Don’t Overthink Prompting

Prompting is Important, But Don’t Overthink It

When it comes to working with computers, you’ve probably heard the phrase “garbage in, garbage out”. And this is true for working with AI – the quality of your prompts matters a lot.

Have you ever struggled to get the right response from an AI tool? You’re not alone. The key lies in how you prompt the AI. But here’s the thing: you don’t need to overthink it either. Experimentation and learning by doing are the real keys to success.

Three Approaches to Prompting

Ethan Mollick, a professor at Wharton and a leading voice on AI in business, wrote about two approaches to prompting: conversational and structured. His insights have shaped my thinking on this topic, and I’ve found his approach to prompting incredibly valuable. And I want to add another approach: simple prompting.

1: Simple Prompting

Before diving into conversational or structured prompts, let’s talk about simple prompting.

Simple prompting involves giving very straightforward instructions to the AI. This is useful when you’re getting started or when the task is uncomplicated. Here’s why simple prompting is effective:

• It’s quick. You don’t need to spend time crafting detailed prompts.
• It’s easy. Anyone can do it, making it accessible for beginners.
• It’s effective. For many tasks, a simple prompt is all you need.

An example of a simple prompt might be:

• “Summarise this article.”
• “Generate a list of marketing ideas.”
• “Translate this text into Spanish.”

This method helps you get immediate, actionable results without overcomplicating the process.

2: Conversational Prompting

Mollick points out that for most people, most of the time, using very conversational prompts with AI is enough.

I’ve found this to be true in my own experience. I often use conversational prompting in a very literal sense as I use voice dictation to give long, detailed prompts to the AI. This approach works surprisingly well for me.

Here’s why it works:

• It’s natural. You’re just talking to the AI as you would to a colleague.
• It allows for A LOT of detail. You can provide a lot of context without worrying about structure.
• It’s flexible. You can easily refine and redirect as you go. I often add instructions at the start and then repeat and refine them at the end.

An example of a conversational prompt might be:

• “I’m working on a blog post about AI prompting. Can you give me some tips on different prompting methods? Also, include examples for each method. If you need clarification, ask me questions one at a time!”

3: Structured Prompting

While conversational prompting is great for many tasks, there are times when a more structured approach is necessary.

Mollick describes this as turning the AI into a tool that does a single task well in a way that’s repeatable and adapts to its user.

In my own work, I’ve developed a simple framework called GOAL to help with structured prompting:

• G – Goal: Clearly define what you want to achieve.
• O – Output: Specify the format and style of the response you need.
• A – Additional Context: Provide background information and relevant data.
• L – Look at the Output: Review and refine the response.

Structured prompts can be very useful, as detailed in my blog post on prompt frameworks. It covers several frameworks that can help you craft precise and effective prompts.

A structured approach is particularly useful when:

• You need consistent results across multiple queries.
• You’re working on a complex task that requires specific information.
• You want to create a template that others in your team can use.

Finding Your Balance

The key is to find the right balance between these three approaches.

Here’s what I recommend, based on both Mollick’s insights and my own experience:

1. Start simple. For most tasks, begin with a simple approach. It’s quick and often surprisingly effective.
2. Move to conversational prompting for more detail. If you need to provide more context, switch to conversational prompts.
3. Add structure when needed. If you’re not getting the results you want, or if you need more consistency, switch to a more structured approach like GOAL. For more on structured prompting, see my detailed guide on prompt frameworks.
4. Experiment and iterate. This is crucial. Don’t be afraid to try different methods. What works for one task might not work for another.
5. Use voice dictation for complex prompts. If you need to provide a lot of context, try using voice dictation. It allows you to give detailed prompts without getting bogged down in typing.
6. Create templates for repeated tasks. If you find yourself doing similar tasks often, create a structured prompt template. This can save you time and ensure consistency.

Making It Work for You

Here are some actionable steps to improve your AI prompting:

• Just start using it. Mollick emphasises that the best way to learn is by doing. Use AI for everyday tasks. Don’t overthink it – just chat.
• Try the GOAL method. Next time you have a more complex task, use the GOAL structure. More details on this can be found in my post on prompt frameworks.
• Use voice dictation. You might be surprised at how much more context you provide when speaking versus typing.
• Create a prompt library. Start saving prompts that work well for you.
• Share and learn. If you’re working in a team, share your best prompts and techniques.
• I wrote more about structured prompting here.

The AI landscape is evolving rapidly. What works today might be outdated tomorrow. That’s why it’s crucial to keep experimenting and learning.

So, dive in and start experimenting. Yes, the quality of your prompts matters, but don’t let that hold you back. The beauty of AI is that you can always refine and try again. The more you use it, the better you’ll get at crafting effective prompts.

The future of work is AI-assisted, and the best way to prepare is to start using these tools now. Don’t worry about getting it perfect – just get started. If you need support with AI, reach out – I’d be happy to help.