The Do’s and Don’ts of Writing Prompts for AI

Alright, digital explorers! Ever felt the itch to have a chat with an AI? You probably already know, and perhaps have tried,  OpenAI’s ChatGPT, but do you know how to truly  command this digital beast to achieve the best results. No need to press the panic button, we’ve got you covered. Think of this as your no-nonsense guide to getting chatty with ChatGPT. 

We’ve been in the trenches, experimenting with all the tools and making all the rookie mistakes so you don’t have to. Now, we’re here to share what we’ve learnt to help you navigate your AI interactions more effectively.

But before you strap in and start your AI-convo adventure, here are some do’s and don’ts of writing prompts for AI.

Do: Be specific. Be clear!

In the world of marketing and storytelling, precision is everything. Imagine launching a campaign with the vague directive ‘promote our product.’ You might get results, but they’ll likely be as broad and generic as asking for ‘a story’ in ChatGPT and then receiving a run-of-the-mill fairy tale in return. For instance:

Vague Prompt: ‘Write about space.’
Clear Prompt: ‘Write a 200 word introduction about the challenges astronauts face during long space missions.’

Vague Prompt: ‘Discuss technology.’
Clear Prompt: ‘Provide a concise overview of the evolution of smartphone technology from 2010 to 2020.’

Don’t: Assume it’s all-knowing and always right

While AIs are impressively advanced and can tackle an array of problems, sometimes they might slip on a banana peel (digitally speaking, of course). Believe it or not, AIs can have their hiccups and blunders. Even the most sophisticated AI might draw an incorrect conclusion or miss subtle nuances in a prompt. Take these for example:

Prompt: ‘Who was the first person to walk on the moon?’
Potential AI Response: ‘The first person to walk on the moon was Julius Caesar.’
Issue: This is historically inaccurate. The first person to walk on the moon was Neil Armstrong in 1969 during the Apollo 11 mission. Julius Caesar was a Roman general and statesman from ancient history, having no connection to space exploration.

Prompt: ‘Why do we sneeze?’
Potential AI Response: ‘Humans sneeze to give their brains a quick reboot or to update their internal software.’
Issue: A fun idea, but not accurate. Sneezing is a reflex action to expel irritants from the nose or throat.

Do: Embrace the back-and-forth banter

It’s not all about firing questions and expecting magic! Think of it as weaving a tapestry of dialogue where each thread contributes to a bigger picture.

Engage in a flowing chat with ChatGPT, by layering your queries on previous responses. This approach can uncover insights and avenues of thought you might not have initially considered. It’s very much like a dance—sometimes it’s beneficial to let the AI take the lead, offering suggestions or guiding the conversation. At other times, be assertive with your prompts and guide ChatGPT to your desired destination. 

Making the most of ChatGPT requires a bit of back-and-forth, listening, adapting, and refining until you hit that sweet conversational spot. Check out what we did with ChatGPT:

User: Hey ChatGPT, can you help me brainstorm marketing campaign ideas for a new eco-friendly trainer?

ChatGPT: Sure thing! How about a campaign titled ‘Step into the Future’ that emphasises the trainers’ sustainable elements?

User: I like that! But can we make it more fun and engaging for a younger audience?

ChatGPT: Absolutely! How about ‘Green Feet, Neat Beats!’? Pair the sneakers with music, perhaps hosting a green concert or partnering with eco-conscious artists for limited edition designs?

User: Oh, that sounds cool! But what if we focus on the journey? Like, each step taken in the sneaker contributes to a greener world?

ChatGPT: Love where you’re going with that! ‘Every Step Counts’. We can create a digital counter tracking the cumulative steps taken by all wearers, symbolising their collective journey towards a sustainable future.

(Example generated from ChatGPT)

Do: Harness ChatGPT’s flexibility to mirror your voice

A unique feature of ChatGPT is its ability to adapt to various styles and tones on command. You can set the tone or ask ChatGPT to act a certain way (like writing in iambic pentameter). This comes in particularly handy if you’re working with multiple clients all with their own unique brand voice. 

Suppose you’re managing a campaign for two clients: one, a luxury watch brand targeting an elite audience, and the other, a fun and collaborative PR and marketing agency targeting SMEs in an array of different sectors. 

For the luxury watch brand, you might ask ChatGPT to: ‘Write in a sophisticated and elegant tone.’ followed by ‘and describe the timeless elegance of our newest watch collection.’ 
Potential AI Response: ‘Crafted with precision and an eye for detail, our newest collection embodies the very essence of luxury. Each timepiece is a testament to our unwavering commitment to excellence, resonating with the age-old adage that true elegance stands the test of time.’

Whereas for the PR and marketing agency, you could say: ‘Write in a casual, playful yet informational tone targeting SME businesses looking for marketing support’. You might ask ChatGPT to: ‘Talk about how messaging workshops are useful for you and your business’. 
Potential AI Response:  ‘So, what’s the deal with messaging workshops? Imagine you’ve got this product or service (which, of course, you do ’cause you’re fabulous!). But you’re struggling to shout it out to the world in a way that makes people go, ‘Whoa! I need that!’ That’s where messaging workshops come into play.’

(Example generated from ChatGPT)

Do: Play around with prompts! 

The more you experiment, the better you’ll understand AI’s strengths and quirks. Imagine ChatGPT as a sandbox of potential, and each prompt you give is like moulding a unique sandcastle. The joy isn’t just in the result but in the process of discovery and creation itself. Perfection isn’t achieved on the first try. By playing around with different prompts, you’ll not only uncover what ChatGPT is capable of but also its limitations.

Don’t: Use ChatGPT for super sensitive or critical tasks 

A quick cross-check never hurt anyone! Behind the sleek interface and rapid responses, ChatGPT operates based on algorithms and the data it’s been trained on (i.e. what’s available on the internet). While extensive, this data might not always capture the latest updates or nuanced contexts. In essence, while ChatGPT can make you more efficient, do not trust its outputs blindly. It always needs a human to validate it, to be its moral compass and emotional guide. 

The AI Chronicles Continue…

Navigating ChatGPT is a journey of precision, conversation, and experimentation. Always provide clear prompts and embrace the tool’s flexibility. While it’s fun to explore its vast capabilities, remember it’s not infallible. Keep a human touch on critical tasks, and together, you and AI can craft remarkable content.

So, you’ve just had a front-row seat to the first act of our AI series! Curious minds, brace yourselves; as we journey further, we’ll be testing an array of AI platforms, each with its own unique purpose. 

Eager to journey alongside us? Keep an eye our for our upcoming AI Landing Page. It’s the one-stop shop for our AI adventures, keeping you in the loop every step of the way.