Customer Service AI Agent Deployment Project Manager Simulation — TELUS Communications

The scenario

TELUS Communications serves roughly 20 million customer connections across Canada and handles several million consumer support contacts a year through contact centres in Metro Vancouver, Edmonton, Calgary, Scarborough, and Rimouski. Over the spring, a Salesforce Agentforce service agent was piloted against 500 curated chat conversations on the consumer wireless line. It resolved 71% of them without a human, scored 4.4 out of 5 on post-chat satisfaction, and cost a fraction of a live interaction. The pilot report went to the board with a single recommendation: scale it. The pilot's success is why this project exists, and it is also the problem. Those 500 conversations were drawn from three of the cleanest intent categories in the queue: balance checks, data-usage questions, and plan comparisons. Production traffic is not curated. It arrives in French and English, from customers who are already angry, on billing disputes where the agent has to reason about money it can commit the company to. TELUS also owns TELUS Digital, whose AI engineering group sells this exact capability to other enterprises, and not everyone in the building agrees that a Salesforce agent should have won the work. You have been brought in as Senior Project Manager to take the agent from 500 conversations to the full consumer wireless chat line by quarter end, holding the 55% deflection target that the VP of Customer Experience committed to the board in June. The technology mostly works. What does not work is the agreement about what working means: the deflection number rewards containment, the contact centre measures resolution, the collective agreement protects the jobs the deflection number displaces, and the regulator holds TELUS to whatever the agent tells a customer. You are being asked to scale a success that nobody has agreed on the definition of.

What you'll do as the project manager

• →Move the Agentforce service agent from a 500-conversation pilot to 100% of consumer wireless chat and in-app messaging traffic by Day 27
• →Achieve and hold a 55% deflection rate on in-scope intents without degrading post-contact satisfaction below 4.0 of 5
• →Ship a human-handoff design that transfers full conversation context to a live agent within 20 seconds of a trigger condition
• →Establish guardrails and an approved commitment boundary so the agent cannot offer credits, adjustments, or plan terms outside a defined policy set
• →Deliver bilingual English and French coverage at equivalent measured quality before full traffic cutover
• →Complete technological-change notice and consultation with USW Local 1944 ahead of any change to bargaining-unit workflows

Project management skills you'll build

The challenges you'll navigate

• •The 55% deflection target was set from curated pilot data and may not be reachable on production intent mix
• •French-language quality is unmeasured, and official-languages expectations do not allow it to trail English into production
• •The agent can state commitments that bind TELUS under the Wireless Code, and no approved boundary has been defined yet
• •Changes to bargaining-unit workflows trigger technological-change obligations eight months before bargaining opens
• •Salesforce staffing quality on the production engagement is unproven, because the pilot ran with a different team

Technology & stakeholders

You'll manage 7 stakeholders, including Nadine Chartrand (VP, Customer Experience & Digital Channels), Colleen Fitzhugh (Director, Delivery Governance), Wesley Petrunia (Director, Contact Centre Operations — Western Canada), and more.

What you'll walk away with

A verified, shareable record of a completed enterprise project — plus the PMO deliverables you produced along the way (charter, project plan, SteerCo deck, closure document). It's real, demonstrable project management experience you can put on your resume and speak to in interviews.

Frequently asked questions

Related simulations