User guides

Prompting for beginners: why context matters

Toqan doesn’t know what’s in your head

AI responses can be quite generic. Providing context helps Toqan to give you a response that’s closer to what you want. These examples show the difference in output. The original prompts don’t necessarily produce bad results but the more contextualised prompts give an output that’s more specific to your ask.

Example 1: an apology email

The original prompt: Write an email to my colleague to say sorry for my rude behaviour.

With more context: Write a sincere email to my colleague to say sorry for my rude behaviour this morning. I was frustrated because I felt she wasn’t listening to my concerns about our new projec. I’m worried I will be blamed if things go wrong. But I shouldn’t have responded with such rudeness.

Example 2: a thank you email to the team

An okay prompt: Write an email to my team to say thanks for their hard work on the Black Friday campaigns.

An option with more context: Write an email to my operations team to say thank you for the many extra hours they put into preparing for Black Friday sales. They sacrificed time with their families to help us prepare for this major event and their efforts are massively appreciated. We have seen a 7% increase in productivity thanks to the work they did.

Example 3: call for ideas

In this example, you’ll see how adding in the goal changes the output – it becomes much more focused on helping the entrepreneurs to stay in the game over a longer period.

The first prompt: Give me ideas for a cool digital project about entrepreneurship.

With more context: Give me ideas for a cool digital project about entrepreneurship that I can pitch at our next strategy meeting. Our CSI focus this year is to help talented entrepreneurs in Brazil to grow their businesses. We want to help them with advice on how to get started, how to stay on track and how to flourish, especially in the first few years when they struggle most.

You don’t need to write a beautiful prompt to give context – providing facts in bulleted form will be enough to help Toqan give a more personalised response.

Another hot tip here: if the contextualised prompt doesn’t give you the perfect answer, stay in the thread and give Toqan more info. Ask it to revise by keeping the additional information in mind.

This website uses essential cookies which are necessary to provide our site and services.

We only use essential cookies to ensure the functionality of the website, and to remember choices you have made to provide a more personalized experience (e.g., storing your contact details if you wish to join the waitlist). These cannot be deactivated.

For more details on the cookies we use and their purpose, please refer to our cookie notice for more details. Additionally, please refer to our privacy policy to see how your personal data is used in relation to the Toqan website.