Signals — Track What Callers Say
Written By Anand Raj
Last updated 2 months ago
Signals are the things you want your AI Agent to listen for on every phone call — beyond the standard call summary and workflows. If Performance tells you what the Agent did, Signals tell you what your callers are saying about the topics that matter to your business.
Pro tip: Set up Signal tracking on day one. The system starts building stats from the moment a keyword is added — it cannot back-fill insights for conversations that already happened.
What you can track with Signals
Signals are designed for the questions you can't answer from raw call logs alone, for example:
Services you don't offer (yet) — track how many callers ask for a service you've never advertised. Real demand data to decide whether to add it.
A new provider or staff member — track what callers are saying about a recently onboarded provider (positive, negative, requesting them by name).
Operational pain points — long wait times, callback requests, rude staff complaints, missed appointments.
Revenue opportunities — refill requests, insurance verifications, weekend / evening slot demand.
Default vs Business-Specific Keywords
Open the settings icon on the Signals page to manage your tracked keywords. There are two types:
Default Keywords (System) — a curated set we ship with every business: Long Wait, Missed Call, Callback Request, Check-In Experience, Missed Appointment, Paperwork, Emergency, Connect To Human, Negative Review Threat, Rude Staff, Morning Slots, Evening Slots. These are auto-tracked and read-only.
Business-Specific Keywords (Custom) — your own keywords. Just describe what you want to track in plain English (e.g. "I want to track if a customer is complaining about Dr. John") and click Generate. The AI will turn that into trackable keywords like Accepting New Patients, Prescription Refill Request, Insurance Verification, Lab Results Enquiry.
Click Save to apply. From that moment forward, every new call is scanned for these keywords.
Signals — List View
The List view shows one row per call that matched at least one tracked signal. For each call you see:
Source — the caller's phone number.
Date — when the call happened.
Summary — short AI-generated description of what the caller enquired about.
Persona — New Customer, Returning Customer, Vendor, Spam, Unknown, etc.
Sentiment — Satisfied, Frustrated, Churn Risk, or neutral.
Severity — Low, Medium or High urgency.
Keywords — which tracked signals fired on the call.
Workflows — what the AI did as a result (Book Appointment, Note Taking, Call Transfer, Reschedule Call…).
Action — open the full call detail to listen / read the transcript.
Use the Persona, Sentiment and Severity filters at the top to drill into a specific slice — for example "all High severity, Frustrated, Churn Risk calls in the last 30 days."
Signals — Insights View
Switch to the Insights view to move from individual calls to aggregate trends. You get:
Sentiment counters — Satisfied, Neutral, Frustrated, Churn Risk totals for the selected date range.
Breakdown across all keywords — split of matched calls by Persona, Sentiment, Severity and Workflow. Click any keyword chip to refocus the breakdown on just that signal.
Top Keywords by Volume — the most-mentioned themes across all conversations (Long Wait, Rude Staff, Word of Mouth, etc.). This is the fastest way to spot what callers keep bringing up.
Use the date range picker (top right) to compare any window — last week vs. last month, before vs. after launching a new service, and so on.