This deck covers the new IBM Voice Gateway product which introduces a next generate IVR system that is conversational and built on IBM Watson cognitive services. It can also transcribe calls in real time to enable agent assist type applications. Use capabilities like the Watson Conversation service, Speech To Text and Text To Speech, the IBM Voice Gateway is built on cloud native principles.
Transforming Data Streams with Kafka Connect: An Introduction to Single Messa...
Bring IBM Watson to your telephone
1. InterConnect
2017
Bring Watson to Your
Telephone: Introducing the
IBM WebSphere Connect
Voice Gateway for Watson
Tom Banks: IBM Offering Manager and Technical
Evangelist
Email: tom.banks@uk.ibm.com
1 3/22/17
Brian Pulito: IBM Senior Technical Staff Member
Email: brian_pulito@us.ibm.com
Twitter: @brianpulito
IBM Voice Gateway
Galeal Zino: Founder NetFoundry
Tata Communications
2. 2 3/22/17
Please note
IBM’s statements regarding its plans, directions, and intent
are subject to change or withdrawal without notice at IBM’s
sole discretion.
Information regarding potential future products is intended to
outline our general product direction and it should not be relied
on in making a purchasing decision.
The information mentioned regarding potential future products
is not a commitment, promise, or legal obligation to deliver
any material, code or functionality. Information about potential
future products may not be incorporated into any contract.
The development, release, and timing of any future features
or functionality described for our products remains at our sole
discretion.
Performance is based on measurements and projections
using standard IBM benchmarks in a controlled environment.
The actual throughput or performance that any user will
experience will vary depending upon many factors, including
considerations such as the amount of multiprogramming in
the user’s job stream, the I/O configuration, the storage
configuration, and the workload processed. Therefore, no
assurance can be given that an individual user will achieve
results similar to those stated here.
3. 3
Customer services today
Web based support is going through a cognitive bot revolution
• Making it cheaper and faster to handle common queries
• Handling queries 24 hours a day, 365 days a year
Call center challenges
• Cost of optimization
• Wait time reduction
• Improve customer satisfaction
• Easier access to internal documents and info
Why not take the cognitive chat bots we are using online, and
integrate them into telephone based support?
4. 4 3/22/17
Watson Voice Services Today
• No connectivity to telephony (no support for SIP and RTP)
• Text To Speech and Speech To Text only support Web
Sockets or REST
• No IVR integration capabilities (e.g. MRCP)
• No orchestration of services to provide a telephony voice
interface
• No integration with SIP trunking or enterprise Session Border
Controllers
5. 5 3/22/17
Bringing Watson to telephone based support
Bringing Watson to call centers will:
• Increase the number of calls that can be handled automatically
• Lower queue times, saves agents for complex calls
• Improve customer service experience and satisfaction
• If the customer’s inquiry or problem is resolved in the first call, only 1% of
those customers are at risk to go to your competitors
• Assist call center agents during a call
• Remove the need for agents to be experts on all topics
• Get useful info to agents when they need it, quickly and automatically
6. 6 3/22/17
The Value Proposition
• Call center agents take time to train, and have high turnover
• IVRs have 51% satisfaction with customers
• 7.3% of customers who use IVRs consider competitors after their experience
• Watson can provide a better customer experience while handling more
calls without needing to transfer out to a person
7. 7 3/22/17
Introducing the new IBM Voice Gateway
The IBM Voice Gateway for Watson will be released on March 24th, 2017
What is it?
