PwC telecom client
NLU migration and experience redesign for 4 troubleshooting bots
Through PwC, I supported one of the largest telecom companies in the United States. This was the prime opportunity for a fresh outlook on conversation design through an increased focus on contact center operational excellence and migration of their IVR off of Nuance. From thought leadership, strategy, design, engineering, and testing, PwC and I supported every step of the process. Through this incredible partnership, we migrated from Nuance to Kore, completely redesigned high-touch flows, retooled NLP training from the ground up, and built a strong enablement program for lasting results beyond the engagement.
Capabilities leveraged
Experience
Content strategy
Flow design
Integration logic
Transcript analysis
Language
Model tuning
Quality assurance
Product
User research
Roadmap planning
Organization
Design operations
Challenges addressed
01
Inadequate intent and speech recognition: Limited accuracy in understanding caller inputs, leading to suboptimal containment rates.
02
Incomplete deflection-driven strategies: Reliance on explain-and-deflect patterns without robust segue support, driving escalations and hindering collaborative interactions.
03
Ambiguous or outdated copy: Highly technical or imprecise language diminishing user trust and sentiment.
04
Gaps in integrations: Significant channel shift and elevated escalation triggers stemming from insufficient or nonexistent task integrations such as device installation help.
05
Dispersed model development: Rapid, decentralized model training fueling overlapping intents, conflicting patterns, and increased maintenance overhead.
06
Unoptimized voice prompts: Unnecessarily verbose or loosely structured voice prompts elevating cognitive load and reducing customer satisfaction.
Timeline
Activation logic steps
Talkthrough
Offline port deflection
$18M
In new cost savings, for a total of $183M.
5%
In additional IVR containment, for a total of 51%.
86.5M
Yearly IVR call volume.