a task.
Reader I will pilot one low-risk task with clear examples and a review step.

Robot Then measure whether it improves quality, time, or consistency.

Takeaway: Pilot a low-risk task with human review, then measure quality, time, and consistency.

Summarisation: Reducing longer material into its most important points.
Transformation: Rewriting content into a different tone, format, audience, or structure.
Classification: Sorting inputs into useful categories.
Human review: A person checking and approving AI output before it is used.