• Next generation (Cognitive) Interactive Voice Response system
• Cloud native, Microservice solution (Docker images)
• Horizontally scalable
• Deployable to on premise networks or the cloud
• Provides a SIP endpoint for connecting to, and orchestrating between,
Watson services
8. 8 3/22/17
Customer calls
support
Watson Speech To Text, Text To
Speech and Conversation services
orchestrated by gateway to talk with
customer
Call routed to Voice
Gateway via SIP Trunk or
Session Boarder Controller
Transfer out to agent
when necessary
Use case 1: Cognitive Self-Service Agent
IBM Voice Gateway
9. 9 3/22/17
Self-Service Agent Demonstration
Text To Speech
Watson Services
Conversation
(Dialog, NLC)
Speech To Text
SIP Trunk
SIP Orchestrator
Media Relay
Twilio used
for demo
IBM Voice Gateway
WS
10. 10 3/22/17
Self-Service Agent: Single Turn Orchestration Flow
Text To Speech
Watson Services
Conversation
(Dialog, NLC)
Speech To Text
1) Caller Ask Question
2) Question streamed to STT over WS
3) Text Utterance Returned
4) Text sent as Ask
5) Ask response
6) Ask Response text sent to TTS
7) Synthesized audio returned
8) Audio response streamed to caller
IBM Voice Gateway
12. 12 3/22/17
Use case 2: Cognitive Agent Assistant
Customer calls support and is
connected with an agent
Speech To Text used to convert streams
to text then analytics used to provide
real-time feedback to agent
SBC forks media to the Voice
Gateway
SBC
Agent uses real-time cognitive
feedback to help customer
Watson
Services
Session Border
Controller
IBM Voice Gateway
13. Agent Assistant Demonstration
Watson
Speech To Text
Analytics via Watson
(NLC + Retrieve and Rank)
SIPREC
Customer (RTP / G.711)
Transcribed Text
published over MQTT
Web Browser
Agent
Desktop
MQTT
Broker
PowerMedia XMS
(Transcoding)
WebRTC Gateway
(WebSphere Liberty)
Agent
Agent (RTP / G.711)
Customer
WebRTC
Forking SBC
NetFoundry SIP
Trunk
IBM Voice Gateway
14. Self-Service Agents
Point of sale agent
Pre-screening agent
Credit checking agent
Support agent
Booking agent
Agent Assistants
Point of sale agent assistant
Real-Time trouble shooter
Agent coaches and trainers
Healthcare diagnostic assistant
Transcriptions
15. Omni-Channel Cognitive Agents
WhatsApp
Facebook Messenger
Slack
Twitter
Customers connect to BOTs
hosted in the Watson
Conversation Service
Watson
Conversation
Speech To Text, Dialog
and Text To Speech
services orchestrated to
talk and chat with customer
IBM Voice Gateway
16. Use case 3: Cognitive IVR
IVR
Customer calls
support
Watson Speech To Text, Text To
Speech and Conversation
services, integrated with IVR
application server
Call routed to Cognitive Gateway via
MRCP 2.0 from the existing IVR
provider
Transfer out to agent
when necessary
The MRCP 2.0 standard
allows for tight integration into
existing IVR solutions and
extends the life into the
cognitive world.
IVR providers like Avaya,
CISCO, Genesys and
Blueworx
MRCP
IVR
Gateway
18. IBM Voice Gateway Microservice Composition
Composed of two co-dependent Docker containers
SIP Orchestrator (SO)
Microservice
• Runs on WebSphere Liberty
• Handles all SIP signaling
• Watson Conversation orchestration
• Controls the media relay through a
web socket connection per call
• Acts as a SIP endpoint or a SIPREC
server depending on use case
Media Relay (MR)
Microservice
• Runs on Node.js
• Transcodes between RTP and raw
PCM (for Watson services)
• Orchestrates Speech To Text (STT)
and Text To Speech (TTS)
• Publishes text utterances to SO
• Synthesizes text utterance from SO
IBM Voice Gateway
19. SIP Trunks,
SBCs,
SIP Load Balancer,
SIP Endpoints,
Etc.
Media Relay Container
(ports 16384-32767)
SIP Orchestrator Container
(ports 5060 or 5061)
Offer SDP Answer SDP
Play Text Transcoded Text
Web Socket API
Web Socket API
Node.js
WebSphere Liberty
SIP/SIPREC
RTP/G.711
Web Socket APIs
1919
Microservices
Reporting
(Splunk, REST)
Transcriptions to
MQTT Broker
Watson
Conversation
Service
Watson Speech
To Text Service
Watson Text To
Speech Service
REST APIs
MQTT
Microservice Architecture
IBM Voice Gateway
20. 20 3/22/17
Customizing the Voice Gateway when acting as a self-service agent
Gateway API exposed to the Watson Conversation service for:
• Call control (hangup, transfer)
• SIP header access for logging (Request, To and From URIs, etc.)
