Pride, prejudice, and prompt engineering


Danielle Wintrip CM

B2B & Professional Services Marketing
Authentic AI for Marketing Course

May 28, 2025

Dear Reader,

Lately, I have been seized with a desire to write in the Victorian novel style, which is how generative AI would have found its own expression if it hadn't been trained on the Internet. And if it had, we would have sentences like this one, which, to my mind, is the second most perfect sentence ever written.

"It is a truth universally acknowledged, that a single man, in possession of a good fortune, must be in want of a wife."

The trainers of large language models, such as ChatGPT, Claude (a French name!), or Notebook LM, fed these models a vast amount of words, and although we don't know exactly what they were, we do know that we get a lot of ordinary results as a result.

Ordinary content is okay. Ordinary content written awkwardly is okay. But boring content isn't OK. As my favourite author, Leo Tolstoy said, "Boredom: the desire for desires."

And herein we find today's learning about AI.

Desire something before you use AI.

Let’s call this the “Pride and Prompt” principle.

Before you type your prompt into ChatGPT, know your heart’s desire. Otherwise, you’re asking your AI assistant to choose a gift without telling it who it’s for, what the occasion is, or whether the recipient is a bookworm or a bonsai enthusiast.

Here’s your Prompt Etiquette Primer, fit for a regency ball or a Tuesday afternoon writing sprint:

S: Specificity

A good prompt is like a good gift—tailored, thoughtful, and clearly suited to the recipient.
Instead of: “Write a paragraph saying goodbye to a retiring colleague.”


Try: “Write a warm, 100-word note for a retirement card that celebrates someone’s creativity, work ethic, and legendary coffee-making skills.”

D: Desire

What do you want the final product to do? Make someone laugh? Clarify a complex idea? Persuade your reader to RSVP "yes" to your weird little garden party?


Instead of: “Explain climate change.”
Try: “Explain climate change to a ten-year-old using a birthday party metaphor.”

C: Context

Without context, AI flounders. Give it the who, what, and why. Set the stage like you’re throwing a dinner party and the AI needs to know who’s coming and what they can eat.


Instead of: “Write a story.”
Try: “Write a short bedtime story (under 300 words) for a child who loves dragons but is scared of the dark. End with a reassuring line.”

So the next time you prompt your AI assistant, be like Elizabeth Bennet: precise, principled, and never vague. After all, if you're going to summon a genie, it helps to know what your first wish is.

Yours truly,

Danielle

By the way, another Authentic AI for Marketing Course starts on Monday, June 2. It's five Mondays in a row, from 10 am-12 pm, live online. Click the button to learn more and register.

15 Beverley Street, Toronto, ON M5T 1X8
Unsubscribe · Preferences

Danielle Wintrip, Wintrip Communications

Read more from Danielle Wintrip, Wintrip Communications

Danielle Wintrip CM B2B & Professional Services Marketing Authentic AI for Marketing Course May 21, 2025 Dear Readers, I wrote a LinkedIn post today. LinkedIn offered to rewrite it with AI, so I said yes. C’était terrible. Terrible in French sounds so much better than in English. With a mouth full of marbles, the algorithm took my somewhat awkward but authentic phrasing and turned it into a marble puddle of muddle. It read like the drudgery of trudging through the mud on a hike in an emo...

Danielle Wintrip CM B2B & Professional Services Marketing Authentic AI for Marketing Course Hi [First Name], Working with ChatGPT can feel like talking to a teenager. Sometimes effusive, sometimes emo, and often not living up to its side of the conversation. Generative AI tools like ChatGPT can be brilliant, but only when we know how to ask for what we want. So, I created SEEQR©—a simple framework to help you think more clearly with AI, not just at it. SEEQR© isn’t another prompt formula,...

Hi Reader, Ever typed something into ChatGPT and gotten back a weird, flat, slightly-too-cheerful mess? I'm hosting a free webinar on Friday, April 11, at 11 a.m. ET called "How to Talk to GPT" It's for curious humans who want to feel a little more in charge when using ChatGPT—whether for business, everyday life, or just the joy of figuring out how this thing works. We'll cover: What most people get wrong when writing prompts Small changes that make a big difference in what GPT gives you back...