10 AI Prompting Techniques That Actually Improve Marketing Output

A 2025 study of 9,000 AI users found a 6x productivity gap between people who use AI effectively and people who use it at a surface level. The tools are the same. The difference is skill. Dasha Shakov, Head of Marketing at Proton AI, has been building with AI since ChatGPT launched in November 2022. In this session, she shares the 10 prompting techniques that produce the biggest improvement in output quality — including a meta-prompting approach that delivers a 17% quality lift over standard prompting, and a "gaslighting" technique that forces AI to think critically instead of agreeing with you. These aren't theoretical frameworks. They're the techniques Dasha uses to run a full marketing team's output across content, field, product, and outbound functions. What you'll learn: - Why meta-prompting — asking AI to write the prompt before you write the prompt — improves output quality by 17% - How to describe the outcome you want rather than just the task, and why it changes what you get back - Why front-loading context (who you are, your audience, your constraints) before your instructions produces better results - The gaslighting technique: telling AI that someone criticized your work forces genuine critical thinking instead of validation - How to "steal expertise" by uploading reference materials and asking AI to apply a specific expert's thinking to your problem - Why breaking large tasks across multiple conversations produces better results than one long session - How dictating instead of typing (150 words per minute vs. 40) changes how you interact with AI tools entirely