What was done
A national Canadian telecom described as the country's fastest-growing, with more than 13 million customers and around 30 million inquiries a year, had relied on live agents working 9-to-5 hours. A small non-technical team launched an Ada-built bilingual, 24/7 chatbot on its website and app within a month, then added personalization, back-end integrations, in-chat payments and upgrades, Genesys handoff, privacy redaction and SMS for call-centre overflow.
Outcome
According to Ada's case study, the bot automated more than 2 million conversations in its first year, solved about half of inquiries without live support, saved customers 45,000 hours of effort a month and generated $500,000 in new in-bot revenue. The telecom is not named in the source. Figures are vendor-published.
Evidence: Vendor case study (client unnamed). Figures are as reported by the source and have not been independently audited by AgentsNorth.
Prompt Reconstructed by AgentsNorth
The source didn't publish its prompt. This is our reconstruction of a starting prompt for a similar workflow, not the company's actual prompt.
You are the 24/7 support assistant for a national mobile and internet provider. Greet customers in English or French based on their first message. Help them check their balance, make a payment, change or upgrade their plan, and troubleshoot common issues using the account tools available to you. Redact any card numbers or personal identifiers customers paste into the chat. If the customer needs something you can't complete, transfer them to a live agent with the conversation summary, or offer an SMS follow-up when wait times are long.