Aurora Insights
The AI that reviews every conversation while your team sleeps.
Every night, Aurora Insights analyzes your team's conversations and delivers a prioritized list of broken promises, disgruntled customers, and unanswered questions. So you can start your day with the best to-do list you've ever had.
Asset
8 Important
14 News
Customer expresses frustration after two follow-ups without response regarding their order.
The agent promised a quote on Tuesday, but it hadn't been sent by Thursday.
Customer compared prices with a competitor; agent did not respond to the point.
What it detects
Four types of findings to get you started.
Aurora Insights reviews every conversation in your inbox and flags what needs attention. Plus, two automatic detections that close themselves: inactive conversations and lost transfers.
Broken Promises
“I’ll send you the quote on Friday.” Friday came and went, and nothing arrived. Aurora detects when an agent has promised something and hasn’t delivered.
Angry Customers
When a customer goes from "I have a question" to "this is the third time," AI identifies it before it escalates.
Unanswered Questions
“Do you ship to Monterrey?” Among five other messages, it gets buried. Aurora finds it and puts it on the list for the next day.
Custom Rules
Describe in one sentence what you want to detect, and the AI will apply it every night. Use as many rules as you need.
The problem
What is not seen, is lost.
Your team closes hundreds of conversations every week. The ones that require attention always get lost in the noise. Aurora Insights separates them for you, before it hurts.
Most customers leave because of a bad experience, not because of price.
It costs more to attract a new customer than to retain a current one.
It's the time limit before a lead goes cold
How it works
Three steps. Zero manual effort.
Set it up once. The AI does the rest every night.
Connect your channels
Aurora already has the entire conversation. WhatsApp, Instagram, Messenger, TikTok — everything that passes through the team's inbox is a candidate for review.
The AI analyzes overnight.
At 3 AM, Aurora reviews up to 5,000 conversations, applying built-in rules and those you define. Without slowing down your team.
Your team opens the list
Findings by priority — Critical, Important, Informative — with summary and suggested action. Mark as Reviewed, Discarded, or Action Taken and continue.
Integrated rules
The five things we already looked up for you.
They're enabled by default. You can turn them off whenever you want. Without writing a single line.
Broken Promises
It detects when an agent promised a quote, to return a call, or to send information and failed to do so within the agreed timeframe.
Important
Angry Customers
Identify conversations where the customer expressed significant frustration or anger. Address these yourself if necessary.
Reviewer
Lost Transfers
Active conversations without an assigned agent and with pending customer messages. They close automatically when someone takes them.
Important
Inactive Conversations
Customer messages left unanswered for more than 24 hours. It resolves itself as soon as your team responds.
Important
Unanswered Questions
Specific customer questions that the team didn't answer. The AI distinguishes between a real question and a greeting, so it only highlights what matters.
Important
Custom Rules
If you can describe it, AI can detect it.
You don't need to set up anything complicated. Just write in one sentence what you want your quality team to look for each night, and Aurora applies it to all conversations throughout the day.
-
Natural languageNo code, no booleans. How would you tell it to a colleague?
-
Define the priorityCritical, Important or Informative — you decide how urgent.
-
No extra cost per ruleAs many rules as you need. Tokens are charged per conversation, not per rule.
"Tick when a customer mentions a competitor or asks to compare prices."
"Dial when a customer complains about a discount that the agent didn't give them a number for."
"Dial when someone asks about the Pro plan but don't schedule a call."
Asset
8 Important
14 News
Third message without a response. Customer mentions canceling the annual contract.
Appointment scheduled for today at 10am — no confirmation sent to the client.
Inquiry about availability at the Guadalajara branch, not answered.
Triage in seconds
A tray for everything that slips through.
Each finding includes a summary of what happened and a suggested action. Your team leader logs in, opens the chat with a click, and decides whether to respond, dismiss, or note that it has been addressed.
-
Filters that matterBy owner, priority, rule, date. Three clicks and you only see yours.
-
Self-resolutionIf the team responded before you saw the finding, it closes automatically.
-
Visibility by roleEach agent sees only their own conversations. The admin sees the entire team.
-
Monthly summaryWhich rule triggered the most results? Which agent had the most findings? Data-driven coaching.
Who uses it
For teams that close one hundred conversations a day.
If you have more than one agent and a shared inbox, Insights pays for the month with the first customer it rescues.
Real estate
Where a missed visit costs more than a month of software. Detects clients who requested an appointment and never received confirmation.
Educational centers
Every parent who inquires about registration and doesn't receive a response represents a lost spot. Insights detects unanswered questions before the registration deadline.
Automotive agencies
High average ticket prices, long sales cycles. Detect when a prospect has cooled off between quote and test drive.
Frequently Asked Questions
What people ask us before activating it.
How much does it cost to use Aurora Insights?+
Insights uses Aurora Tokens — one per conversation analyzed. The first analysis is free (up to 1,000 conversations). After that, you pay with the tokens included in your plan or by adding extra packages.
When is the analysis performed?+
Every night, once a day. The scan starts in the early morning, and the results are ready when your computer opens Aurora in the morning. If your account was inactive for a few days, the next scan will catch up by reviewing data from up to a week prior.
Does AI read all messages?+
Read the most recent messages in each analyzed conversation. For most, it's the entire thread; for very long conversations, only the recent context. Your data stays within Aurora—it's not used to train external models.
Who can see the findings?+
Each agent sees findings only for their own conversations. Administrators see everything. Rule settings and the monthly summary are visible only to administrators.
How do I create a custom rule?+
On the Rules tab, click “Add rule” and describe in one sentence what you want to detect. Choose the priority and save. It will start working in the next nightly run. To write good prompts: be specific. “Mark when a customer asks for discounts” works better than “find sales problems.”
What if I have conversations in different languages?+
The AI detects the language of each conversation and writes the summary in the client's language. If you work with clients in both Spanish and English, you'll see findings in both languages in your inbox.
It starts tonight
Tomorrow morning you open your prioritized list.
Activate your first review without a card. Up to 1,000 free conversations. No tokens consumed.