Voice Biometrics London 2008:

spleenypuddleSecurity

Nov 29, 2013 (3 years and 7 months ago)

68 views

Voice Biometrics London 2008:
Planning Solutions for the Future
Dan Miller
Senior Analyst
Opus Research
©
2008 Opus Research, Inc.
Opus Research Background

Founded in 1985 
Market analysis and
strategic planning

First study of voice
biometrics in 2003

Track global deployments
and vendors

Conversational Access
Technologies (CAT) 
Morphing into Contextual
Communications
Voice Biometrics 2008: Planning
Solutions for the Future
2
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 3
Welcome to VBC2008 
Objectives 
Define the power of multiples

Share implementation experience

Scout out solutions

Get answers to your questions

Network

Todays Agenda
Introduction-Setting the stage

Focus on the user experience

Look at the big picture: Identity Management

Discuss experience in financial services

Networking reception
©
2008 Opus Research, Inc.
Page 4
My Objectives

Understand Voice Biometrics 
Definition of Terms

Market size, forecast and drivers

Contrast with other biometrics

Implementation Issues
Alternative approaches

New entrants

Integration issues
Voice Biometrics 2008: Planning
Solutions for the Future
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 4
Voice Biometrics and Speaker
Verification 
Voice Biometrics is a technology

Captures an utterance from a live caller

Compares it to previously stored voiceprint

Produces a score

Speaker Verification is an application

Employs a biometric engine plus business logic

Enrolls customers by obtaining voice prints

Compares live utterances to voice prints to
produce a passor failresponses
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 6
Components of Robust Speaker
Verification 
Core Verification Engine
Receives voice sample (utterance); compares it to
a voiceprint (template)

Confirms who said it

Core Recognition Engine
Compares utterance to ASR grammar

Determines what was said

Business Logic
Decides if the caller passes or fails

Dictates required next steps
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 7
What is a Voice Print?
Physical Characteristics The unique physical traits of the
individuals vocal tract, such as
shape and size.
Behavioral Characteristics The harmonic and resonant
frequencies, such as accents, the
speed of your speech, and how
words are pronounced and
emphasized.
Voiceprint -
Together these physiological and behavioral
factors combine to produce unique voice patterns for every
individual
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 8
Speaker Verification vs. Speaker
Identification
For Verification:

User claims an ID

Application matches voiceprint to that claim

For Identification:

No claim of identity

ID System tries to detect closest matchof
captured utterances to voiceprint from a
population of registered users
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 9
Text Dependent vs. Text Independent 
Applications that require a specific pass
phrase are Text Dependent

Require training

Customarily involve enrollment

Text Independent applications can
use any utterance
Simplify enrollment

Support conversational authentication
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 10
Evaluating Biometric Effectiveness
Threshold (Sensitivity)
Errors
F
a
l
s
e

R
e
j
e
c
t

R
a
t
e

(
F
R
R
)
F
a
l
s
e

A
c
c
e
p
t

R
a
t
e

(
F
A
R
)
Equal Error Rate (EER)
Two Critical Criterion
:

False Accept: Grants access
to imposters

False Reject: Denies access to
legitimate users
Equal Error Rate 
Measure of overall
effectiveness

Used for comparative purposes
Failure to Enroll (FTE) 
Percentage of prospective
users unable to join
©
2008 Opus Research, Inc.
The rest of a complete solution 
Involve other factors and infrastructure
One-time passwords

Knowledge-based authentication

Physical tokens

Compliance and mandates
Internal: rules, policies and procedures

External: Industry specific mandates and
standards
Voice Biometrics 2008: Planning
Solutions for the Future
Page 11
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 12
Where Voice Biometrics Fit 
PWR (Password Reset)
Tens of thousands served

Proven ROI

Strong proof of conceptIn the intelligent
customer front door

How may I help you?

