ASMR an Accenture Federal Service Company – “We determined to leave our ShoreTel deployment behind and migrate to a cloud solution that includes Microsoft Skype for Business, but we still needed a call center to replace our ShoreTel Workgroup application and we felt AWS Connect could do the job”, said Manager Jason Agcaoili of ASM Research a Washington D.C. area Government service organization.
The ShoreTel workgroups, something we typically describe as a “poor man’s” call center, provided a lot of features ASM wanted to maintain. For example, the ability for the supervisors to see the status of each of their customer service queues. Agents also wanted to see the status of the rest of their team and supervisors need real time updates on the status of callers in queue.
These requirements were outside the AWS CCP soft phone, but they were critical to our ability to manage the various support teams in our deployment. We also need to create holiday schedules and matching customer greetings that would alert callers to why we were closed. Being able to close a specific queue for a team meeting was also on our “wish list”.
The architecture uses a single AWS Instance and over 10 DNIS numbers to manage and route incoming calls. The AWS Connect instance is integrated with Microsoft Skype for Business. To accomplish the list of requirements provided by ASMR, DrVoIP made use of of a full range of AWS Services that included:
- Kenesis Streams
- Agent Streams
- S3 Buckets
- AWS Connect Salesforce Connector
Core Features Developed with AWS services
Holiday Schedules and Ad-hoc Closings
To create a Holiday list of closing, a DynamoDB table was created. On incoming calls, a Lambda function was launched to check this list to note any holiday closings that might over ride the normal hours or operation. Additionally, a Supervisor is able to close a queue for a team meeting on a scheduled or demand basis. This closing would be checked by a Lambda function similar to the holiday check and would also retrieve the custom prompt created by the supervisor to play to callers during the closing period using Polly text to speech!
This is a graphical description of the current callers in queue. We use a Lambda function to update this list to both add and remove the caller. A modified “Agent Whisper” step is used to remove the caller from the list when the call is presented to an Agent.
Agent Monitoring and Coaching
FMG required that in addition to the standard AWS Connect feature of allowing a supervisor to monitor a call, DrVoIP had to create a facility to enable the supervisor to “coach” the agent by whispering in the agents ear. This was created using Kinesis media streams and custom “‘sockets” for Agent to Agent calling.
The ability to enable the supervisor to broadcast a text announcement to all Agents was required. This being a non-standard feature of AWS, was created by DrVoIP.
Recording retrieval and playback
Each Agent, granted the permission from a Supervisor was enabled to download and playback their own voice recordings. This non standard AWS feature was created using Media Streams and a custom GUI! Supervisors are enabled to download and playback all recordings. Recordings are stored in S3 buckets.
Option to offer Call Back
Using AWS Contact flows, DrVoIP created a call back option offered to callers waiting for an agent to become available. This option included the ability to turn this option off based on a schedule that is managed by the supervisor of that queue.
Agent to Agent Calling
Generally an Agent can call another agent only when the agent had an active phone call. DrVoIP implemented an option that enables agent to agent calling without requiring an active phone call. All Agents can see the status fo their team mates through a custom Agent dashboard.
Text Notification to “on call” after hours
Using a combination of Lambda, DynamoDB, Pinpoint and SNS callers into the center after hours were offered the option to request a call back from an on call representative. The representative was drawn from the DynamoDB table, contacted via Pinpoint and SNS and provided with the details of the inbound caller using contact flow attributes!
Searchable Disposition Codes
At the end of each phone call, the Agent is presented with a drop down list of “disposition codes” that can tag the phone call for later retrieval!
Real Time Reporting
Each Agent desktop had a browser based Agent Desktop that shows the status of each Queue that Agent is a member of in addition to a Global tab that shows statistics for all Queues in the deployment.
The AWS Connect dashboard has access to a wide variety of “historical reports” that can be configured and then scheduled to be run as required to meet the requirements of the stakeholders. For a detailed list of available reporting elements click here!
Platform Migration Strategy
DrVoIP Recommends a “warm port” strategy when migrating from one platform to another. Warm porting is the use of temporary phone numbers on the AWS Connect instance that are used during the design, deployment, testing and “go live”. This enables us to “call forward” existing DNIS numbers to the appropriate number on the AWS platform. In this way, the old system now becomes a “fall back” option should it become necessary for any reason. Later, a port date is established with the carrier currently handling the DNIS numbers and when that port can be verified the new platform is already in place and this numbers can then be swapped out with no service interruption.
Brian Cox, IT Director – DrVoIP has been taken a very hands on approach to implementing our needs as we work to develop a replacement helpdesk solution. Our Helpdesk staff is very particular in their demands and DrVoIP has shown great flexibility in producing or adapting solutions to meet these. Additionally they provide regular status updates are always available to hop on a conference call to hash out any issues. Additionally I feel their knowledge of both the AWS product as well as our previous product (Shoretel) has been a great asset in easing the transition, as they already know how our current solution works and what we will want from AWS in regards to mimicking certain features. Click to Download Case Study Details