Advanced Speech and Language Technology for Complex Customer Care - - PowerPoint PPT Presentation

▶
advanced speech and language technology for complex
SMART_READER_LITE
LIVE PREVIEW

Advanced Speech and Language Technology for Complex Customer Care - - PowerPoint PPT Presentation

Oct 31, 2023 599 likes •785 views

Advanced Speech and Language Technology for Complex Customer Care Automation and Self-Service Roberto Pieraccini Chief Technology Officer SpeechCycle, Inc. 26 Broadway, 11 th Floor New York, NY 10004 roberto@speechcycle.com What is

slide-1
SLIDE 1

Advanced Speech and Language Technology for Complex Customer Care Automation and Self-Service

Roberto Pieraccini Chief Technology Officer SpeechCycle, Inc. 26 Broadway, 11th Floor New York, NY 10004 roberto@speechcycle.com

slide-2
SLIDE 2

What is SpeechCycle?

SpeechCycle is the leading provider of 3rd Generation speech applications for digital service providers.

Started in 2001; located in NYC. Rapidly growing software company (about 70 people) Telephone based automatic spoken dialog systems for complex customer care. Deployment models: on-demand and on-premise managed service. Processing millions of complex support calls every month for the largest cable and telecommunication operators in the US and Australia. Full automation rates up to 40% Experts in speech recognition, speech science, software engineering, advanced voice interaction design, and the strategic value of speech systems in the contact center.

slide-3
SLIDE 3

A brief history of commercial spoken dialog systems

GENERATION FIRST SECOND THIRD Time Period 1994-2001 2000-2005 2004-today Type of Application Informational Transactional Problem Solving Examples Package Tracking, Flight Status Banking, Stock Trading, Train Reservation Customer Care, Technical Support, Help Desk. Architecture Proprietary Static VoiceXML Dynamic VoiceXML Complexity (Number of DMs) 10 100 1000 Interaction Turns A couple 5-10 More than 10 Dialog Modality directed directed + natural language (SLU) + mixed initiative directed + natural language (SLU) + intelligent mixed initiative

slide-4
SLIDE 4

LangTech 2008

4

February 29, 2008

The Problem

  • Cost of customer care is on a steep increasing curve; this is

especially true in the Digital Service Provider arena

– Devices and services are becoming more and more complex—more things can go wrong – Number of customer grows year by year – …and so the number of agents needed to provide quality support – Outsourcing and offshoring reached a point of diminishing returns

  • Quality of customer care is on a decreasing curve

– Long waiting queues – New products and services offered every so often – Agents are not always up to date – Turnover makes agent training difficult and costly – Difficult to maintain consistent quality of service – Infrastructures are not always up to date

slide-5
SLIDE 5

LangTech 2008

5

February 29, 2008

…been there

The switchboards were something to behold, with many, many operators sitting in long rows plugging countless plugs into countless jacks. The cost of adding new subscribers had risen to the point unforeseen in the earlier days, and that cost was continuing to rise, not in a direct, but in a geometric ratio. One large city general manager wrote that he could see the day coming soon when he would go broke merely by adding a few more subscribers.

AT&T, Early 1900s

slide-6
SLIDE 6

LangTech 2008

6

February 29, 2008

Customer Care Automation Value Proposition

  • Value for the customer (the provider)

– Reduced costs

  • Value for the final user (the subscriber)

– No waiting in a queue – Consistent quality of customer care

  • Value for the technology vendor (i.e. us)

– Revenue - Profit

slide-7
SLIDE 7

LangTech 2008

7

February 29, 2008

Technical Support Automation is Difficult

  • Acquisition of knowledge
  • Keeping up to date with new products and services
  • Emotional state of callers
  • Problem identification
  • Caller mental model
  • Instructing non technical savvy callers
  • Uncontrolled events
  • Challenging acoustic environment
  • Cultural barriers against speaking to machines
  • Callers do not trust automated systems can help them
  • Callers are not cooperative
slide-8
SLIDE 8

LangTech 2008

8

February 29, 2008

SpeechCycle’s Approach to Automated Technical Support

  • Speech recognition over the telephone with sophisticated Voice

User Interface

  • Detect call reason using advanced natural language technology
  • Ask simple questions when needed
  • Don’t ask questions...if possible. Integrate with provider’s

customer information systems

  • Instruct caller to follow simple diagnostic and troubleshooting

steps if needed.

