Find the product ideas hiding in your customer reviews

You’ve got 50 reviews sitting in your inbox and a nagging feeling you’re missing something important. Here’s exactly how to find it in under a minute: paste 5 of those reviews into Gemini and get back 3-5 clear feature requests, ranked by how often customers mention them.

That’s it. No spreadsheet, no rereading, no guessing what customers “really meant.”

Sound familiar?

You collect reviews because you’re supposed to. But when you actually read them, it’s just noise — “app is slow,” “confusing,” “wish it did more.”

You know there’s useful stuff in there. But turning vague complaints into an actual to-do list feels like a whole afternoon project.

This is exactly what AI is good at: spotting patterns across messy text that would take you forever to find by hand. It takes about a minute.

The 60-second fix

1. Grab 5 real reviews or complaints

Copy them from anywhere — app store reviews, support emails, a survey, even DMs. Doesn’t matter how messy or short they are.

2. Open Gemini and paste this prompt

Here are 5 customer reviews for my product. Read them and group the complaints into 3-5 actionable feature requests or product improvements.

For each one, give me:
- A short title for the feature/fix
- How many of the 5 reviews mentioned this issue (even indirectly)
- One suggested next step

Reviews:
1. [paste review 1]
2. [paste review 2]
3. [paste review 3]
4. [paste review 4]
5. [paste review 5]

3. Read the grouped list, not the raw reviews

Gemini does the pattern-spotting for you. You just react to the summary.

Real example

Input reviews:
1. “Love the app but it logs me out every day, so annoying.”
2. “Why do I have to sign in again every single time I open it?”
3. “Great features, wish there was a dark mode though.”
4. “Checkout is confusing, I almost gave up buying.”
5. “Constantly getting logged out on my phone, very frustrating.”

Gemini’s output:

1. Fix session/login persistence — Mentioned in 3 of 5 reviews. Next step: investigate token expiration settings, especially on mobile.

2. Simplify checkout flow — Mentioned in 1 of 5 reviews, but flagged as a near-lost-sale. Next step: run a quick usability test on the checkout page.

3. Add dark mode — Mentioned in 1 of 5 reviews. Next step: low priority, consider for a future release based on demand.

Five vague complaints just became a mini product roadmap — with a clear #1 priority (fix logins) before you even finished your coffee.

Level up

• Paste 20+ reviews at once and ask Gemini to also rank issues by “urgency vs. effort” so you know what to tackle first.
• Add: “Also tell me the overall sentiment (positive/negative/mixed) for each theme” for a quick sentiment snapshot.
• Ask it to write the feature requests as tickets, formatted like: “As a user, I want ___ so that ___” — ready to drop into your project tool.

FAQ

Do the reviews need to be perfectly written or formatted?
No. Copy-paste them exactly as customers wrote them — typos and all. Gemini reads meaning, not grammar.

Can this work for feedback that isn’t a formal “review,” like emails or support tickets?
Yes. Any text where a customer describes a problem or wish works the same way for feature request extraction.

How many reviews do I need for this to be useful?
Five is enough to see a pattern start to form. More reviews (10-50) give more reliable groupings, but even five beats reading them one by one with no structure.

Today’s check-in

Try this with 5 real reviews and reply in the comments with the #1 feature request Gemini found for you — you might be surprised what rises to the top.

Tomorrow: Got one good idea but no time to write 10 posts about it? Day 67 shows you how to turn one core idea into two full weeks of content in a single AI conversation.