Case Study Contact Center at Scale
A Case Study from the Hospitality Industry
In need of modernization, a hospitality company turns to Digiclarity.
In 2017 one of the most respected international hotel and resort brands found itself facing a series of telephony issues. While its competitors had integrated newer technologies such as voice recognition and AI bots into their systems, they had yet to innovate beyond push-button DTMF.
Updating this older system required coding support from its call-center provider. Unfortunately, disruption to the provider’s services at the time reduced the necessary technical support for making this change. This not only impacted their ability to drive business requests, it also jeopardized the speed with which any sort of change could go to market.
As a trusted partner since 2012, Digiclarity was retained to address their urgent telephony needs. We were tasked with extending our successful track record with its business by ushering the company’s worldwide call-center infrastructure into the modern age as quickly as possible.
Digiclarity Taps into the Power of Amazon Web Services
Their need for a change in direction coincided with AWS’s release of Amazon Connect. The cloud-based service utilized the same call-center platform used by the online retailer’s own expansive customer-service team. We immediately recognized this as a strong potential solution for their 5,000+ call-center agents around the world.
Our methodology for migrating such platforms follows a logical sequence designed to cause as little disruption to the system as possible:
We first determine the minimum viable product with the necessary complexity for a full test. This lets us operate more nimbly and with less risk than if we were to dry run with larger teams or departments.
As employees adjust to any new system, they will naturally have questions about capabilities and need help troubleshooting unfamiliar issues. By immediately training help-desk support after sufficient testing, we create a safety valve for all agents and therefore smooth the transition significantly.
After a successful test case and training session with the organization’s help desk, we scale the migration across the organization. This can include thousands of employees around the world depending on the business.
Having fully scaled, we then move into automation, data analysis, machine learning, and AI. At this point we are able to offer numerous benefits, such as customer profiling for easier desk routing based on call history.
Following this approach, we partnered with them on a pilot program to evaluate the speed and effectiveness of Amazon Connect integration. They chose a specific support desk for the pilot given the wide range of requests the desk received relative to its small size, as well as its requirement that agents be all over the globe, allowing us to test audio latency and response times. This allowed us to sufficiently determine the system’s capabilities while keeping the transfer limited in scope as a test run.
The program was by all accounts a major success. Migration of the desk onto Amazon Connect improved reporting and offered a phone panel specifically tailored to the desk’s needs. Most importantly, back-end administrators and developers required no niche technical training thanks to Amazon Connect’s user-friendly graphical interface.
They subsequently green-lit conversion of all call-center desks worldwide in 2018.
Case Study at a Glance
-
Client One of the world's largest hospitality providers offering 5,800+ properties and 939,000+ rooms across 17 brands in 114 countries.
-
Timeline January 1 - December 8, 2018
-
Challenge Transition global call center to Amazon Connect at scale.
-
Solution Custom Reservations and Customer Care software.
-
AWS Services Amazon Connect, Kinesis, DynamoDB, Aurora, Redshift, EKS, AWS Lambda
-
Results 5,000+ global call-center agents speaking 13 languages on 6,500+ inbound phone lines transitioned; Migration in under 1 year; $430,000 in annual cost savings.
Architecture used for processing CTRs at scale (click to enlarge):
