YouTube's Missing Feature (That This Agent Just Solved)

I have a weird habit when I watch YouTube videos.

Before I even press play, I scroll down to read the comments. Partially to see what people think -- but mostly to figure out if the video is actually worth my time.

Views and likes get me to click the video, but almost instinctively, I know comments are where the real signal is. That's where you find out if the video delivers on its promise.

But reading between high-quality comments and trolls is messy and takes effort.

One of this week's top agents on agent.ai -- YouTube Video Conclusion -- just solved this problem for me.

So today, I want to break down:

• What this agent actually does (and why it matters)

• The "laughs per minute" parallel

• How to use it for free

—@dharmesh

What this agent actually does

The YouTube Video Conclusion agent is simple: it jumps to the end of any video, extracts the main conclusion, then analyzes the comment section to see if viewers actually agree with what was said.

Raise your hand if you've ever skipped to the end of a video just to get the main point. (Guilty!)

This agent does that automatically, then goes a step further by reading the room -- aka the comment section -- to see if the crowd is cheering or booing.

You paste in a YouTube URL and you get:

• The video's main conclusion -- summarized simply

• A Confidence Score (0-10) -- how confident the conclusion is in its own claims

• Summary of Comments -- key themes and sentiments from the comment section

• A Resonance Score (0-10) -- how much the audience actually agrees with the video's message

• Action Items -- next steps for viewers

I tested it on MKBHD’s video about the new NEO humanoid robot. It immediately surfaced that while the video concluded the product was overhyped, viewers were split -- some praised the skepticism, others felt the demo footage being remotely controlled made the critique unfair.

The Resonance Score of 8/10 showed Marques's critique landed well overall, despite the split.

That's the insight you can't get from view counts or likes.

As a viewer: I know if this video is worth my time, almost like a Rotten Tomatoes score for YouTube.

As a creator: You get valuable feedback without reading through troll comments.

Think of this as "Laughs Per Minute" metric

Comedians have this metric they obsess over: laughs per minute (LPM).

It's simple -- count the laughs, divide by runtime, and you know if your material is working. Instead of guessing if a joke landed based on “vibes”, it roots it in real data.

The YouTube Video Conclusion agent provides creators with a similar feature: a resonance score that indicates whether your content is resonating or falling short.

You can have a million views, but if the comments are full of people disagreeing with your thesis or pointing out errors, that's valuable feedback for next time. Right now, most creators have no efficient way to measure this at scale without reading every comment one by one.

YouTube Studio does let you filter comments by "contains question" or subscriber status (you can search, review, and respond). But there's no positive/negative scoring or sentiment charts, and no way to see at a glance if your message actually landed.

Eventually, I could see a comment sentiment analysis metric could be the north-star for creators, much like how LPM has become the north-star for comedians. And from a viewer's perspective, it's even more valuable, since you get a quick signal about whether a video is worth your time.

Thing is, YouTube has the data, the ML capabilities, and the platform to bake this feature right in for creators. So it makes sense for them to eventually build this.

But until then, agents like this one fill the gap.

Try it yourself

If you're a creator, researcher, or just someone who wants to understand if a video's message actually resonated with viewers, you can try the agent here.

It's particularly useful for:

• Content creators analyzing their own videos

• Researchers studying public opinion on topics

• Anyone trying to avoid clickbait videos or wasting time on content

The agent isn't perfect -- comment sections can be chaotic, and sentiment analysis has its limits. But it's progress.

As a viewer, it's like having a friend who's already watched the video tell you if it's worth your time. As a creator, it's feedback without the trolls.

Let me know what you think. And if you’re building agents on agent.ai for your real-world problems, I’d love to hear about them.

—Dharmesh (@dharmesh)