AI Tips #4: Simplify to Amplify Your ChatGPT Prompts
Discover how simplifying your approach into a step-by-step process can lead to more effective, relevant, and precise interactions with ChatGPT and other AI tools
Have you ever felt frustrated that you’re only able to get ChatGPT, Copilot, or Claude to accomplish simple or straightforward of tasks, and that they completely fall down when it comes to more complex problems? The secret sauce to working effectively with AI chatbots lies in breaking down complex challenges. Let's dive in! 🚀
Why it Pays to Break Down Complex Problems
As I frequently note in my writings, it’s useful to think of AI chatbots as an intern - a very smart and capable one, but with limited experience of how your specific subject matter, your organisation and how you might prefer to work.
When it comes to complex problems, there are therefore two potential issues one might face when working with these AI “interns”.
The first challenge stems from the fact that complex challenges can potentially be interpreted or solved in a very broad range of ways. There is therefore naturally a lot of room for ambiguity, and if only a basic “one-liner” prompt is used, chances are the AI will adopt an approach that you may not agree with or may not be appropriate.
Secondly, the more complex the problem, the more likely it is that the AI itself may make a mistake or an assumption that is not suitable to your context. This can particularly be an issue when it comes to complex logic or math challenges.
Both of these issues can be addressed by breaking the complex challenge down into a series of smaller, more manageable steps. This has the benefit of allowing the AI to work through the problem step-by-step, which increases its accuracy, and also gives one the opportunity to scrutinise its logic flow and assumptions.
Using AI to Break Down Complex Problems
If you have a good idea of how to break the problem down into a series of manageable steps, then go ahead and do so. Otherwise, you can ask the AI itself to take a stab at breaking problem down. Let’s assume that your boss has given you the task of developing a cost cutting (ahem!) value optimisation plan for the Sales team that you are leading.
Rather than simply telling the AI chatbot to provide you with a “value optimisation” plan, you could instead task it with identifying the steps required to develop a robust plan, and then walking you through each of those steps. If you’re looking for a specialised tool that can do this, you’ll be happy to know that I covered Goblin MagicToDo in an earlier post.
Once you’ve identified the steps, you would then collaborate with the AI to work through each step one by one, until you reach the end result. This is a technique known as prompt chaining, where the outputs of the first step, is used as the input for the second step, and so on and so forth.
Example #1: Building a Financial Model
Let’s get into the role of a financial analyst that has been tasked to develop a financial forecast model, which is by no means a simple task! Given my background in Finance, I’ve decided that I am able to layout the steps for this task without the help of AI.
My prompt would therefore look something like the one below:
“You will play the role of a financial analyst. Your objective is to develop a month-on-month profitability and cashflow forecast model over a 24-month period to support the launch of a new social networking / messaging product known as "WhotsApp". We will work through each of the following steps together, one step at a time, until we arrive at the completed financial model.
These are the steps we will take:
Develop advertising and partnerships revenue forecast from launch expectations and anticipated growth rates;
Develop initial platform development cost forecasts;
Develop ongoing fixed costs forecasts including infrastructure (e.g., hosting) and administrative costs (e.g., office space);
Develop ongoing variable costs forecasts including operational (e.g., customer support) and marketing and advertising costs;
Conduct profitability and break-even analyses;
Conduct scenario analyses by varying critical assumptions.”
As the user, you would then work through each of these steps in turn with the AI in an iterative process until you land on the final output that you are happy with.
Example #2: Developing a Marketing Campaign
In this second example, our task is to develop an email marketing campaign to again support the launch of “WhotsApp” (I know, this is not the most inspired example). As Marketing is not a one of my key disciplines, I will get the AI to take the lead in crafting the steps.
My prompt would therefore look something like the one below:
“You will play the role of a marketer. Your objective is to develop an email marketing campaign to support the launch of a new social networking / messaging product known as "WhotsApp". You will first lay out the steps required to develop such a campaign. You and I will then work through each of the steps that you have laid out together, one step at a time, until we arrive at the completed email marketing campaign.”
Conclusion
The beauty of chain prompting lies in its versatility and adaptability across various challenges, be they in Finance, Marketing, or any other field requiring nuanced problem-solving. Such an approach, together with collaborative iteration with AI, allows you to navigate through each facet of the problem, ensuring that every aspect is thoroughly considered and aligned with your objectives.
As always the goal is not to replace human insight but to augment it with AI. As you work through your problem together with ChatGPT, Copilot or Claude, remember that you should remain the final arbiter of the inputs, assumptions, and decisions undertaken at each step of the process.
Justin Tan is passionate about supporting organisations and teams to navigate disruptive change and towards sustainable and robust growth. He founded Evolutio Consulting in 2021 to help senior leaders to upskill and accelerate adoption of AI within their organisation through AI literacy and proficiency training, and also works with his clients to design and build bespoke AI solutions that drive growth and productivity for their businesses. If you're pondering how to harness these technologies in your business, or simply fancy a chat about the latest developments in AI, why not reach out?