
Marketing Meetup Podcast Learn AI prompting for marketers: strategies that actually work in 2025
Oct 17, 2025
Anna Carina Berkman, founder of Markitect and senior AI trainer, dives into the transformative skill of AI prompting for marketers. She shares the COCO framework for crafting effective prompts and explores common mistakes to avoid. Learn about 'lazy prompting' and how to create reusable prompts for dynamic marketing needs. Anna also discusses personalizing chatbot outputs and addresses the risks of generative AI, including bias and data privacy. Her insights equip marketers with practical strategies for maximizing AI's potential in 2025.
AI Snips
Chapters
Transcript
Episode notes
ELIZA Demo Shows Early Chatbot Limits
- Anna demonstrated ELIZA, a 1966 Rogerian psychotherapist chatbot that loops and gaslights users.
- The example shows chatbots existed decades ago but were limited without modern internet-scale data.
Avoid Vague Prompts That Produce Generic Output
- Avoid vague prompts like 'summarize this' or 'write this better' and instead provide objectives, audience, tone, and CTA.
- Narrow prompts or add context to prevent generic, overused outputs.
AI Output Needs Human Judgment
- AI output is only step one; apply your expertise to judge, edit, and give feedback for better iterations.
- Treat AI as a 24/7 success partner but always apply critical thinking.
