One Concept, One Example, One Question

After four issues exploring the questions in my Practical Delegation Test, this week begins a series on its Delegation Modes, starting with exploratory use.

The concept

The first question in my Practical Delegation Test asks whether I can describe the outcome clearly. When I can't, AI can still help me understand the problem well enough to decide what to do next.

Exploratory use fits work where the task itself is still being discovered. You might use AI to make sense of unfamiliar material as you work out what needs doing. It can also challenge the way you've framed a problem or suggest possibilities you haven't considered.

I judge that use by whether it opens up possibilities worth considering or exposes an assumption I need to check. A suggestion I eventually reject can still help me see why my original understanding was incomplete.

I'm still responsible for deciding which suggestions deserve investigation and what evidence would support them. AI's suggestions don't establish what's true or which option deserves a commitment, however convincing the wording. I still need to decide whether I know enough to define the next task.

The example

Imagine an organiser whose beginner sewing classes have attracted fewer people than expected. They think moving the classes online might help, but haven't established why attendance is low.

Rather than hand the task to AI, the organiser uses it to help identify possible reasons for poor attendance. They ask for competing explanations and questions to put to potential participants, giving AI the context they have while pointing out gaps.

AI suggests that the timetable might exclude interested people. The organiser hadn't considered its other suggestion: the advert could give the impression that people need some sewing experience. They use these possibilities to prepare questions about what people understood from the advert and what would prevent them attending.

Through a local community group, the organiser speaks to people the classes were intended to reach, including some who hadn't attended. Several knew the classes were for beginners but assumed they'd need their own sewing machine. Machines are provided, although the advert doesn't say so. AI hadn't raised that practical barrier, and moving online could make it harder for those people to take part.

The organiser decides to test an advert that explicitly says “machines provided” before changing how the classes run. They'll compare enquiries and attendance at the next session to see whether that clarification helps.

The question

For a task you're working on, what untested assumption could AI help you challenge, and what evidence would change your mind?

Related

• The Practical Delegation Test offers four questions to help you decide whether and how to use AI on a real task.

Elsewhere this week…

• I've published The Practical Delegation Test in Social Housing, bringing the four questions together to explore how housing providers can decide what work to give AI and what responsibility stays with the people delivering the service.

Support

Everything I publish here is free to read, and always will be.

I maintain this newsletter and my website in my own time. If something here has been useful to you, and you'd like to help cover the costs behind that work, you can support me through Ko-fi.

Generative AI is a powerful tool. I openly use it to conduct research, challenge my thinking, generate drafts, and review my work. AI never owns what I think, or my personal judgement. Everything I publish is a reflection of my own thoughts and opinions.