Tell me your pass phrase

Point solutions
©
2008 Opus Research, Inc.
Biometrics in Everyday Life
Voice Biometrics 2008: Planning
Solutions for the Future
Page 13
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 14
In these contexts, voice is:

Tokenlessauthentication

ID proofing

PIN replacement

Voice signatures

Spoken tokens
©
2008 Opus Research, Inc.
Improving user experience
Voice Biometrics 2008: Planning
Solutions for the Future
Page 15
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 16

Strengthens existing, weak security practices

Embodies additional factor (something you are)

Cant be lost or stolen

Does not require private information

Can be combined with other technologies

Can get stronger with use of random phrasing to
counter replay attacks

Ripe for retrofitprojects
The voice biometricedge
©
2008 Opus Research, Inc.
Global market for biometrics

Steep assent (per
ABI)

Broad mix of
markets

Narrow set of
applications
Voice Biometrics 2008: Intro
and Tutorial
17
©
2008 Opus Research, Inc.
The alternative biometrics 
Fingerprint

Face

Hand

Iris

Voice

Signature
Voice Biometrics 2008: Intro
and Tutorial
18
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 19
Building the Case for Voice 
Well suited for remote authentication
Suitable for any telephone

No specialized client hardware required

Supports mobility

Cost effective to implement
Leverages Web and IVR investments

Integrates with existing authentication
infrastructure

User friendly
Especially text independent

Requires no behavioral changes
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 20
The Inflection Points
1980s-90s20002010
Novel technology
Point solutions
Improved engines
Integrations with
security policy
Conformance with
Compliance Issues
Hooks into security
Standards across
modalities
Attention to user
End-to-end approach
Text-independence
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 21
as Others See Us

Hype Cycle 2005
Biometric User ID
more than 10
years away

Beyond the
Peak of Inflated
Expectations

Just behind
Corporate
Blogging
©
2008 Opus Research, Inc.
This years model

Voice not on
the curve

Biometric
payments are

Speech recfor
mobile
approaching
peak
Voice Biometrics 2008: Planning
Solutions for the Future
Page 22
©
2008 Opus Research, Inc.
Forecast for voice biometrics
Voice Biometrics 2008: Planning
Solutions for the Future
Page 23

Licenses

Services

Transactions

No more
hockey
stick
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 24
State of the industry 
Over 3 million registered
Quiet, policy-driven adoption

Three times as many people have iPhones

Customer-facing just starting
BellCanada approaching 1 million enrollees

Many commercial banks on hold

Mobile commerce should be a boon

Govt, healthcare, payment cards next
©
2008 Opus Research, Inc.
Voice Biometrics 2008:
Addressing Implementation Issues
25
Strategic Objective

Balance:
Security

User
experience

Cost
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 26
Tactical imperative 
Leverage existing
infrastructure
Help desks

Contact centers

Data centers

Policy engines

Security

Web services
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 27
Haalp! Verbatim challenge 
Security Technologies include (among
other things):
Matrix Cards

Tokens and integrated token cards

Risk-based authentication models for self-
service channels

Voice biometrics for call center and
telephone banking

Fingerprint and vein biometrics for branch
banking
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 28
Making best efforts 
To replace or retrofit existing systems
Compliance: multifactor mandate

Convenience: anytime, anyplace any phone

Cost-savings: Works for PWR!

To condition the market
Familiarize enterprise employees

Train, involve integrators

Reinforce new ways of doing things
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 29
Best efforts (continued)

Block and tackle
Highlight importance to customers

Demonstrate suitability

Encourage enrollment

Sustain trust
Enlist recognized integrators

Demonstrate longevity

Raise profile of solutions
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 30
Marketing challenges

A technology is not a product
Packaging is crucial

Support a must

Integration required

A product is not a solution
What problem(s) are we solving?

What are the alternatives?

Why should customers switch?
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 31
Addressing User Experience Issues 
End-to-end approach
Simplified enrollment

Component of the front door

Integrated with existing callflows

Conversational authentication
Text-independent or familiar phrase

Agent-assisted

Simplified
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
32
What needs to be proven 
For the enterprise decisionmakers
Efficacy of the technology

Life-expectancy of solutions providers

Lack of disruption/ease of deployment

For end-users
Trust in solution

Ease of use

Convenience
©
2008 Opus Research, Inc.
Voice Biometrics 2008: Planning
Solutions for the Future
Page 33
Which is why we are here 
Beyond Basics
Why ID proofing is so important

How applications scale

What strategies are working

What comprises an end-to-endapproach

To Implementation
Get to know prospects and partners

Build the ecosystem
©
2008 Opus Research, Inc.

Questions?
Voice Biometrics 2008: Planning
Solutions for the Future
Page 34