• Speech control (switch language, STT confidence and alternates)
• Music on hold and audio wav file playback to caller
• Error message handling (transfers, error utterances)
• Contact center integration
State variables (prefixed with vgw) in JSON context
State variables (prefixed with vgw) in JSON context
Examples include vgwHangup, vgwMusicOnHoldURL
Examples include vgwHangupReason, vgwBargeInOccurred
Watson ConversationIBM Voice Gateway
21. 21 3/22/17
Reporting from the Voice Gateway
Reporting events published from the voice gateway
REST server
(Splunk HEC)
Conversation Turn Events *
Splunk event headers
Input text
Output text
Intents
Entities
Context state variables
Conversation ID
Turn counter
And more…
* JSON from Conversation response
Call Detail Record Events
Splunk event headers
Start and Stop time
Call length
SIP details (e.g. SipTo, SipFrom)
Number of turns
Number of barge ins
Watson transaction latencies
Intents details (#, list of all)
And more…
IBM Voice Gateway
22. 22 3/22/17
Securing the IBM Voice Gateway
The voice gateway supports the following:
• Connecting to Watson over TLS
• Connecting to Watson via a forwarding proxy
• Mutual authentication for REST and Web Sockets
• Supports for SIPS
• Supports trust and key store configuration
• Default trust store with certs for Watson connectivity
IBM Voice Gateway
23. HA using Kubernetes (k8s)
• Kubernetes used for container orchestration and scheduling
• A Voice Gateway pod consist of a single SIP Orchestrator and Media Relay container
• Pod deployed in Net Host mode to expose VM IP addresses and ports
• Only one Voice Gateway Pod per node (enforced with k8s configuration)
• External SIP Load Balancing of k8s pods
• SIP LB statically updated with POD address details
• POD must be queisced of all active calls before taking out of service
• Sample kubectrl scripts available here:
• https://github.com/WASdev/sample.voice.gateway.for.watson
• IBM Spectrum Conductor for Containers tested
• Recommended for an on premise Docker substrate
23
24. IBM Voice Gateway HA using Kubernetes
24
Etcd
Config Management
k8s APIServer
Kube Master
Scheduler
WVG POD
Media Relay
SIP
Orchestrator
WVG POD
Media Relay
SIP
Orchestrator
WVG POD
Media Relay
SIP
Orchestrator
Session Border Controller / External SIP Load Balancer SIP signalling
RTP Media
VM / Node VM / Node VM / Node
IBM Voice Gateway
25. IBM Spectrum Conductor for Containers
3/22/17 25
Distributed Key -
value
Install
Proxy
Logging/Reporting
CI/CD
Tenant
management
Network
Authentication
Authorization
Persistent
Volume
GUI
Image Registry
Service
Discovery
Service Load
Balancing
Trouble Shooting
HA
Open Source + Integration + Value Add
Resource
Management
LSF Community
Edition
Application Catalog
27. 27 3/22/17
Enterprise integration requirements
• Integrate with contact centers and enterprise telephony systems
• Compliance requirements (e.g. HIPAA, PHI, PCI and PII)
• Integration with existing Systems of Record
• Voice analytics
• Voice security (e.g. hybrid connectivity, authentication, etc.)
• De-identification of Watson Conversation requests
28. Enterprise Telephony Integration with Watson
Conversation
(Dialog + NLC)
Speech To Text
Text To Speech
Watson Services
Service Orchestration
Engine (SOE)
Customer Specific APIs
(for PHI, PCI and PII)
Web Socket over TLS
Web Socket over TLS
REST over TLS
SessionBorderController
Contact Center / IVR
Hybrid-Cloud Solution
Pattern
Enterprise Network
IBMVoiceGateway
MPLS,
NetFoundry,
Public Internet
29. 29 3/22/17
Service Orchestration Engine (SOE) Sample
Conversation Ask Request
Watson
ConversationIBM Voice Gateway
Custom API
Services
Service
Orchestration
Engine
Modified Ask Request
Conversation Ask ResponseModified Ask Response
Why most deployments will need an SOE:
• De-identification of Watson Conversation requests
• Personalize the Watson Conversation responses
• Use telephony features like Caller-id, DTMF, etc.
• Integrate with API’s to enhance the user interaction
• Exploit Voice Gateway features using state variables
• Voice security integration using DTMF or biometrics
SOE Banking Sample Application shows:
1. How to call API’s to enhance WCS responses
2. How DTMF can be used to process user input
3. How to log the interactions to WCS
4. How to approach building your own SOE
5. How to enable user security
6. How to personalize the interactions
https://github.com/WASdev/sample.voice.gateway/tree/master/soe
30. 30 3/22/17
IVR and Contact Center Integration
Voice gateway integrates with existing IVRs and
Contact Centers using SIP initiated call transfers:
• Transfer to voice gateway from IVR
• Transfer to Contact Center from voice gateway
Session Border Controllers can handle SIP REFER
messages and perform the transfer:
• SBCs catch refer and perform the transfer
Agent dashboards integrate via:
• Metadata exchanged in SIP signaling
• Conversation history published from an SOE
Session
Border
Controller
IBM Voice Gateway
Contact
Center
1
23
4
1. Call arrives at SBC
2. SBC routes call to Voice gateway
3. Customer request transfer and a
SIP REFER is sent to SBC.
