WHAT PROMPTS SHOULD I BE ASKING CHAT GPT?

20 Apr

Written By AWITA

How do we engage with AI without sleepwalking into a future we don’t understand? Led by Martina Batovic, this event taught us how AI can help reduce administrative burden and support data-informed decision-making.

This second session taught us how to effectively prompt ChatGPT and evaluate outputs. By understanding the strengths and limitations of AI tools you’ll be able to shift focus to higher-value decision-making.

Members get exclusive access to AWITA events and the recording.

Catch up on the event recording here

5 TAKEAWAYS

1. Define the AI’s role and boundaries upfront

Treat AI like a domain specialist rather than a generic chatbot. Assign it a clear role (e.g., museum curator, PR strategist, market analyst) and specify tone, depth, and constraints. This reduces generic responses and improves relevance, especially in nuanced fields like art, communications, or strategy.

2. Prioritise verifiable, grounded outputs over fluency

Don’t optimise for how polished the answer sounds - optimise for accuracy. Instruct AI to avoid guessing, explicitly flag uncertainty, and provide sources or references where relevant.

3. Use AI as a multi-stage workflow partner

Treat AI as part of an end-to-end process rather than a one-time generator. Use it iteratively across stages: generating ideas, structuring thinking, drafting content, refining language, and supporting execution. This turns AI into a thinking partner rather than just a content tool.

4. Match the AI tool to the task

Different tools excel in different areas - some are better for speed and general tasks, others for depth, reasoning, or enterprise integration. Choosing the right tool for the job leads to higher-quality output than relying on a single system for everything.

5. Maintain consistency through account-based usage

AI becomes more useful when it builds context over time. Logging in to a account helps it learn your preferences, tone, and working style, leading to more relevant and personalised outputs. This is particularly valuable for ongoing projects and organisational use cases.

AWITA