How I find topics
worth talking about
5 ways to research better content ideas with AI (and the prompts to try yourself).
Ask AI to give you 30 content ideas, and it’ll happily deliver
“5 mistakes everyone makes in your industry.” “3 things I wish I knew sooner.” “The ultimate guide to X.”
Sounds familiar? The problem is that AI can generate endless ideas without actually finding anything interesting.
When I research content for businesses and brands I help grow on social media, I prefer starting with something real: a conversation, a surprising study, a new development, or a question people keep asking.
Find something worth talking about first. Then figure out what to say about it.
That’s what we’re doing here.
There’s an AI agent for this
I actually built an AI agent called Media Scout to help me with this part of the job. It digs through news, studies, real stories and audience conversations, organizes the findings, and helps check the sources.
So if you’d rather skip the manual digging, you can check it out. But if you want to learn how to do it yourself, keep reading. Here are five ways to start using AI tools you already have.
Find what people are actually talking about
Instead of guessing what your audience cares about, go looking for what they’re already discussing.
Reddit threads, community forums, public discussions and comment sections are full of questions, frustrations and disagreements.
These can be much more useful than a list of generic content ideas.
prompt · try it yourself
Example
You’re a social media strategist and discover business owners repeatedly asking why their content gets views but no sales.
Now you can explore the difference between visibility and conversion, the audience you’re attracting, or how content connects to an offer.
One real problem. Several useful angles.
Look for disagreements, not just answers
Every industry has advice that gets repeated until everyone treats it as a rule.
“Post every day to grow.” “Always follow a morning routine.” “Automate everything you can.”
Maybe that advice works. Maybe it doesn’t work for everyone.
The exceptions are often where the interesting content lives.
prompt · try it yourself
Example
Instead of repeating “Post consistently,” investigate whether posting frequency actually improves results across different platforms and business types.
Your content might become: Is posting every day really necessary for your business?
A more interesting question than another generic consistency tip.
Find fresh studies, news and real stories
This is one of my favourite parts of research.
A surprising study, an unusual business decision, a new discovery — anything that makes you stop and think, wait, that’s interesting.
The challenge is that this material is scattered across publications, research databases, company announcements and websites. AI with web access can help gather it.
prompt · try it yourself
Example
You’re researching AI and consumer behaviour and discover a study about how people respond to AI-generated content.
Instead of immediately claiming “People hate AI content,” investigate what was actually measured, who participated and what the findings mean.
The real story might be about trust, transparency or how people perceive creativity.
Check where the claims actually come from
You’ve probably seen statistics like: “90% of consumers prefer…”, “Businesses that do X grow three times faster…”, “Scientists have discovered…”
Sometimes they’re accurate. Sometimes they’re outdated, taken out of context or entirely made up. And AI can repeat them just as confidently.
A link isn’t proof. The source needs to support the claim.
prompt · try it yourself
My quick fact-checking checklist
- Who originally said it?
- When was it published?
- What evidence supports it?
- Does the source actually say what the claim suggests?
- Is there important missing context?
You don’t need to turn every social media post into an academic paper. But if you’re using research to support your expertise, the facts should hold up.
Find what everyone else is missing
Competitor research usually goes something like this: open someone’s profile, find their most popular posts, figure out how to make something similar.
There’s value in understanding what performs well. But I think a more interesting question is: What aren’t they talking about?
Which audience questions are being ignored? Which perspectives are missing? What recent developments haven’t received much attention?
prompt · try it yourself
Example
You’re a nutrition professional and notice that most accounts in your field discuss protein intake in the same general way.
But your research reveals unanswered questions about protein needs for different age groups.
That’s an opportunity to create something more specific and useful than another high-protein meal list.
Okay, you’ve found something interesting. Now what?
Before turning a finding into a post, I like to organize it into a simple research card. Nothing fancy.
You can paste your research into AI and ask it to complete these cards for you. Then review the results and decide which findings are worth developing.
And don’t skip the “What’s my perspective?” part. Two people can read the same study and create completely different content.
Research gives you the material. Your expertise is what makes it worth following you for.
Or, you could let Media Scout do the digging
If you’ve tried a few of these prompts, you’ve probably noticed that proper research involves quite a few steps.
Searching different sources. Comparing findings. Checking claims. Collecting links. Organizing everything into something you can actually use.
I enjoy the discovery part. The repetitive digging? Not so much. That’s why I built Media Scout Agent.
It’s a downloadable AI research workflow for Claude and Codex that helps you:
- Find relevant news, studies, real stories and emerging ideas.
- Explore audience questions and critical perspectives.
- Gather organized research findings with source links.
- Review verification labels and supporting evidence.
- Identify material worth developing into content.
You provide your research direction, audience and goals. Media Scout handles much of the gathering and organization.
One last thing
One thing I’ve noticed while helping businesses and brands grow their social media is that the best content ideas aren’t always the trendiest.
The best ones happen when something people are already talking about overlaps with what genuinely excites you, makes your blood boil, or gives you something meaningful to say. That’s where I’ve found the best content comes from.
Happy creating! 🧋
Iuliia Sousa
Founder, POV Shifters