  • Perform diagnostic and troubleshooting steps automatically if
  • possible. Integrate with diagnostic systems.
  • Automated reporting and call performance classification
  • Use log data to monitor and continuously improve automation and

caller experience

– Speech – Language – Logic

slide-9
SLIDE 9

LangTech 2008

9

February 29, 2008

The Doctor Analogy

  • Know your patients even before they talk

– Medical records

  • Let your patients talk

– What’s your problem?

  • TAKE THE INITIATIVE
  • Then ask simple clarification questions

– Does it hurt when you laugh?

  • Take measurements, run tests
  • Prescribe cure
  • If it does not work:

– Try something else

  • If it does not work:

– Send to a specialist with an updated medical record.

slide-10
SLIDE 10

LangTech 2008

10

February 29, 2008

The Doctor Analogy applied to Automated Technical Support

  • Know your callers even before they talk

– Integration with Customer Account DB

  • Let callers talk

– Problem identification with natural language

  • TAKE THE INITIATIVE
  • Then ask simple clarification questions

– Directed dialog

  • Take measurements, run tests

– Integration with diagnostic tools

  • Prescribe cure

– Step by step resolution

  • Problem solved?

– How did we do?

  • If it did not work:

– Try something else

  • If it did not work:

– Escalate to human agent

slide-11
SLIDE 11

LangTech 2008

11

February 29, 2008

The long term vision for automated technical support

Please describe your problem My internet connection is slow Sit back and relax. I will fix it for you and call you back when I am done! SPECH ONLY FULLY INTEGRATED Customer account information Home device information Diagnostic tools Network state

IMMERSIVE CALLER EXPERIENCE

slide-12
SLIDE 12

LangTech 2008

12

February 29, 2008

Confidential

The Continuous I mprovement Cycle

Call-Flow Design Call-Flow Testing Prompts and Grammars Management Real-time VUI

  • ptimization and

learning Reporting Call Data Analysis Natural Language grammar development, tuning and testing Speech Recognition Performance analysis Dynamic Content Refresh

