BANKING
Client
Emirates NBD is one of the largest banking groups in the Middle East, and aims to be the most innovative. It has about 9,000 staff.
Project partners were Avaya (US) and Servion Global Solutions (India).Program/Project
Introduce Natural Language Call Steering across all business lines and customer segments, as part of the “EVA voice-based virtual assistant” initiative.
Support English and Arabic speakers.My Role
User Interface Consultant/Mentor to the UI Design Lead.
My Responsibilities
Reviewed client’s 150-page Program Requirements Specification. Discussed potential issues with Nuance Senior Director and Regional Manager.
Supported UI Design Lead with preparation/presentation of Kick-off and Requirements workshop materials.
Mentored her with semantic tagging of c. 60,000 pilot and post-deployment caller utterances.
Acted as her wingman during several multi-day Tagging Review workshops in Dubai, UAE (United Arab Emirates) with client/partner staff, including Product Manager, Business Analysts, and senior call center staff.
Coordinated some more complex custom aspects, including:
> mechanism to support smarter call handling, based on customer’s account information
> negotiation of data interface between Nuance module and Servion wrapper code
Liaison with Servion Lead Business Analyst for any non-trivial technical/QA questions.> mechanism to support smarter call handling, based on customer’s account information
> negotiation of data interface between Nuance module and Servion wrapper code
Challenges & Opportunities
Some very strong personalities within the client organization. Meetings were sometimes lively.
Large stakeholder team, so workshops typically had 20+ attendees.
Due to communication challenges between parties during the pre-sales phase, client had an expectation of a Siri-like virtual agent, which was sub-optimal for a Call Steering application. This made persona discussions more complex than usual.
The UAE is a very culturally diverse nation, so English grammar needed to be adapted to accommodate a large Indian expatriate population, and Arabic grammar for a broad range of dialects. I worked closely with one of Nuance’s most experienced international Speech Scientists – his focus was mostly acoustic/language models, mine mostly semantic.
Semantic tagging of early caller requests,
in preparation for a client workshop
Analysis of early caller path and intent, split by Line of Business,
for presentation to Business Owners
Client/Nuance team,
following the final sign-off workshop