Voice
Call transcripts and interaction details are captured as part of the customer history.
Project · Applied AI & Contact Center
SupportBuddy is a concept for connecting fragmented customer interactions across voice, email, and chat so agents can understand the active issue quickly, without manually piecing together the full history.
The problem
A customer may begin with a call, send a follow-up email, and later contact support through chat. Those interactions can be stored in separate places, leaving the next agent to search manually and rebuild the story during a live conversation.
This creates repeated explanations for customers, slower resolution for agents, and a less consistent support experience.
The approach
SupportBuddy uses available customer identifiers, interaction timing, active-ticket context, and AI summarisation to help agents understand the most relevant history before responding.
Call transcripts and interaction details are captured as part of the customer history.
Relevant email conversations are linked to the same customer context.
Chat interactions become part of the active support journey instead of remaining isolated.
High-level workflow
Use available identifiers such as phone number, email address, account context, and interaction timing.
Review open tickets and recent related interactions while excluding closed or irrelevant history.
Create a concise issue summary, recent activity timeline, and useful next-step context for the agent.
Deliver the result to the agent through an interaction workflow or screen-pop experience.
Expected value
Less time spent searching across disconnected systems
Better context for follow-up interactions
More consistent customer experience across channels
Faster issue understanding for support agents
Clearer handoffs between teams and interactions
Confidence-based review for incomplete or ambiguous matches
What this project explores
SupportBuddy explores multi-channel correlation, AI summarisation, customer identification, confidence scoring, ticket context, and delivering useful information at the moment an agent needs it.