DATA

slide-13
SLIDE 13

LangTech 2008

13

February 29, 2008

Handling complexity

CanYouSeeModem yes no WaitGetBrand continue ExpertPrompt ~UserOption.help ~InteractionError MistookRouter true ExpertPrompt nobrand NeedMakeModel ~exit ~exit CallerConfirmedCabledataModem=true TopEnd_CompareModemVars TopEnd_MakeModel success CustomerAccount_Success SupportedModem Supported OhMyModem client cox !!C/C/C!! Unsupported Rpt_OutOfScopeExit=UnsupportedModem HandleTransferOkay ~exit TopEnd_VerifyModem CustomerHasMultipleModems true ModemHasBatteryBackup true ModemHasBattery Backup true TopEnd_NoBrandModem WaitUnplugModem ModemProximity continue ~UserOption.help yes no WaitIdentify ~UserOption.help ~InteractionError continue ~UserOption.help WaitUnplugModem_deux continue StillNeedHelpIdentifying no yes ProximityBad ~exit yes no no ModemProximity_deux CantReachOrFindPowerCord_deux
  • therModemUnplugProblem=true
  • ther
~UserOption ~InteractionError Thanks ProximityBad Escalate ~exit reach find CanYouReachModemPowerCord yes WaitUnplugModem_deux continue no yes Rpt_WantAgent=true ~InteractionError NeedHelpIdentifying_2 HelpUnplugModem NMUnplugModem MistookRouter true WaitUnplugModem_2 ~InteractionError ~UserOption.help continue OtherProblemCapture_2 TransferQueue=normal HandleTransferOkay ThxAndMoreInfo Blurb ThxAndMoreInfoBlurb ThxAndMoreInfoBlurb HelpUnplugModem CycleModemRecently_1 yes no ~InteractionError WaitUnplugModemAnd Coax continue FirstAction=CameBackPrompt ~UserOption.operator ThxAndMore InfoBlurb_1 CallerUnplugged Coax=true ~InteractionError WaitGetPenForVoIPSteps AskIfShouldReadVoIPStepsWithoutPen ~UserOption.help Good continue yes no VoIPStep1of4 ~InteractionError AskIfShouldRepeatAllVoIPSteps ~UserOption.help no ~InteractionError VoIPStep2of4 ~InteractionError.NoInput VoIPStep3of4 ~InteractionError.NoInput VoIPStep4of4 ~InteractionError.NoInput yes ~UserOption.repeat VoIPStepsEpilogue VoIPStepsRepeatPreamble NB: For CIP4.0 compatibility, in VoIPStepNofN activities, moved RepeatPrompts definitions to HelpPrompts; in all questions, changed RepeatThat grammar use to “help”; removed RepeatPrompts (5158.wav) from AskIfShouldReadVoIPStepsWithoutPen, and (5164.wav) from AskIfShouldRepeatAllVoIPSteps. We should undo these changes when CIP4.1 is ubiquitous. HasRouter true ~InteractionError.NoInput heardVoIPSteps heardVoIPSteps true true ~InteractionError.NoInput VoIPStepsPreamble_router VoIPStep2of4_router VoIPStepsPreamble VoIPStep1of4 VoIPStep3of4 ~InteractionError.NoInput VoIPStep4of4 ~InteractionError.NoInput ~InteractionError.NoInput abort
  • perator
Motorola Linksys Toshiba yes no no yes yes no no yes yes no no yes yes no yes no no yes yes yes no ConfirmZoom no yes ConfirmNetgear no yes ConfirmCom21 ConfirmWebstar ConfirmDLink ConfirmAmbit ConfirmBestdata ConfirmCybersurfer ConfirmToshiba ConfirmTerayon ConfirmRCA ConfirmMotorola ConfirmLinksys NeedMakeModel NeedMakeModel NeedMakeModel NeedMakeModel NeedMakeModel NeedMakeModel NeedMakeModel ConfirmCMX no yes netgear linksys motorola rca NeedMakeModel ModemGroupName Best data Bestdata Zoom CMX DLink Scientific Atlanta Com 21 3Com Netgear Ambit Cybersurfer RCA RCA NEW rca Terayon no IsCMX_1 yes no modemtype=cmx LeftThree modemtype=3com BottomTwo ConfirmArris yes no Arris modemtype linksys netgear WaitLinksysHaveCableLight ~UserOption.help DoesLinksysHaveCableLight no continue MistookRouter=true continue ThanksForChecking WaitNetgearHaveStatusLight continue ~UserOption.help DoesNetgearHaveStatusLight yes no Mistook Router! modemtype netgear LinksysToo NetgearToo UnplugThatModem JustAMoment ComputerCount=multiple lightsAgainVisit=net MistookRouter=true ComputerCount=multiple modemtype linksys lightsAgainVisit=net continue yes RouterBrand=linksys RouterBrand=netgear If caller mistook router, we identify modem properly, then power it down. WaitGoFindModem_3 WaitGoFindModem_4 CallerConfirmedCabledataModem true dlink WaitDLinkHaveStatusLight continue ~UserOption.help DoesDLinkHaveStatusLight yes no RouterBrand=dlink WaitGoFindModem_5 continue MistookRouter=true ComputerCount=multiple modemtype dlink lightsAgainVisit=net DLinkToo TopEnd_ModemDown TopEnd_GetModem TopEnd_GetModem TopEnd_GetModem
  • Hundreds of pages of call-

flow (300-400 for typical call flow)

  • Thousands of nodes
  • Several thousands of prompts
  • Hundreds of grammars

Call Flow as Software

Use software engineering principles to create and maintain large call-flows

  • Hierarchical structure
  • Reuse, libraries
  • Object oriented principles
  • Configurability
  • Source control

Call Flow as Software

Use software engineering principles to create and maintain large call-flows

  • Hierarchical structure
  • Reuse, libraries
  • Object oriented principles
  • Configurability
  • Source control
slide-14
SLIDE 14

LangTech 2008

14

February 29, 2008

Symptom Identification with Statistical Spoken Language Modeling (SSLU)

SSLU tools

Internet Symptoms 0.00% 2.00% 4.00% 6.00% 8.00% 10.00% 12.00% 14.00% in te rn e ts lo w s e c u rity

