Marketers; What difference does your choice of AI make?
I’m just gonna put it out there; I find the variety in the world of automotive brands baffling. Whether you’re after a roaring SUV or a nippy city car, you can pretty much take your pick of brands. So it really comes down to personal preference, as a Ford and a Mercedes-Benz serve the same base function of getting you around on the roads.
The point I’m slowly getting to is that you have your pick of AI tools to choose from. Whether it’s text, images, video or audio, there’s almost the full A to Z of tools out there. But do they all really serve the same purpose?
Let’s put them to the test
In the name of science, I decided to run a little test. I wrote 25 subject lines for a lead generation campaign. As you’d expect, a few were duds, some were pretty good;
The 25 subject lines
To start off, below is my list of subject lines - look them over, and see how you would rank them;
- Take the AI plunge
- Enough AI talk. How do you use it?
- Test your AI knowledge; then try your own
- Are you smarter than an AI?
- What the hell is a meta-prompt?!
- All prompts are equal (but some are more equal than others)
- What sort of prompts are you writing? When is a prompt not just a prompt?
- How to think about AI prompts
- Garbage In Garbage Out; writing AI prompts
- 6 types of AI prompt to try
- Level up your AI prompts
- Are you still zero-shot prompting?
- Understanding the different kinds of prompt
- Looking to level up your prompts?
- 6 kinds of prompts and an AI quiz
- How’s your AI eye?
- Got an eye for AI?
- How to write prompts like an expert
- What kind of prompts are you writing?
- What does a generative Ai need to know?
- Are you giving AI enough to work with?
- This is why AI is giving you generic results
- Generic AI results? Here’s why
- What makes a great AI prompt?
I didn’t really have a favourite. Enter AI.
I fed this list into 3 separate AI engines, Gemini, ChatGPT and Perplexity, and asked them all the exact same question:
“[LIST]. The intended audience is UK-based B2B marketing managers who are considering using generative AI for content, but haven't yet started. The desired outcome is for them to click on a link to a blog that explains the different kinds of AI prompts. Please rank the subject lines from most to least effective, and explain your reasoning.”
What I got back was 3 different rankings:
| Writing order | Subject Line | ChatGPT ranking | Perplexity ranking | Gemini ranking |
|---|---|---|---|---|
| 1 | Take the AI plunge | 17 | 19 | 18 |
| 2 | Enough AI talk. How do you use it? | 12 | 5 | 17 |
| 3 | Test your AI knowledge, then try your own | 7 | 20 | 15 |
| 4 | Are you smarter than an AI? | 8 | 22 | 25 |
| 5 | What the hell is a meta-prompt?! | 25 | 21 | 21 |
| 6 | All prompts are equal (but some are more equal than others) | 24 | 18 | 19 |
| 7 | What sort of prompts are you writing? | 14 | 9 | 12 |
| 8 | When is a prompt not just a prompt? | 11 | 11 | 20 |
| 9 | How to think about AI prompts | 4 | 10 | 1 |
| 10 | Garbage In, Garbage Out; writing AI prompts | 20 | 12 | 9 |
| 11 | 6 types of AI prompt to try | 5 | 1 | 4 |
| 12 | Level up your AI prompts | 2 | 3 | 5 |
| 13 | Are you still zero-shot prompting? | 16 | 16 | 8 |
| 14 | Understanding the different kinds of prompt | 23 | 8 | 3 |
| 15 | Looking to level up your prompts? | 15 | 15 | 6 |
| 16 | 6 kinds of prompts and an AI quiz | 6 | 25 | 16 |
| 17 | What does a generative AI need to know? | 19 | 23 | 23 |
| 18 | Got an eye for AI? | 18 | 24 | 24 |
| 19 | How to write prompts like an expert | 3 | 2 | 14 |
| 20 | What kind of prompts are you writing? | 13 | 17 | 13 |
| 21 | What does a generative Ai need to know? | 10 | 13 | 22 |
| 22 | Are you giving AI enough to work with? | 9 | 7 | 7 |
| 23 | This is why AI is giving you generic results | 21 | 6 | 10 |
| 24 | When is a prompt, not just a prompt? | 22 | 14 | 11 |
| 25 | What makes a great AI prompt? | 1 | 4 | 1 |
While there’s quite a variety across the list as a whole, 8 options were given the same ranking by 2 out of 3 of the AIs, though noticeably not always the same 2!