4. SBC re-routes the call to an agent
through the Contact Center.
32. Global, Turnkey Solution for Enterprises and Contact Centers
to Leverage Watson Voice Cognitive Services
Voice from
anywhere
Interoperate PSTN & SIP voice to Watson
Secure, reliable delivery to Watson over Internet
Watson Cognitive
Services
https://www.NetFoundry.io
IBM Voice Gateway
33. 33 3/22/17
Details on the new NetFoundry AppWAN
TM
for Watson
The NetFoundry AppWANTM for Watson was released on March 20th, 2017
What is it?
• Connects business and contact center sites to Watson Services
• Integrates with the IBM Voice Gateway for turnkey Watson interop
• Cloud native, microservice solution
• Horizontally scalable and deployable on premises
https://www.NetFoundry.io
34. The IBM Voice Gateway provides Watson
interoperability, saving enterprises and contact
centers from needing expensive telephony
interworking projects to consume Watson services.
Make it Easy for Customers to Consume Watson Services
NetFoundry connects sites to Watson with top
security & reliability over software defined
networks, accessed from any Internet connection,
enabling customers to connect to Watson without
multi-month, expensive MPLS extension projects
IBM Voice Gateway
35. Enterprise Telephony Integration with Watson
Conversation
(Dialog + NLC)
Speech To Text
Text To Speech
Watson Services
Watson Custom Proxy
Customer Specific APIs
(for PHI, PCI and PII)
SessionBorderController
Contact Center / IVR Enterprise Network
Hybrid-Cloud Solution
Pattern
Enterprise Network
NetFoundryAppWANTM
GatewayForWatson
NetFoundryAppWANTM
GatewayForWatson
IBMVoiceGateway
This shows the hybrid cloud pattern. IBM Voice and NetFoundry gateways install on-premises or in the
cloud as Docker containers or VMs, scaling horizontally with agility, reliability and DevOps efficiency
36. 36 3/22/17
Self-Service Agent in Cloud Hosted Deployment Pattern
Text To Speech
Conversation
(Dialog, NLC)
Speech To Text
Watson
Services
IBM Data Center
IBM Voice Gateway
Enterprise
Network
NetFoundryAppWANTM
GatewayForWatson
Session
BorderController
NetFoundryAppWANTM
GatewayForWatson
RTP
SIP
Contact
Center
37. IBM Voice Gateway
Use Case #2: Cognitive Agent Assistant in Hybrid Pattern
Watson
Speech To Text
Analytics via
Watson
(NLC + Retrieve
and Rank)
SIPREC
Customer (RTP / G.711)
Transcribed Text
published over MQTT
Web Browser
Agent
Desktop
MQTT
Broker
Forking SBC
Agent
Agent (RTP / G.711)
NetFoundry
SIP Trunk
Customer
NetFoundry SDN
over Internet
Watson assists
the contact
center agent in
serving the
customer
NetFoundry encrypts all data, blocks all unauthenticated data from
entering customer and cloud sites, and improves performance
Contact Center Environment
NetFoundry
AppWANTM
for Watson
gateway
NetFoundry
AppWANTM
for Watson
gateway
38. Security
As demonstrated by
Mirai and similar attacks,
security is increasingly
the top challenge of a
hyperconnected world
TCO
It is expensive and time
consuming to extend
private networks and
infrastructure to the
Watson cloud
Performance
Most security solutions
inhibit performance
whereas real-time
Watson services require
top performance
3 Key Challenges for Watson Connectivity
39. Security
Makes Watson customer
networks “dark” to
security threats. Encrypt
and secure the Watson
data across any network.
TCO
Eliminates private
network and telco CPE
costs. Turnkey solution
for SIP/telephony
interoperability
Performance
Delivers enterprise grade
reliability and
performance for Watson
interworking
NetFoundry + IBM Voice Gateway
Helps Solve the 3 Key Challenges for Watson Connectivity
IBM Voice Gateway
40. NetFoundry AppWANTM enables cloud services to leverage service-specific
software defined networks over any Internet access providers. The result is
businesses securely and reliably take cloud services from prototype to
production, without network and infrastructure build-out delays and costs.
