Read Every Comment Under Your Competitor's Last 5 Launches

You can already see what your competitor is posting. That's the easy part.

The hard part is reading 400 comments under their best TikTok. Noticing three of their last ten posts mention a product line you don't carry. Realising the comments worth reading are in Korean and Portuguese.

You don't do it. You screenshot the chart and call it research.

What changed

Give Anna a competitor's handle. She pulls the posts. Opens the top performers. Reads every comment. Tells you what's being said.

In any language. With the quotes.

You don't connect anything. You don't export a CSV. The data lands as a dataset tab, like a spreadsheet upload would.

The prompt

If you ask Anna one thing this week, ask this:

"Pull the comments on the last 5 launch posts from @[competitor]. Run sentiment. Find the themes. Pull the products people name. Translate anything that isn't English."

This is the prompt because it's the work you can't do by hand and won't pay an agency to do every month.

A typical answer:

"Two negative themes recur across the five posts. 31% of negative comments mention shipping windows — concentrated in German and Korean. 18% mention colour accuracy in the photos. Both only show up 48 hours after the post drops, not in launch comments. Three product mentions you don't track keep coming up: Glossier Boy Brow, Rare Beauty Soft Pinch, and a Korean brand called Tirtir. The MUA tagged in two posts has 4× the comment-conversion rate of the others."

That's a content brief, a product brief, and a partnerships brief. Three paragraphs.

Why this is the prompt

Most social tools show you what's already visible. Followers. Post counts. Engagement rates.

The signal is in the comments. That's where teams stop.

• Volume. 400 comments × 5 posts × 3 competitors = 6,000 lines. Nobody reads 6,000 comments.
• Language. Half of any creator's audience comments in something other than English. Most teams skip them.
• Pattern. One negative comment is noise. The same complaint under three launches, in two languages, is a roadmap. You can't see the pattern without reading the 6,000.

Anna reads the 6,000. Tells you what recurs. Quotes the comments that prove it.

What you do with it

Three moves.

Brief your team. "Their last five launches got the same complaint about shipping from international audiences. Their domestic comments don't mention it. We can win there." That ends a meeting instead of starting one.

Find the products you should be carrying. When the same product shows up across competitors' comment sections, it's a category signal. The top five are obvious. Items 12-25 are the surprises.

Find the influencers worth talking to. Rank tagged creators by comment-conversion rate. The right names surface. Not by follower count — by who actually drives engagement.

What you stop doing

• Opening six tabs to count posts and compare engagement
• Pasting comments into spreadsheets to eyeball sentiment
• Running individual comments through a translator
• Building quarterly competitor decks that are stale by Tuesday
• Saying "they seem to be growing on TikTok" without the comments to back it

You can also ask

The launch-comment prompt is the starting point. Once your data is in, the same toolset answers a few related questions:

• "Pull the last 30 TikTok posts from @[handle]. Rank by engagement rate."
• "Pull @[handle]'s last 20 posts from TikTok, Instagram, and YouTube. Where are they investing? Where are they coasting?"
• "Pull the top 50 TikTok posts using #[hashtag] this week. Which themes are growing?"

These work. They're not the prompt to start with.

FAQ: AI competitor comment analysis

Which social platforms does Anna pull comments from?

TikTok, Instagram, YouTube, X (Twitter), Facebook Pages, and Threads. The connector handles public posts; private accounts and paid-only platforms (LinkedIn organic) are out of scope.

Do I need to log in as the competitor to pull their comments?

No. Anna reads public comments through each platform's public-data surface (or the platform's official API for read-only access). You connect your own account once for rate-limit handling; you never need credentials for the competitor's account.

How many comments can Anna analyse in one pass?

Typical run: the last 5 posts of a competitor, each with up to a few thousand comments — a few thousand to ~25,000 comments per analysis. For higher volumes, run multiple smaller pulls so Anna keeps her statistical and sampling guarantees intact.

How does Anna handle non-English comments?

She detects language per comment, translates to English for sentiment classification, and reports sentiment and theme breakdowns by language. The translation is for analysis; original text stays in the report so a native-speaker reviewer can verify.

Can Anna tell genuine engagement from spam or bot comments?

She flags comments that match common spam patterns (link-stuffing, repeated text, suspicious account age signals) and either excludes them from sentiment scoring or reports them separately. You can override either decision.

What is the legal status of analysing a competitor's public comments?

Public comments are public. Anna respects each platform's terms of service for read-only access and does not store the comments long-term — they are pulled per analysis. If a platform's TOS changes, the connector follows.

Can I track a competitor's launches over time?

Yes — run the same prompt on each launch and Anna remembers the comparison baseline. The "Q3 launch vs Q1 launch" prompt becomes one sentence after the first run.

Does this replace a brand-monitoring tool like Brandwatch or Sprout Social?

Partially. For lightweight competitor and launch tracking, it removes the need. For enterprise social-listening (millions of mentions, deep historical archives, alert routing), dedicated tools have features Anna does not. Many DTC operators only ever needed Anna's slice.