What about the reasoning?
I was also interested in WHY the various subject lines scored as they did. Each AI provided reasoning: ChatGPT and Gemini went through the whole list individually, while Perplexity only commented on the top 5. All of them provided summaries.
ChatGPT’s reasoning
1. What makes a great AI prompt?
Why: It poses a clear, outcome-oriented question that directly appeals to those curious about improving their AI content process. It promises insight into quality, which is both intriguing and educational.
2. Level up your AI prompts
Why: “Level up” implies improvement and growth; a message that resonates well with marketers looking to get ahead. It’s concise and action-oriented.
3. How to write prompts like an expert
Why: This subject line combines the promise of expert guidance with a “how-to” approach, positioning the content as a valuable resource for beginners and those looking to enhance their skills.
4. How to think about AI prompts
Why: It signals a deeper dive into strategy, inviting readers to explore a new perspective on AI content, a compelling prospect for decision-makers considering a new approach.
5. 6 types of AI prompts to try
Why: The list format is attractive because it promises bite-sized, actionable insights. The “6 types” also suggests variety, which can be enticing for someone looking for options.
6. 6 kinds of prompts and an AI quiz
Why: Similar to the previous entry, the numbered list and quiz element introduce interactivity and engagement; key for sparking curiosity.
7. Test your AI knowledge, then try your own
Why: This subject line invites participation and self-assessment, which can be a fun, engaging hook for marketers beginning to explore generative AI.
8. Are you smarter than an AI?
Why: It’s playful and thought-provoking. Although it might lean a bit more toward entertainment, it still creates curiosity about the subject matter.
9. Are you giving AI enough to work with?
Why: This question prompts readers to consider whether they’re effectively leveraging AI and nudging them to explore further solutions.
10. What does a generative AI need to know?
Why: It hints at the behind-the-scenes requirements of successful AI content, which can spark the curiosity of those wanting to optimize their prompts.
11. When is a prompt not just a prompt?
Why: This is clever and philosophical, inviting a deeper look into the art and science behind prompt writing. Even if it might be a bit abstract for some.
12. Enough AI talk. How do you use it?
Why: The opening statement suggests a break from jargon and focuses on practical application, appealing to busy marketing managers seeking actionable advice.
13. What sort of prompts are you writing?
Why: This engages the reader directly by prompting self-reflection on their current practices, although it’s slightly more ambiguous.
14. What kind of prompts are you writing?
Why: Similar in intent to the previous line, but the slight variation might resonate differently. It’s still an invitation to evaluate current approaches.
15. Looking to level up your prompts?
Why: This is a softer version of “Level up your AI prompts.” It’s friendly and suggests improvement, though it’s a bit less direct.
16. Are you still zero-shot prompting?
Why: While it speaks directly to AI techniques, “zero-shot” might be too technical or niche for those who haven’t yet started using generative AI.
17. Take the AI plunge
Why: It’s a call to action, but its brevity leaves some ambiguity about what the reader will gain: ideal for some but possibly too vague for others.
18. Got an eye for AI?
Why: This is playful and hints at insight, yet the pun might not translate as effectively for all readers in a professional context.
19. How’s your AI eye?
Why: Similar to the above, but the wordplay might be perceived as too casual or less direct, which can be less engaging for a B2B audience.
20. Garbage In, Garbage Out; writing AI prompts
Why: Although it uses a well-known phrase in the industry, the negative connotation (“garbage”) might deter readers who are just starting out and looking for a positive learning experience.
21. This is why AI is giving you generic results
Why: It hints at a problem-solution dynamic but starts off on a negative note. Marketing managers might prefer a more positive framing that offers empowerment rather than critique.
22. Generic A results? Here’s why
Why: The shorthand “Generic A” could be confusing, and like the previous line, it focuses on a negative outcome rather than offering immediate improvement.
