Resolve customer calls proactively with AI agents
TELUS uses BigQuery, Gemini for customer experience, and Gemini to automate 30% of customer calls and resolve them 87% faster. Task: TELUS ingested 22 million annual customer calls into BigQuery, using Sensitive Data Protection to scrub private details. The Gemini platform analyzed this data for sentiment and root causes, while Gemini for customer experience used multimodal AI to resolve issues across chat, voice, and video. Human agents received real-time guidance from AI agents for complex cases.; tools: Gemini. Human role: The original post does not say how a person stayed involved.; limitations: Implementation requires a unified data environment in BigQuery and Sensitive Data Protection for compliance. The source does not specify the exact human approval workflow for autonomous resolutions, though frontline staff receive AI coaching. Integration depends on migrating legacy scripts to natural language understanding..
Tools
Published on Agent Work Atlas:
Overview
TELUS uses BigQuery, Gemini for customer experience, and Gemini to automate 30% of customer calls and resolve them 87% faster.
Workflow
TELUS ingested 22 million annual customer calls into BigQuery, using Sensitive Data Protection to scrub private details. The Gemini platform analyzed this data for sentiment and root causes, while Gemini for customer experience used multimodal AI to resolve issues across chat, voice, and video. Human agents received real-time guidance from AI agents for complex cases.
Reported result
The system automated 30% of customer calls and resolved issues 87% faster, saving $53.9 million in annual operational costs. Previously, only 1% of calls were analyzed; now all 22 million interactions are reviewed to identify root causes and optimize responses.
Human role
The original post does not say how a person stayed involved.
Limitations
Implementation requires a unified data environment in BigQuery and Sensitive Data Protection for compliance. The source does not specify the exact human approval workflow for autonomous resolutions, though frontline staff receive AI coaching. Integration depends on migrating legacy scripts to natural language understanding.
Try it
I want to replicate TELUS's AI customer service workflow. I have [replaceable input: raw customer call transcripts or chat logs] stored in [replaceable input: BigQuery dataset]. First, apply Sensitive Data Protection to scrub PII. Then, use Gemini to analyze sentiment and root causes. Next, configure Gemini for customer experience to generate response recommendations for [replaceable input: target customer segment]. Finally, present a sample of 5 analyzed interactions with proposed resolutions to me for approval before scaling. Do not send any messages or make changes without my explicit confirmation. If data is missing, ask for the specific schema needed.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.