Painting of ducks in the water.

Koodo Mobile’s Port-In Support [Chatbot]

Role: Conversation Designer
Tools Used: Google Dialogflow, Mural, Google Sheets, Figma
Duration: 3 weeks
Team: 1 Conversation Designer, 1 Product Manager, 1 Developer

Problem Statement

Customers trying to transfer their phone number to Koodo often experienced failed port-ins due to missed SMS authorizations or mismatched account details. This led to:

  • High call volumes to support
  • Frustrated customers
  • Delays in onboarding

The goal was to design a self-serve chatbot experience that would:

  1. Identify port-in issues quickly
  2. Guide users to the correct resubmission flow
  3. Reduce agent handoff for known issue
Our Approach

We followed a user-centered, agile conversation design process, using the following steps:

  1. Discovery & Research

We:

  • Reviewed customer service transcripts
  • Interviewed support agents
  • Identified the top friction points:
    • “I didn’t get the text.”
    • “Not sure what happened to my transfer.”
    • “Do I need to start over?”

2. Mapping User Intents in Mural

Using Mural, we:

  • Mapped the user journey from failed transfer to resolution
  • Defined 3 key intents:
    • FailedPortIn
    • ResubmitPortIn
    • CheckPortInStatus

We also outlined fallbacks for vague requests like:

“My number isn’t working” → route to clarification

3. Flow Design in Mural

We prototyped the conversation flow visually in Mural:

📍 Trigger Intent:
“I’m trying to move my number to Koodo but it’s not working.”

🔁 Bot Response Flow:

  • Clarifies if it’s a port-in issue
  • Educates the customer on why transfers fail
  • Offers to resubmit with Yes/No buttons
  • Provides direct link to the port-in resubmission form
  • Confirms submission and offers further help

We added branching logic for:

  • Yes/No answers
  • User confusion or errors
  • Agent handoff when needed

4. Building in Dialogflow CX

In Dialogflow CX, we:

  • Created flows with state-based nodes
  • Used event handlers to manage Yes/No quick replies
  • Implemented rich content for web chat (buttons, links)
  • Tagged each node with fulfillment actions for analytics

Example snippet (Dialogflow CX):

5. Testing & Iteration

We:

  • Tested with internal teams via the Dialogflow web simulator
  • Used a spreadsheet of test cases (e.g., missed SMS, wrong provider info)
  • Improved fallback handling when users typed in incomplete info
  • Ensured tone stayed friendly, helpful, and on-brand

6. Results & Outcomes

After internal deployment:

  • 60% of port-in related queries were resolved without human agent intervention
  • Average resolution time dropped from 8 mins to under 2 mins
  • Increased resubmission form completions by 40% in the first month

Key Takeaways

  • Visual conversation design in Mural helps uncover edge cases before building
  • Dialogflow CX’s modular flow management makes iterating fast and scalable
  • Focusing on clarity + user intent drastically reduces confusion
  • Rich content (buttons/links) improves conversion and guidance

Maryam Salam