23. Understanding the different kinds of prompt
Why: This subject line suffers from a grammatical hiccup (“prompt” instead of “prompts”) and feels less engaging due to its lack of energy or a clear benefit.
24. All prompts are equal (but some are more equal than others)
Why: While witty, the reference might come off as too informal or even off-putting for a professional audience seeking straightforward insights.
25. What the hell is a meta-prompt?!
Why: Although it can spark curiosity, the use of “hell” and the overly casual tone make it less suitable for a professional B2B audience in the UK, where a more measured tone is often expected.
ChatGPT’s summary
The top-ranked subject lines clearly articulate a promise of improvement, expert advice, or actionable steps, all of which resonate well with busy marketing managers who are in the early stages of exploring generative AI for content. In contrast, subject lines that lean too casual, negative, or ambiguous tend to be less effective in this context.
Gemini’s reasoning
1. How to think about AI prompts
Why it’s top: This is highly practical and addresses a fundamental need. It positions the blog as a guide to understanding the core concept, which is crucial for beginners. “How to” is a classic, effective phrasing.
2. What makes a great AI prompt?
Why it’s second: Directly addresses the quality of output, a key concern for marketers. It implies that the blog will provide actionable advice on improving results.
3. Understanding the different kinds of prompt
Why it’s third: Clearly states the blog’s content and its educational value. “Understanding” suggests a comprehensive and informative approach.
4. 6 types of AI prompt to try
Why it’s fourth: Offers a tangible takeaway (6 types) and encourages experimentation. “Try” is an action-oriented word.
5. Level up your AI prompts
Why it’s fifth: Appeals to the desire for improvement and advancement. “Level up” is a common and understandable phrase.
6. Looking to level up your prompts?
Why it’s sixth: Similar to the previous one, but phrased as a question, which can be less direct.
7. Are you giving AI enough to work with?
Why it’s seventh: Raises a potential pain point (insufficient input) and suggests a solution.
8. Are you still zero-shot prompting?
Why it’s eighth: While relevant, it might be too technical for complete beginners. It assumes some familiarity with AI prompting concepts.
9. Garbage In, Garbage Out; writing AI prompt
Why it’s ninth: Uses a familiar phrase that resonates with data-driven marketers. It emphasizes the importance of input quality.
10. This is why AI is giving you generic results
Why it’s tenth: Addresses a common frustration and offers a solution.
11. Generic A results? Here’s why
Why it’s eleventh: A shorter version of the previous one, but less impactful.
12. What sort of prompts are you writing?
Why it’s twelfth: A simple, direct question that encourages reflection.
13. What kind of prompts are you writing?
Why it’s thirteenth: Very similar to the previous one, small change in wording, but similar effect.
14. How to write prompts like an expert
Why it’s fourteenth: Appeals to the desire for expertise, but might be intimidating for complete beginners.
15. Test your AI knowledge, then try your own
Why it’s fifteenth: Includes a quiz element, which can be engaging, but might not be the primary focus for this audience.
16. 6 kinds of prompts and an AI quiz
Why it’s sixteenth: Same reasoning as 15.
17. Enough AI talk. How do you use it?
Why it’s seventeenth: A bit more casual and direct, which can be effective, but might not be suitable for all B2B audiences.
18. Take the AI plunge
Why it’s eighteenth: Uses a metaphorical phrase, which can be less clear and direct.
19. All prompts are equal (but some are more equal than others)
Why it’s nineteenth: A clever reference to “Animal Farm,” but it might be too abstract and not immediately clear to everyone.
20. When is a prompt not just a prompt?
Why it’s twentieth: A bit too philosophical and unclear for a straightforward marketing email.
21. What the hell is a meta-prompt?!
Why it’s twenty-first: While attention-grabbing, it’s too informal and might alienate some B2B professionals.
22. What does a generative AI need to know?
Why it’s twenty-second: bit vague.
23. How’s your AI eye?
Why it’s twenty-third: Very informal, and unclear.
24. Got an eye for AI?
Why it’s twenty-fourth: very similar to the previous one, and also very unclear.
