- Integrating Amazon Connect call flows with your CRM.
- Leveraging Machine Data to derive insights from Amazon Connect derived data.
- Make your call center agent’s lives easier by providing relevant information to them ahead of the call.
- Call your Bitcoin broker with the click of a single button. Literally.
PRESENTATION BY:
- Patrick Hannah | VP of Engineering @ CloudHesive
5. Amazon Connect Overview
• ACD-as-a-service
• ProvisionToll Free orToll Numbers
• Port in/port out existing numbers
• Outbound Caller ID Name + Number
• Pre-configurable speed dial numbers
• Environment ready within minutes from provisioning (regardless of size)
• Near-infinite capacity
• Simple pricing:
• Per minute, per leg (Inbound from customer, outbound to agent, platform usage)
• Per day, per number
• Supports Data encryption at rest with Key Management Service (KMS)
• Reports, Recordings
• Supports Data encryption in flight with customer supplied key
• Credit-card-in-IVR-collection
6. Amazon Connect Users
• Integrates with your existing technology footprint
• Active Directory (or use Directory Service)
• User provisioning requires less than 30 seconds of administrator time (point and click), can
be automated through API and supports bulk provisioning
• Connectivity
• WebRTC stream can be routed over Direct Connect
• Granular role/permission assignment; Does not need to conform
Traditional Administrator/Supervisor/Agent roles
• Empower your business to run their business
• Supports hierarchies (Supervisor, Lead,Tier 3,Tier 2,Tier 1, etc.) for
reporting purposes
• Administrative Changes are Logged
• Agent Status Changes are Logged
7. Amazon Connect Queues
• Supports Hours of Operations
• Define a maximum number of calls for overflow capabilities
• Supports prioritization and skill based routing (routing profile)
• Supports custom statuses (On Break, etc.) for reporting and routing
purposes
8. Amazon Connect Metrics
• Per Call
• CustomerTrace Record (CTR)
• Call Recording
• Sent to S3
• Administrator/Supervisor
• Trailing since midnight, Past n hours
• 30 minute granularity
• Built in reports/dashboards
• Real time + Historical
• Detailed data extracts
• CSV to S3, transferred hourly
• Basic
• CloudWatch
• Per Agent
• Agent Event
• Streams API
9. Amazon Connect Contact Flows
• IVR-as-a-service
• Each leg of the call can be influenced:
• What caller first hears
• What caller hears when in queue
• What caller hears when connecting to an agent
• What agent hears when connecting to a caller
• Enable/disable recording of caller/agent legs
• Play custom audio or text to speech (numerous languages and voices)
• Menus + Digit collection
• Percentage allocation, look ahead routing, error handling, timeout handling, default routing
• Speech to text/natural language processing (via Lex)
• Programmatic functions (via Lambda) – perform real time data lookup against nearly anything, based on
call data, pass results back to the call
• Pass data to Agent (via JavaScript API), CTR (via Kinesis) or lookup historically via
Administrator/Supervisory view
10. Interfaces
• Initiators
• Alexa Skill (API: Call Control + Data Stream)
• Alexa Calling (Literally, call)
• Click-to-Call (API: Call Control + Data Stream)
• Buttons/Sensors (API: Call Control to Dummy Number + Data Stream)
• Apps (API: Call Control + Data Stream)
• Phone (Literally, a call)
• Once initiated interact via:
• DTMF
• Voice (Bot)
• Voice (Agent)
• What this allows us to do:
• Combine modes to start an interaction out on a device, move to voice and ultimately move to an agent
13. Conclusion
• Moving a Contact Center, while a seemingly arduous task can be done on a
shortened timeframe (6-8 weeks for 800 users) with minimal hard cost (pay per
use)
• Adding additional interfaces (IOT originated contacts, as an example) can be
done in a componentized manner, again with minimal hard cost
• And finally, adding ML based services, such asTranslate can be done with
minimal heavy lifting, adding near real time voice analysis to your call recordings