Transnational online service

Learning Program

Chatbot Analytics
done properly

Most teams deploy chatbots and then guess at what's working. This program teaches you to read the data — session drop-off rates, intent resolution, handoff triggers — and act on what you find. 8 weeks, 3 live group sessions per week, cohort size capped at 14.

8
weeks total
14
max per cohort
24
live sessions
Dashboard showing chatbot session analytics and performance metrics

Program Structure

Six modules, one clear thread

Each module builds on the previous one. You won't find isolated lectures — every session connects to a running project your group works on together, using real chatbot data from anonymised production environments.

1

Metric foundations

What chatbot metrics actually measure versus what they appear to measure. We cover containment rate, CSAT correlation, and why a 91% resolution rate can still indicate a broken bot.

Weeks 1–2 · 6 sessions
2

Conversation flow analysis

Reading session transcripts at scale. Participants learn to identify fallback loops, intent misclassification clusters, and the 3 conversation patterns that reliably predict high abandonment.

Weeks 2–3 · 5 sessions
3

KPI design for teams

Building dashboards that non-technical stakeholders can use without misreading. Covers threshold-setting, anomaly flagging, and which 6 metrics belong in a weekly ops review.

Weeks 3–4 · 4 sessions
4

Handoff and escalation data

When a bot transfers to a human agent, that event carries data. This module focuses on extracting signal from handoff logs — timing, trigger phrases, agent resolution time — to close gaps in automation.

Weeks 4–5 · 4 sessions
5

A/B testing in live bots

Running controlled experiments on deployed chatbots without disrupting production. Participants design and evaluate 2 live tests during this module using the group's shared test environment.

Weeks 5–7 · 3 sessions
6

Reporting and decision cycles

Translating analytics into decisions that get implemented. The final module covers structuring monthly performance reviews, presenting findings to leadership, and building a continuous monitoring habit.

Weeks 7–8 · 2 sessions

Format

Live

Group sessions via video, recorded for async review

Weekly load

~6h

Sessions + independent practice exercises

Language

EN

Materials available in written English throughout

Access

Global

Participants join from any timezone, scheduling rotates

Where participants land after eight weeks

The program doesn't promise transformation. It gives you a repeatable process for reading chatbot performance data, finding the problems worth fixing, and communicating what you find to the people who control the roadmap.

Structured audit process

A documented 5-step method for auditing any chatbot deployment, applicable across platforms — not tied to one vendor's tooling.

Dashboard templates

Ready-to-adapt templates for weekly and monthly reporting, built around the 6 metrics that matter most in operational reviews.

Group practice network

Cohort members typically continue exchanging data interpretations and test results for months after the program ends — the group dynamic persists.

Skill area Before the program After the program Difficulty level
Reading session transcripts Manual, time-consuming, inconsistent Pattern-based review in under 40 minutes per audit Foundational
KPI selection Defaulting to platform-provided metrics Custom metric sets aligned to business objectives Intermediate
Stakeholder reporting Raw data exports or verbal summaries Structured reports with clear decision prompts Intermediate
A/B test design Ad hoc changes without control conditions Controlled experiments with defined success criteria Advanced
Escalation data analysis Rarely examined, treated as support overhead Systematic handoff log review integrated into weekly ops Advanced
Reading session transcripts
Before: Manual, time-consuming, inconsistent
After: Pattern-based review in under 40 minutes per audit
KPI selection
Before: Defaulting to platform-provided metrics
After: Custom metric sets aligned to business objectives
Stakeholder reporting
Before: Raw data exports or verbal summaries
After: Structured reports with clear decision prompts
A/B test design
Before: Ad hoc changes without control conditions
After: Controlled experiments with defined success criteria
Escalation data analysis
Before: Rarely examined, treated as support overhead
After: Systematic handoff log review integrated into weekly ops