Gemini’s summary
Key considerations for UK B2B Marketing Managers:
- Practicality: They want to know how to use AI to improve their work.
- Clarity: Avoid jargon and overly technical terms.
- Relevance: Focus on the benefits of effective AI prompting for content creation.
Reasoning for putting “Are you smarter than an AI?” at the bottom:
- While it’s attention-grabbing and plays on curiosity, it’s more of a clickbait-style subject line.
- It doesn’t directly address the practical needs of B2B marketing managers who are looking for actionable information on AI prompts.
- It may come across as gimmicky or less professional than the other options.
- It does not directly relate to the blog’s content, which is teaching the different kinds of prompts.
- Essentially, it’s more about sparking a fleeting moment of intrigue rather than conveying valuable information.
Perplexity’s reasoning
1. 6 types of AI prompt to try
This subject line is direct, actionable, and promises specific, valuable information. It appeals to the target audience of B2B marketing managers who are considering using AI but haven’t started yet, as it offers a clear, structured approach to getting started.
2. How to write prompts like an expert
This subject line appeals to the audience’s desire for expertise and improvement. It suggests that the content will provide valuable skills that can elevate their AI usage to a professional level.
3. Level up your AI prompts
This option is concise and action-oriented. It implies that the reader can improve their current skills, which is appealing to those who may have basic knowledge but want to enhance their capabilities.
4. What makes a great AI prompt?
This subject line poses a question that directly addresses the core issue of prompt quality. It’s likely to pique the interest of those who want to ensure they’re using AI effectively.
5. Enough AI talk. How do you use it?
This subject line acknowledges the abundance of AI discussions and pivots to practical application, which is likely to resonate with managers who are ready to move beyond theory to actual implementation.
Perplexity’s summary
These top choices are effective because they:
- Directly address the topic of AI prompts
- Promise practical, actionable information
- Suggest ways to improve skills or knowledge
- Are clear and concise
- Avoid jargon or overly technical language
- Appeal to the desire for expertise and effectiveness in using AI
The lower-ranked options tend to be less specific, use puns or idioms that might not resonate with all readers, or don’t clearly indicate the value of the content in relation to AI prompts.
The three combined
Taken together, it was clear that all 3 AI models preferred direct, action-oriented language. It is not surprising that you really want your email recipient to follow your call-to-action right away before your message gets lost among the roughly 150 others that they’ll get that day.
The shared negatives were jargon and informality. This speaks to the broad nature of the prompt that I used; “B2B marketing managers” is a very generic segment. Particularly in the case of formality/informality, there will definitely be industries where you have leeway to be a little more tongue-in-cheek and industries where you need to be straight-faced at all times. If you understand your own particular audience well enough, you’ll know where they sit on this spectrum.
So, what’s the lesson here?
I think there are two.
First off, each AI has its own quirks, and it’s worth taking the time to understand them and what that means, which tools you use and how.
Secondly, the role of humans in content might be more important than ever. We know our audiences and what they’ve responded to in the past, so we need to direct our AI assistants to focus their immense knowledge in particular ways.
Ok, just one more; AI can’t be surprising. It can only work from what’s gone before, so you won’t get a “Think Small” or “Eat Mor Chikin” from it. Subject lines and headlines that grab our brains with both hands and give them a shake remain a purely human domain.
To hear more from Spotler, visit spotler.com today, or come and meet our expert team at IRX at the NEC in Birmingham, 13th & 14th May 2025.
Want more like this?
Want more like this?
Insight delivered to your inbox
Keep up to date with our free email. Hand picked whitepapers and posts from our blog, as well as exclusive videos and webinar invitations keep our Users one step ahead.
By clicking 'SIGN UP', you agree to our Terms of Use and Privacy Policy
By clicking 'SIGN UP', you agree to our Terms of Use and Privacy Policy
Other content you may be interested in
Want more like this?
Want more like this?
Insight delivered to your inbox
Keep up to date with our free email. Hand picked whitepapers and posts from our blog, as well as exclusive videos and webinar invitations keep our Users one step ahead.
By clicking 'SIGN UP', you agree to our Terms of Use and Privacy Policy
Brought to you by