• “Partnering with NetFoundry gives our joint customers greater options for delivering quality of
experience across the Internet,” said Sanjay Uppal, CEO, VeloCloud. “NetFoundry gives
businesses improved performance and security across the Internet.”
• “We are seeing an increased demand from our customers for greater security, control, and
network configurability for their IoT private networks,” stated Bryan Lubel, President of
Integron. “Our partnership with NetFoundry allows us to satisfy those requirements.”
https://www.netfoundry.io
NetFoundry Overview and Testimonials
42. 42 3/22/17
Many companies are already working with the Voice Gateway
• Many POCs underway (including production usage)
• Many minutes processed
• Early prototypes have shown a 30% advantage over traditional IVRs
to completely handle a call
• POCs underway in many different languages
43. 43 3/22/17
How to get started
• Link to Docker images in DockerHub (developer only license)
• https://hub.docker.com/r/ibmcom/voice-gateway-so/
• https://hub.docker.com/r/ibmcom/voice-gateway-mr/
• Link to documentation in Knowledge Center
• https://ibm.biz/VoiceGatewayDoc
• Link to samples in public GitHub Repository
• https://github.com/WASdev/sample.voice.gateway
• Get questions answered at dwAnswers forum
• https://developer.ibm.com/answers/topics/voicegateway/
• Marketplace link (includes video demonstration)
• https://www.ibm.com/us-en/marketplace/voice-gateway
44. 44 3/22/17
Become part of our community on Slack!
Chat on Slack
Then join the #ibmvoicegateway channel here:
https://ibm-cloud-tech.slack.com/messages/ibmvoicegateway
First register with the ibm-cloud-tech team here:
http://ibm.co/2mblgXr
45. 45 3/22/17
Conclusion: Why Cognitive Voice Agents
Better than traditional IVRs
• Understands natural language, agents are conversational
• Fewer opt-outs means lower contact center cost
• Can handle deeper, more complex conversations
• Can handle multiple intents in a single call
Deep Analytics lead to improved KPIs
• All aspects of system contribute to analytic processing. For example:
• Barge-in correlation points to areas to focus training and improvements
• Opt-outs correlated to points in conversation lead to improved flows
48. 48 3/22/17
Notices and disclaimers
continued
Information concerning non-IBM products was obtained from the
suppliers of those products, their published announcements or other
publicly available sources. IBM has not tested those products in
connection with this publication and cannot confirm the accuracy of
performance, compatibility or any other claims related to non-IBM
products. Questions on the capabilities of non-IBM products should be
addressed to the suppliers of those products. IBM does not warrant the
quality of any third-party products, or the ability of any such third-party
products to interoperate with IBM’s products. IBM expressly disclaims
all warranties, expressed or implied, including but not limited to, the
implied warranties of merchantability and fitness for a particular,
purpose.
The provision of the information contained herein is not intended to, and
does not, grant any right or license under any IBM patents, copyrights,
trademarks or other intellectual property right.
IBM, the IBM logo, ibm.com, Aspera®
, Bluemix, Blueworks Live, CICS,
Clearcase, Cognos®
, DOORS®
, Emptoris®
, Enterprise Document
Management System™, FASP®
, FileNet®
, Global Business Services®
,
Global Technology Services®
, IBM ExperienceOne™, IBM SmartCloud®
,
IBM Social Business®
, Information on Demand, ILOG, Maximo®
,
MQIntegrator®
, MQSeries®
, Netcool®
, OMEGAMON, OpenPower,
PureAnalytics™, PureApplication®
, pureCluster™, PureCoverage®
,
PureData®
, PureExperience®
, PureFlex®
, pureQuery®
, pureScale®
,
PureSystems®
, QRadar®
, Rational®
, Rhapsody®
, Smarter Commerce®
,
SoDA, SPSS, Sterling Commerce®
, StoredIQ, Tealeaf®
, Tivoli®
Trusteer®
,
Unica®
, urban{code}®
, Watson, WebSphere®
, Worklight®
, X-Force®
and
System z® Z/OS, are trademarks of International Business Machines
Corporation, registered in many jurisdictions worldwide. Other product
and service names might be trademarks of IBM or other companies. A
current list of IBM trademarks is available on the Web at "Copyright and
trademark information" at: www.ibm.com/legal/copytrade.shtml.