For all the talk about LLMs writing ads and AI agents building creative scripts, there’s one simple truth we’ve learned over and over again at Zedge: the best-performing creative always starts with research.
And not just desk research or pulling a few quotes from a survey – we’re talking about emotionally accurate, context-rich, audience-first exploration. Ads work when they speak with people, not at them. So the closer we get to the user’s actual mindset, the better our creative outcomes.
Here’s the simple 3-step process we now use at Zedge, and how creative teams can adopt it using tools that are smart, fast, and practical.
Step 1: Get Inside Their Head (Gigabrain)
If you want to sound like your audience, you have to spend time social listening. That’s why we start with Gigabrain.
Gigabrain scrapes Reddit and YouTube comments to uncover exactly how people talk about the things they care about - annoying ringtones, phone clutter, customizing their lock screen to match their mood. It’s social listening on steroids, giving you a window into users’ frustrations, desires, and inside jokes.
Let’s say we’re researching Gen Z’s reaction to AI-generated wallpapers. We plug in keywords like “AI lock screen aesthetic” or “wallpaper apps for ADHD brains”, and Gigabrain pulls up a buffet of real conversations. You quickly hear what people love, what they complain about, and what they wish existed but hasn’t been built yet.
Why we use it:
- Gets us emotional language for ad copy
- Helps us spot unexpected angles (e.g. “phone vibe resets”)
- Great for quickly building empathy
Step 2: Check the Facts (Perplexity)
Once we’ve captured consumers’ emotional truth, we turn to Perplexity to fill in their rational gaps.
Perplexity is our go-to for rapid research with receipts. It combines the ease of a chatbot with the rigor of a research assistant -- surfacing answers with source links so we can dig deeper.
When we recently explored “Do specific wallpapers and ringtones affect user mood or behavior?” Perplexity helped us:
- Validate studies linking digital environments to mood
- Find articles on emerging personalization trends
- Compare competing apps and user behaviors
It’s ideal for crafting ad concepts that need a little more backbone, whether you’re justifying a user benefit, writing a hook, or framing a product insight.
Bonus: It’s lightning fast and doesn’t hallucinate nearly as much when you anchor your prompts clearly.
Step 3: Simulate a Real Interview (ChatGPT)
This last step? It’s where things get spooky good.
Once we understand what people are saying (Gigabrain) and why it matters (Perplexity), we use ChatGPT to simulate actual 1:1 interviews with our target segment.
Here’s the exact prompt we use (feel free to steal it):
I work at Zedge, a mobile personalization platform used by millions of people globally to customize their phones with wallpapers, ringtones, and more. My team is currently exploring new user segments, and we want to gain deeper, practical insights into a specific group.
The segment we’re focusing on is: Gen Z Android users who frequently switch up their phone aesthetic to match their mood.
I’d like to conduct a simulated interview where you fully step into the role of someone from this segment. Please embody their mindset, daily habits, digital behaviors, preferences, frustrations, and aspirations—especially as they relate to mobile personalization and content discovery.
I’ll ask you questions as if I’m conducting an actual user interview. Please answer from the perspective of this segment, elaborating on your challenges, motivations, needs, and what you wish Zedge could do better or differently.
Are you ready? Let’s get started.
Then we just…ask.
What’s your biggest frustration with wallpaper apps? What’s your go-to source of inspiration? What makes a ringtone fun instead of cringe?
It’s not perfect -- but it’s shockingly useful for surfacing nuance. We’ve used this to shape new ad angles and write creative briefs that let to successful creatives.
Why it works:
- Creates depth and nuance where data alone can’t
- Helps pressure-test ideas with real-sounding feedback
- Fills in “soft data” gaps for emerging or niche audiences
Why This Works for Us
At Zedge, this approach has changed how we are creative. It’s faster than traditional user research, but far more grounded than winging it.
- Gigabrain gives us voice-of-the-user language
- Perplexity helps us understand context and back up insights
- ChatGPT interviews let us simulate human responses, fast
We still validate, refine, and test – but we’re no longer starting from zero. The result? Creative that hits harder, lands faster, and feels personal from the first frame.
Final Word: Don’t Let “Fast” Mean “Shallow”
If there’s one thing AI can’t do for you, it’s care. Research still matters – not because it’s a marketing checkbox, but because it’s the best way to understand the human on the other side of your ad.
Use these tools not to skip steps, but to go deeper, more quickly. That’s where the creative edge lives now.