Your Support Inbox Might Be Your Best Product Research Tool

Product teams spend a lot of time asking users for feedback.

Feedback boards.

Surveys.

Interviews.

Feature request forms.

But while building FeedLog, I've started thinking about another source of feedback:

Customer support conversations.

Because users don't always realize they're giving you product feedback.

A question can actually be a feature request

Imagine a user asks:

How do I connect this to Slack?

Maybe they're asking for documentation.

But maybe you don't have a Slack integration at all.

That's not really a support question anymore.

It's a feature request.

Another user says:

I keep having to export this and clean it manually.

That's not formatted like feedback.

There's no:

"Hello, I would like to submit a feature request."

But there's clearly a product problem hiding inside the conversation.

Support and feedback are usually separated

Most products treat these as two different systems.

Support → answer the user → close conversation

Feedback → collect request → analyze → product decision

But the user doesn't care about that distinction.

They're just talking to you.

So we've been exploring a different model in FeedLog:

User conversation
      ↓
AI answers the question
      ↓
AI understands what the user is actually saying
      ↓
Bug / request / complaint / question
      ↓
Useful product feedback

The support conversation becomes product input automatically.

"Resolved" shouldn't mean "forgotten"

This is the part I find interesting.

A support agent might successfully answer a customer and mark the conversation as resolved.

Great.

But if 30 customers asked the same question this month, that's probably something the product team should know.

Maybe the documentation is bad.

Maybe onboarding is confusing.

Maybe a feature is missing.

Maybe something in the UI is broken.

The individual conversations were resolved.

The underlying problem wasn't.

AI might be more useful after the answer

There's a lot of excitement around AI answering customer questions.

That's useful.

But I'm starting to think one of the more interesting applications comes after the conversation:

• What are users repeatedly struggling with?
• Which feature requests keep appearing?
• Which complaints are increasing?
• What are users trying to do that the product doesn't support?

That's the direction we're exploring with FeedLog.

Not just:

Can AI answer this user?

But:

What can this conversation teach us about the product?

For small SaaS teams especially, that might turn the support inbox into one of the most useful product research tools they already have.