What are my customers complaining about?
Read every complaint without reading every complaint.
You'd think this requires a CX analyst and a weekend with a spreadsheet — Astra clusters every complaint across every channel and ranks by revenue at risk.
The short answer
Astra answers "what are customers complaining about" by ingesting every support ticket (Intercom), every Twitter/X mention of your handle, every App Store and G2 review, and every Discord/Slack message in your community channels. She clusters them by topic (not keywords — semantic clusters), then weighs each cluster by the MRR of the customers who complained. Output is a ranked list: Top 5 complaint themes, # of customers complaining, total MRR represented, sample verbatim quotes, and a verdict (Quick fix / Roadmap / Wontfix). She separates loud-but-low-value (free users on Twitter) from quiet-but-high-value (your $5k/mo customer who emailed once). You stop reacting to whoever yelled loudest and start fixing what's costing you the most actual money.
How Astra actually does it
- 1Ingest every channel
Intercom tickets (last 30 days), Twitter/X mentions of your handle, G2 + App Store reviews, Discord/Slack community channels, NPS verbatims.
IntercomTwitter APIDiscordSlackG2 - 2Cluster semantically
Group by topic, not keyword. 'Login is broken', 'can't sign in', 'OAuth failed', 'session expires too fast' all collapse into one auth cluster.
- 3Weight by revenue at risk
Map each complainer to their Stripe MRR. A cluster of 3 enterprise customers = bigger priority than 40 free-tier complaints, even if it looks quieter.
Stripe - 4Rank and verdict
Quick fix = ≤1 day eng work. Roadmap = real feature change. Wontfix = legitimate complaint but doesn't fit your strategy. Each verdict comes with reasoning.
- 5Write the report
Lark message: top 5 clusters, # of customers + MRR per cluster, 1-2 verbatim quotes per cluster, verdict, and the single complaint to fix this week.
Lark
What it looks like in chat
A weekly Lark message with the top 5 complaint clusters, customers + MRR per cluster, sample verbatims, verdict per cluster, and the single most valuable thing to fix this week.
Weekly Tuesday 10am, plus real-time alerts when a single complaint represents >$10k MRR.
Ask Astra this right now
We'll spin up your workspace, hand the prompt to Astra, and you see the answer in 60 seconds. Free.
Try this with AstraFrequently asked questions
What if customers complain in multiple languages?
Astra clusters across languages — a Japanese complaint about slow API and an English one collapse into the same cluster. The verbatim quotes stay in original language with a brief translation. Useful when you have international users but a one-language team.
What if I don't have Intercom — just email support?
She reads any email inbox you give her access to (Gmail, Front, Help Scout, plain IMAP). The clustering works the same. Most early-stage founders start with Gmail-only and add ticketing later — Astra doesn't care which tool, only that complaints reach her.
Can I tell Astra to ignore certain types of complaints?
Yes. Say 'ignore complaints about pricing — that's intentional' or 'don't surface free-tier complaints' and she filters them out of the weekly report. They still get logged for trend analysis but don't push the weekly priorities.
What does the report actually look like?
One Lark message every Tuesday morning, ~90 seconds to read. 5 clusters max, with the dollar figure at risk, a real customer quote, and a one-line verdict. No 200-page CX report. If you want the full data, ask her for the underlying spreadsheet.
Run your one-person company.
Hire your AI team in 30 seconds. Start for free.
Free to start · No credit card required · Set up in 30 seconds