Observability Set Up

Markdown

View as Markdown


Bringing Engineering Observability into CX

Observability was born in engineering, monitoring system health through signals and alerts.

At Oversai, we bring that same philosophy to Customer Experience (CX).

With Observability, you can:

• Consolidate hard metrics from your sources of truth (Zendesk, Intercom, Salesforce, Aircall, CCaaS platforms, etc.)
• Enrich them with Oversai intelligence (QA Score, AI Sentiment, AI Metrics, and more)
• Define KPI health ranges
• Create structured monitors
• Trigger alerts to the right stakeholders (Team Leaders, CX Managers, Product, Finance, etc.)


🏗 The Three Pillars of Observability

Observability is built on three core pillars:

  1. Metrics

  2. Monitors

  3. Notifications

This article focuses on Metrics and Monitors.
For Notifications, see: Observability Alerts /Notifications


📊 What Is a Monitor?

A Monitor is a structured group of KPIs you want to continuously track.

Think of it as a health dashboard for a specific team or function.

Example – Tier 1 Support Monitor

You might track:
• AHT
• CSAT
• QA Score
• Customer Sentiment
• FRT

Example – Sales Monitor

You might track:
• ARR
• Empathy (AI Metric)
• Total Interactions
• Conversion KPIs

Each Monitor reflects what matters most to that team.


📈 What Is a Metric?

A Metric is an individual KPI that feeds into a Monitor.

Metrics can come from:

• External integrations (e.g., Total Interactions, CSAT from Intercom)
• Oversai-generated intelligence (QA Score, AI Metrics, AI Sentiment)

Metrics can be:

✔ Simple Metrics
✔ Calculated Metrics (custom formulas)


Here’s where the step-by-step guide to configuring your Observability product begins:

⚙️ How to Create a Monitor

Step 1

Go to: Observability → Monitors
Click New Monitor

Step 2 – General Information

Fill in:

• Monitor Title
• Date Range
• Data Field (Created Date or Updated Date)
• Description (optional)
• Optional: Slack notification channel

Click Save to proceed.


🧮 Creating a Simple Metric

Inside your Monitor, click New Metric.

Here, you can choose whether to create a Simple Metric, for example, CSAT from your ticketing system, or a Calculated Metric.

Calculated metrics allow you to perform calculations using multiple existing metrics. For example, you might divide CSAT by Total Interactions to create a custom performance ratio.

This flexibility allows you to build your own tailored KPIs based on what matters most to your team.

🔹 Basic Information

  1. Display Title

  2. Unit (Units, Percentage, Minutes, etc.)

🔹 Metric Configuration

  1. Metric Source
    Select your data source (e.g., AI Metric, CSAT, Total Interactions)

  2. Aggregation
    Choose how the data is calculated:
    • Count
    • Sum
    • Average
    • Median, and more.

🔹 Field Filters

Segment your data using operators like equals or in.

Example:
Include only tickets that have:
• High Priority
• Medium Priority


🔄 Transformations

Apply mathematical adjustments to standardize or customize metrics.

Example:
Convert CSAT (1–5 scale) into a percentage:

(CSAT ÷ 5) × 100

This ensures consistent reporting across teams.


🚦 Health Rules

Health Rules define what “good” looks like.

Direction

Define whether higher or lower values are better.

Optimal Threshold

Example:
CSAT > 90 → 🟢 Healthy

Warning Threshold

Example:
CSAT < 85 → 🔴 Critical

Once configured, Observability automatically applies color-coded ranges.

Click Save to finalize.


🧩 Creating a Calculated Metric

Calculated metrics allow you to create custom KPI formulas.


Important: You must first create the simple metrics.

Steps:

  1. Define Title

  2. Select Unit

  3. Add Formula

Example:
(CSAT ÷ FRT) × 100

  1. Define Health Rules

  2. Save


🖥 Viewing Your Observability Dashboard

Once your Monitors and Metrics are configured, you’re ready to leverage your dashboard.

Go to:


Observability → Analytics

From here you can:

• Filter by date
• View near real-time performance
• Switch between monitors
• Filter by team
• Drill down into individual users

🎯 Immediate Visual Clarity

The dashboard uses structured color coding:

🟢 Healthy
🟡 Warning
🔴 Critical

This allows you to instantly understand performance health without digging into raw data.

👤 Individual Performance View

When selecting a user, you can:

• Analyze their KPI performance
• Detect trends over time
• Identify outlier tickets

Outliers are interactions that significantly deviate from baseline averages — helping you quickly detect anomalies or coaching opportunities.

Clicking on the Agents name, you will be able to drill down


🚀 Why Observability Matters

Observability shifts CX from reactive reporting to proactive performance monitoring.

Instead of manually checking dashboards, you:

✔ Define what matters
✔ Set health thresholds
✔ Monitor continuously
✔ Detect issues early
✔ Align CX, Product, Finance, and Leadership


Need Help?

If you need assistance configuring your Monitors or Metrics, your Account Manager will be happy to support you.

Start simple, iterate, and scale your monitoring strategy as your organization evolves.



Still need help?

Contact us

Observability