AI is now fully embedded in our day-to-day working lives. It’s reshaping how we create, research, plan and execute marketing campaigns. But while the tools we use are accessible to anyone with an internet connection, the output they deliver is only as good as the input they’re given. The old adage ‘garbage in, garbage out’ has never been more true when it comes to using AI tools.
This is the age of prompt engineering – a skillset now essential for anyone looking to truly extract meaningful value from generative AI. In much the same way we taught ourselves how to “google things”, as marketers we need to be able to “talk” to AI effectively. And it starts with mastering the art of the prompt.
What Is Prompt Engineering?
Crafting precise, contextual instructions – called prompts – that help AI generate relevant, high-quality responses. In marketing speak, it’s the equivalent of providing an agency with a creative brief — the more thoughtful your prompt, the better the output.
Common Prompting Mistakes Marketers Make
Like anything in life, practice makes perfect. Prompting is no different. The more you use it the better and easier it becomes. I’ve experimented tirelessly with many tools and learnt from my mistakes. These are the common mishaps I made:
- Being too vague and generic: simple prompts will lead to bland and generic responses. If you’re asking the AI tool to produce content for you, you don’t want it to sound like it was written by AI (even though it was). Be specific and clear in your ask.
- Forgetting who your audience is: Like any marketing campaign, if you’re not clear on who your audience is, your message will not hit the mark. Specify who you’re trying to reach so the AI can tailor tone and style accordingly
- Context is key: Set the scene. Let the AI know who you are, what your objective is and what channels you want to reach.
- Iterate your prompts: AI is not a magic button (although it often feels that way). If your first prompt doesn’t get the result you hoped for, iterate, tweak and optimise.
Helpful prompt framework
A tool I’ve found useful is the P.R.O.P framework:
- Persona
- Request
- Output
- Parameters
Keep these four things in mind when writing your prompt. For example:
“You are a senior B2B marketing leader (persona), suggest 10 topical blog ideas on the role of agentic AI in marketing (request). Include suggested titles and a short description (output). The blog ideas need to be relevant for B2B marketers within the tech industry (parameters).”
Better Prompts, Better Outcomes
If you’re anything like me, your first prompt attempts will look like the Basic Prompt examples I’ve given below. The more you test, fail and iterate, the better you’ll become.
Example 1: Blog Outline
Basic Prompt:
Write a blog about B2B lead generation.
Improved Prompt:
Act as a B2B content strategist. Write a blog outline for senior tech marketers on modern lead generation trends. Include a strong intro, 5 key sections, and a call-to-action focused on demand generation.
Example 2: LinkedIn Ad Copy
Basic Prompt:
Write a LinkedIn ad about our webinar.
Improved Prompt:
You’re a growth marketer writing a short LinkedIn ad to promote a webinar for IT professionals on the use of agentic AI in financial services. Use a bold hook, 2-3 benefits of attending, and a call-to-action. Keep it under 250 characters.
Example 3: Email Campaign
Basic Prompt:
Write an email to promote our new whitepaper on AI.
Improved Prompt:
Write a professional email for mid-level B2B marketers promoting a whitepaper on AI in marketing. Use an engaging subject line, a short intro highlighting the value, and a CTA button. Match the tone to our brand: helpful, confident, and modern.
A few more tips once you’ve grasped the basics
Once you’re comfortable with basic prompting, you’ll grow the confidence to explore of more structured and strategic techniques to get higher-value outputs:
- Chained prompting: Approach complex tasks in stages by breaking it down into a into steps — generate the outline first, then ask for expansion. It allows you to course-correct as you refine your thinking.
- Context stacking: AI models perform best when they have the full picture. Provide the right level of context each time you use it, as it won’t retain things such as your brand guidelines.
- Prompt templates: Create reusable prompt frameworks for blog posts, ad copy, email campaigns etc. that can be shared amongst your team. It will help empower team members to use AI more effectively (like yourself!)
Final Thoughts
With strong prompting skills, AI becomes your trusted collaborator and a handy tool in your B2B marketing tool kit. It’s a skill that will be relevant in any industry or job, and will even be helpful in your personal life (explain inflation to me like I’m 10). I used that prompt, very helpful.
Hugh Bowen, Senior B2B Marketing Leader
