Presentation is loading. Please wait.

Presentation is loading. Please wait.

WEBINAR Fight Fraud And Offer A Better Customer Experience With Voice Biometrics May 18, Call in at 1:55 p.m. Eastern time Andras Cser, Vice President,

Similar presentations


Presentation on theme: "WEBINAR Fight Fraud And Offer A Better Customer Experience With Voice Biometrics May 18, Call in at 1:55 p.m. Eastern time Andras Cser, Vice President,"— Presentation transcript:

1

2 WEBINAR Fight Fraud And Offer A Better Customer Experience With Voice Biometrics May 18, 2015. Call in at 1:55 p.m. Eastern time Andras Cser, Vice President, Principal Analyst Heidi Shey, Analyst

3 © 2015 Forrester Research, Inc. Reproduction Prohibited3 We work with business and technology leaders to develop customer-obsessed strategies that drive growth.

4 © 2015 Forrester Research, Inc. Reproduction Prohibited4 Agenda › Voice biometrics 101: why now, expected benefits, and how it works › Active versus passive voice biometrics › Voice biometrics vendor landscape

5 © 2015 Forrester Research, Inc. Reproduction Prohibited5 Voice biometrics today › Authenticates identity › Detects fraudsters › Streamlines automation › Creates differentiated customer experiences My voice is my password.

6 © 2015 Forrester Research, Inc. Reproduction Prohibited6 Voice biometrics today › Authenticates identity › Detects fraudsters › Streamlines automation › Creates differentiated customer experiences My voice is my password. Speech recognition Active and passive and text- dependent and text-independent authentication Known fraudster identification

7 © 2015 Forrester Research, Inc. Reproduction Prohibited7 Why now? › Password frustrations at a tipping point › Mobile: out-of-band authentication › Mobile: built-in, inexpensive VB reader (microphone!) › FIDO Alliance’s UAF and U2F specifications › Password frustrations at a tipping point › Mobile: out-of-band authentication › Mobile: built-in, inexpensive VB reader (microphone!) › FIDO Alliance’s UAF and U2F specifications

8 © 2015 Forrester Research, Inc. Reproduction Prohibited8 Voice stands out among 2FA options 1.Easy to centralize and no special endpoint reader required 2.Passive VB is non-intrusive. 3.Revocable, re-recordable post breach 4.VB is what you are, do, and know. 5.Deploys easily at scale 6.Adapts to a variety of compliance and privacy requirements

9 © 2015 Forrester Research, Inc. Reproduction Prohibited9 Voice biometrics: a variety of uses Voice authentication Verifying identity Convenience Security Cost saving

10 © 2015 Forrester Research, Inc. Reproduction Prohibited10 Two major ways on how VB works for authentication Text- dependent, or active, VB authentication requires users to use one or more passphrases. Text- independent, or passive, VB authentication uses freeform audio to identify the user.

11 © 2015 Forrester Research, Inc. Reproduction Prohibited11 Passive VB works best with authentication › Shares the known fraudster and authentication voiceprint sample databases › Bypasses the known fraudster lookup if the user has already authenticated him- or herself › Integrates easily and naturally with speech recognition

12 Step 1: Tagging fraudsters’ voice recordings as “known fraudster” samples Step 2: Comparing every phone conversation’s customer sample to known fraudster samples How VB works for identifying known fraudsters

13 © 2015 Forrester Research, Inc. Reproduction Prohibited13 Active versus passive voice biometrics Pros and cons of each approach

14 © 2015 Forrester Research, Inc. Reproduction Prohibited14 Active (text-dependent) VB Small audio Mobile use Explicit verification High accuracy Explicit enrollment Obtrusive No fraudster ID ProsCons

15 © 2015 Forrester Research, Inc. Reproduction Prohibited15 Passive (text-independent) VB No passphrase Lower friction Fraudster detection Larger audio Lower scalability Less suitable for mobile Lower accuracy ProsCons

16 © 2015 Forrester Research, Inc. Reproduction Prohibited16 Two options for where to store the enrollment samples Server (centralized storage) Endpoint (PC, in a mobile application, or trusted execution environment)

17 © 2015 Forrester Research, Inc. Reproduction Prohibited17 Server-based voice print storage Cross-channel One enrollment High security Privacy challenges Slow performance No offline capability ProsCons

18 © 2015 Forrester Research, Inc. Reproduction Prohibited18 Endpoint-based voice-print storage Privacy Offline uses are possible. Added vulnerabilities Requires separate enrollment per device ProsCons

19 © 2015 Forrester Research, Inc. Reproduction Prohibited19 Voice biometrics market landscape Key vendors

20 © 2015 Forrester Research, Inc. Reproduction Prohibited20 Nine voice biometrics vendors to watch (Auraya Systems)

21 © 2015 Forrester Research, Inc. Reproduction Prohibited21 › Spoofing detection › Active and passive VB › Innovation plans for mobile, call center, and fraud detection

22 © 2015 Forrester Research, Inc. Reproduction Prohibited22 › Active and passive VB › Self-tuning and adaptation › Web API for integration

23 © 2015 Forrester Research, Inc. Reproduction Prohibited23 › Synchronizes web and phone sessions in a SaaS solution › Outbound call allows for analytics and data mining. › Plans to fuse with other types of biometrics and to improve cross-channel support

24 © 2015 Forrester Research, Inc. Reproduction Prohibited24 › Offers a software development kit (SDK) › Combined and integrated VB and facial recognition › Focuses on text-independent solutions that have low power and low memory requirements

25 © 2015 Forrester Research, Inc. Reproduction Prohibited25 › Unified VB platforms (mobile, Web, IVR, and call center) › Active and passive VB › Plans to develop new antispoofing algorithms and add device verification technologies

26 © 2015 Forrester Research, Inc. Reproduction Prohibited26 › Focuses on real-time tuning and high-definition audio processing › Workflow engine for integration with other authentication factors and biometrics

27 © 2015 Forrester Research, Inc. Reproduction Prohibited27 › Passive VB, allows en masse enrollments from historical call recordings › Augments scoring with predictive analytics › Can combine known fraudster ID with authentication

28 © 2015 Forrester Research, Inc. Reproduction Prohibited28 › Active and passive VB › Ability to add additional VB engines, integrates with other 2FA and biometrics › Mobile SDKs are available.

29 © 2015 Forrester Research, Inc. Reproduction Prohibited29 › On-premises and SaaS option available › SaaS solution available in English, Spanish, and Chinese › Very scalable solution and proven up to 1 billion calls per year and passes OWASP standards

30 © 2015 Forrester Research, Inc. Reproduction Prohibited30 Best practices for voice biometrics 1.Don’t initially mandate for all, but for accuracy, implement for at least 400 users. 2.Ensure easy enrollment from all channels; check privacy requirements about audio. 3.Enable mobile VB for authentication on multiple channels. 4.Make sure you have a live customer on the line.

31 Thank you forrester.com Andras Cser +1 617.613.6365 acser@forrester.com Heidi Shey +1 617.613.6076 hshey@forrester.com


Download ppt "WEBINAR Fight Fraud And Offer A Better Customer Experience With Voice Biometrics May 18, Call in at 1:55 p.m. Eastern time Andras Cser, Vice President,"

Similar presentations


Ads by Google