The Case Study

A telecom can reduce human transfers while leaving more customers with unresolved problems. That tension sits inside the request to double the reach of self-service within six months: The improvement has to preserve satisfaction, account safety, and access to people. Your first deliverable is a defensible interpretation of that request, with enough precision to expose missing evidence before an architecture creates commitments.

What does 45% mean?

The customer is a fictional telecom operating in four EU countries. It has 18 million subscribers and about 2.5 million support contacts each month across chat and voice, in four languages. The brief supplies no customer name.

Its existing FAQ chatbot cannot take account actions, and only 22% of contacts are fully handled without a human. Customer data resides in an older CRM and a slow, rate-limited billing system.

The mandate is to raise self-service resolution from 22% to 45% within six months without lowering customer satisfaction. Self-service resolution means that the customer’s issue is actually resolved without human assistance under an agreed measurement definition. The existing phrase “fully handled” needs examination before it can serve as a comparable baseline.

The current bot’s inability to act explains an important capability gap. It does not establish how much of the remaining workload is safely automatable, whether customers will use a replacement, or what integrations can support. Those questions require evidence.

Build an opening case record that captures the mandate, constraints, initial interpretation, stakeholder responsibilities, and discovery questions. Keep the telecom scenario separate from other case studies.

Calculate the requested change

Start with the contact denominator, rather than the subscriber population. Assuming monthly volume stays at 2,500,000 and baseline and target use comparable definitions:

• Baseline: 0.22 × 2,500,000 = 550,000 monthly self-service resolutions.

• Target: 0.45 × 2,500,000 = 1,125,000 monthly self-service resolutions.

• Additional resolutions: 1,125,000 − 550,000 = 575,000 per month.

• Percentage-point change: 45% − 22% = 23 percentage points.

These calculations translate the supplied mandate into its conditional operational scale. They describe a monthly target state, not a six-month cumulative total. Multiplying 575,000 by six would assume the entire uplift existed throughout the delivery period, which the mandate does not say.

The denominator counts contacts. One subscriber can contact support repeatedly, so 575,000 additional resolved contacts would not mean 575,000 additional unique subscribers helped. Nor does the arithmetic demonstrate feasibility, savings, or reduced staffing: Those require workload, cost, and operating evidence.

Before presenting the number to the sponsor, decide which interpretation its inputs support.

A correct calculation becomes misleading when its nouns change. “Additional monthly self-service resolutions at unchanged volume” ...