Rebel, everyone’s shouting about AI and social media, but most of it is noise. Nicole Mezzasalma, social strategist at Battenhall, cuts through it with a journalist’s eye for what’s actually useful versus what’s just shiny.
About Nicole Mezzasalma
Nicole began her career as a journalist, spending 15 years in the industry before moving into marketing, and has spent 5 years at Battenhall focusing on social media strategy and innovation.
- LinkedIn: Click here
Treat AI Like A Junior Teammate, Not A Magic Box
Vague prompts get vague, generic results. Nicole’s fix: give AI the same context you’d give a new hire, brand voice, audience, goals, and examples of what “good” looks like.
“Think about AI as a colleague or a junior member of your team. You have to give it instructions. If you just say draft me 10 options of captions, it will come up with crap.”
Nicole Mezzasalma
- Before prompting, paste in a post you love and explain why, then ask AI to match that style
- Include your brand, audience, objectives, and tone in every prompt, not just the task
- Keep updating your AI’s instructions over time as you learn what works
Use AI For Research And Data, Not For Creative Judgment
AI is genuinely brilliant at trend research and crunching large reports fast. It’s not creative, it copies what already exists. Know which job to hand it.
- Give AI a market or topic you know little about and ask for a one-page summary with sources
- Use AI to draft caption variations, then apply your own judgment and voice before publishing
- Always verify AI’s facts and stats before sharing or reporting them, don’t trust without checking
Set Real Boundaries Around Data And Platforms
For a deeper dive into keeping the human element intact, listen to how to use AI in marketing without losing the human touch.
Not everything should run through AI, especially client data or free tools. And audiences react differently to visible AI content depending on the platform.
- Never put sensitive or client data into free AI tools, use paid, secure platforms instead
- Be wary of posting AI images on visual-first platforms like Instagram or Pinterest, audiences push back harder there than on LinkedIn
- Set a clear team policy: what AI is allowed for, and when human review is non-negotiable