  • p

e ra to r in te rn e ts e tu p e m a il in te rn e tc a n tc

  • n

n e c t v id e

  • m
  • d

e m lo g

  • n

in te rn e t a p p

  • in

tm e n t h

  • m

e n e tw

  • rk

e m a ils e tu p in a c c e s s ib le s ite s s e rv ic e c a b le q u e s tio n b illin g v

  • ip

e m a ilc a n ts e n d

  • rre

c ip a d d re s s in te rn e tin te rm itte n t e m a ils e rv ic e c h a n g e a c c

  • u

n t s

  • ftw

a re m u ltip le s y m to m s p a re n ta lc

  • n

tro l c ra s h fro z e n c

  • m

p u t y e s h e lp e m a ils p a m p

  • p

u p s c a n c e l fire w a ll re p e a t e m a ilw e b m a il

37 Unique Symptoms 225,000 utterances + 3,000 utterances/week 37 Unique Symptoms 225,000 utterances + 3,000 utterances/week

Video Symptoms

0.00% 2.00% 4.00% 6.00% 8.00% 10.00% 12.00% 14.00% 16.00% 18.00% S e rv ic e N

  • S

e rv ic e C h a n n e lM is s in g B

  • x

N

  • tW
  • rk

in g P ic tu re N

  • P

ic tu re In te rn e t P ic tu re S n

  • w

y F u z z y L in e P ro b le m P ic tu re P

  • rQ

u a lity O p e ra to r R e m

  • te

N

  • tW
  • rk

in g A p p

  • in

tm e n t P ic tu re M u ltip le P ic S y m p to m H D T V P ic tu re O th e r C h a n n e lO th e r B

  • x

O th e r D V R O rd e rP a y P e rV ie w O th e r S e rv ic e O th e r B

  • x

M

  • x

i R e m

  • te

V a g u e S

  • u

n d N

  • S
  • u

n d B

  • x

W

  • n

tT u rn O n O rd e rO n D e m a n d O th e r T V N

  • C

a b le T V G u id e N

  • tW
  • rk

in g P ic tu re B la c k R e m

  • te

O th e r C h a n n e lC a n tC h a n g e T V O th e r R e m

  • te

P ro g ra m M y R e m

  • te

A c c

  • u

n tB illin g S e rv ic e M u ltip le O u t S e rv ic e V a g u e A c c

  • u

n tS e rv ic e C h a n g e O rd e rO th e r B

  • x

V a g u e S

  • u

n d P

  • rQ

u a lity S e rv ic e O u ta g e P IN P a s s w

  • rd

P ro b le m O rd e rO n D e m a n d V a g u e A c c

  • u

n tO th e r S

  • u

n d V a g u e S

  • u

n d O th e r P ic tu re V a g u e P ic tu re P ix e la te d P ic tu re F ro z e n P h

  • n

e O rd e rV a g u e A s k e d T

  • C

a ll T V W

  • n

tT u rn O n C h a n n e lV a g u e P ic tu re O th e rC

  • lo

rs A c c

  • u

n tV a g u e O rd e rP a y P e rV ie w V a g u e O rd e rP IN P a s s w

  • rd

O rd e rO n D e m a n d E rro r E rro rO th e r C lo c k B

  • x

M u ltip le B

  • x

S y m p to m s

74 Unique Symptoms 210,000 utterances + 2,000 utterances/week 74 Unique Symptoms 210,000 utterances + 2,000 utterances/week

slide-15
SLIDE 15

LangTech 2008

15

February 29, 2008

Contender: Exploring Alternative Strategies

Cumulative average automation rate

14.00% 15.00% 16.00% 17.00% 18.00% 19.00% 20.00% 21.00% 10/26 10/27 10/28 10/29 10/30 10/31 11/1 11/2 11/3 11/4 11/5 11/6 11/7 WITH EXAMPLES OFFER CHOICES ORIGINAL 5000 10000 15000 20000 25000 10/26 10/27 10/28 10/29 10/30 10/31 11/1 11/2 11/3 11/4 11/5 11/6 11/7 Number of calls

slide-16
SLIDE 16

LangTech 2008

16

February 29, 2008

Conclusions

  • The third generation of dialog systems is here
  • Problem Solving: not just form filling or transactions
  • The doctor analogy
  • The power of backend integration and data
  • Call-flow as software
  • Symptom Identification and SSLU
  • Exploration and Learning