Let's get straight to it: AI is no longer just knocking - it's kicking the door wide open for unprecedented productivity gains. We're looking at a projected $15.7 trillion boost by 2030, driven primarily by AI-enabled efficiency.
And people working in product roles will be at the forefront of this major change. During my time as a PMM at Scaleway, I witnessed AI's power firsthand, boosting my overall productivity by 40% and up to 80% for copywriting tasks.
The revolution is already happening, so it's all about being ready for it.
Realizing significant productivity gains with AI can take just a few hours—if you use it properly by mastering the fundamentals.
1. Mastering the basics is everything. 🙇
After chatting with over 20 folks in the product space these past weeks, I've been struck by the divide in user experiences.
On one side, there's a group that can't stop raving about the efficiencies AI brings to their processes:
"It's changed the game for my market research."
"I'm finding new use cases daily. I've barely scratched the surface."
"ChatGPT has become my go-to assistant."
Others… not so much.
“It doesn’t understand what I need.”
“Results are irrelevant.”
“It only states the obvious.”
And you know what? The gap between these two groups boils down to one simple thing: knowing how to properly prompt.
Turns out, a whopping 80% of what you get out of AI hinges on how you prompt it.
So, it's not that AI's lacking; it's the prompts that are falling short.
Here's how to fix that and start crafting prompts that pack a punch. ❤️
2. The ✅ and ❌ of Prompt Writing
What’s a good prompt anyways?
At its core, a prompt is how you communicate with AI, letting it know what you're looking for, whether that's an answer, information, or code. Even with their advanced capabilities, AI tools require clear, direct instructions — they can't read minds, after all. But getting this right is definitely worth the effort.
Do’s
Be Clear: Make your prompt straightforward to remove any confusion, helping AI grasp precisely what you need.
Be Precise: Select your words wisely. For example, instead of saying "optimize performance," specify "improve the mobile app's loading speed to under two seconds."
Be Goal-Oriented: Keep a clear target in mind. Clearly defining your objective gives AI a specific direction to follow.
Be Context-Rich: Provide a clear context. If you're asking about technology, specify the particular field, your target audience, or mention any current trends.
Don'ts
Being too vague: Ambiguity leads to answers you weren't looking for.
Overcomplicating: Keep it simple to avoid confusing the AI.
Assuming it knows everything: AI's not up-to-date on everything unless you fill it in.
Using jargon or complex language: Plain and simple works best.
Asking double-barreled questions: Stick to one point at a time for clearer answers.
3. Example: bad vs good prompts
❌ Not Great
"Tell me about energy drink customers”
GPT4 Output:
Using such a prompt might yield an organized response, but a closer look reveals its limitations. The results are fairly predictable, citing common reasons like taste for drink preference, which hardly offers any new insights for Product Managers or Product Marketing Managers.
✅ Much Better
“You're a product manager at an energy drink company, aiming to understand customer needs to identify new product development opportunities. Identify the primary factors influencing consumers aged 25-35 in choosing energy drinks, focusing on flavor preferences, packaging design, and price sensitivity. Provide recommendations for a product launch targeting this demographic.”
GPT4 Output:
Notice the contrast? The outcome is far more detailed and insightful, shedding light on specific customer preferences. This level of detail equips a Product Manager with the information needed to organize a focus group, validating hypotheses for product development.
By the way, we’ve built a custom prompts library for PMs and PMMs. You can find it here.
Be clear about what you ask. Use straightforward language that leaves little room for interpretation. e.g. instead of "write about technology" specify "write a detailed comparison between the latest iPhone and Samsung".
Use precise examples: Saying "like the article in XYZ publication" can help steer the style or depth of information. You can also refer to previous output generated by the AI.
Give AI context. This might include the industry you're asking about, the target audience, or any recent developments that should be considered.
… but ensure the background information is directly relevant to your prompt. Overloading a prompt with unnecessary details can distract or mislead the AI.
Iterate until you reach perfection. Prompt writing is an iterative process. Don't be afraid to play around to see how they affect the AI's output. This can be particularly useful in finding the right voice for your content or product.
