Coveo - Data Wranglers DC Meetup - 18 July 2017

Post on 23-Jan-2018

131 views 3 download

Transcript of Coveo - Data Wranglers DC Meetup - 18 July 2017

TM

Products & TechnologiesOverview for Data WranglersJuly 2017

Gauthier RobeVice President, Products

Daniel CadochPartner Manager

© 2017 Coveo Solutions Inc - Proprietary and Confidential

The Age of the Customer is here… customers are now more

mobile, consume more reviews, and buy more online than

ever before.

Winning In The Age Of The CustomerForrester ResearchApril 6, 2015

2

Forrester: Winning in the Age of the Customer

© 2017 Coveo Solutions Inc - Proprietary and Confidential3

There is a growing gap in customer experiences

of companies say they offer a superior online experience.

say they have a holistic view of their customers (or are near to it).

of customers that leave blame it on bad online experiences.

say their favorite retailer understands them

What brands think What customers think

69%

81%

51%

37%

© 2017 Coveo Solutions Inc - Proprietary and Confidential4

Every customer journey is unique…CustomerJourneys Awareness Interest Purchase Adoption Advocacy Thirst for more

Social Customer Community

Site

Web sites Apps / stores Self-service Portals Collaboration

© 2017 Coveo Solutions Inc - Proprietary and Confidential5

Every customer journey is unique…

… but only relevant interactions drive business

© 2017 Coveo Solutions Inc - Proprietary and Confidential6

Coveo brings relevance to every interaction customers, partners and employees have with your organization

Knowledge ecosystem

Secure connectivity to reach and unify across the entire enterprise

Predictive analytics

Wisdom from the interactions’ sumtotal cues likely intent

Machine learning

Data science drive “what you need next” recommendations

Content RelevanceBehavior+ =+Apps

In the course of work or digital interactions

Context

Content + Context + Behavior = RelevanceWhy Coveo?

© 2017 Coveo Solutions Inc - Proprietary and Confidential7

The Impact of Relevance

Customers don’t convertand go elsewhere

Agents can handlemore complex cases

Engineers can work onmore complex projects

Salespeople cansell more

Customers can self-serve with ease

Customers can do more,learn more and buy more

Employees end up doing repetitivetasks, resulting in a high churn rate

Incapable of positively impactingupsells and cross-sells

Customer satisfactiongoes down

Poor self-service results in contact center costs going up

POORRELEVANCE

RICHRELEVANCE

© 2017 Coveo Solutions Inc - Proprietary and Confidential8

Relevance is central to personalization

“…for me”

Relevance or relevancy[rel-uh-vuh ns] - noun: the condition of being relevant, pertinent or connected with the matter at hand…[case, account, employee, customer, shipment, part, product, patient…]

▶ Profile, locale

▶ Context, task

▶ Intent, desire

▶ Interactions with your company [clickstream, outcomes, …]

© 2017 Coveo Solutions Inc - Proprietary and Confidential

content intelligence

Positive business impact &

ROI

Negative economic

Impact

.

Siloedsearch

Federatedsearches

0efficiency proficiency

LEADERSLAGGARDS

6Discovers user’s likely intent, by analyzing behavioral data.

Machine learning auto-tunes recommendations and ranking to maximize business outcome, enabling true one-to-one user engagement and upskilling.

Self learning predictiverecommendations

PREDICTIVE[relevance is predictable]

5…related content, experts, products or services pushed to expand user’s knowledge and abilities to do more, buy more, learn more, engage more, etc.

Contextualsuggestions

3

4

…adapt relevance through weighting of ranking factors, query ranking expressions, based on known content characteristics.

…in-product, in-task, and other contextual signals factor the uniqueness of the user and query context and ranks results with higher relevance.

Tunable relevance

Contextual relevance

PROACTIVE[relevance is contextual]

1

2

Unifies and ranks information from multiple sources.

Configurable rich facets, search tabs, folding and security trimming provide the very first step of personalization for users.

Content navigationSecured

unifiedranking

RESPONSIVE[relevance is personal]

CRMM™ stages

The Coveo Relevance Maturity Model™ is a roadmap

REACTIVE[Search with low relevance]

“Relevance is the currency of the digital age.”

© 2017 Coveo Solutions Inc - Proprietary and Confidential

Internal facing: Intelligent Workplaces

“…for me” “…for me”

Customer communities

Personalized websites

Customer self-service

Partner & Dealer portals

eCommerceIn-product intelligence

External facing: Intelligent Engagement

Contact center

CRM & Sales

Employeeportals

Intranets& KM

AI-powered search & recommendation platform [PaaS]

Coveo has several solutions to help you get started

© 2017 Coveo Solutions Inc - Proprietary and Confidential

Insight engines apply relevancy methods to describe, discover, organize and analyze data.

This allows existing or synthesized information to be delivered proactively or interactively, and in the context of digital workers, customers or constituents at timely business moments.

Analysts are recognizing Coveo’s impact

Source: Gartner MQ for Insight Engines 2017

© 2017 Coveo Solutions Inc - Proprietary and Confidential

“Search is about finding answers, content, and documents; adding useful context to apps; and augmenting human intelligence. The most important requirements are the relevancy and completeness of the returned results.”

Relevancy and completeness matter most

Source: The Forrester Wave™: Cognitive Search andKnowledge Discovery Solutions, Q2 2017

The Forrester Wave™ is copyrighted by Forrester Research, Inc. Forrester and Forrester Wave are trademarks of Forrester Research, Inc. The Forrester Wave is a graphical representation of Forrester's call on a market and is plotted using a detailed spreadsheet with exposed scores, weightings, and comments. Forrester does not endorse any vendor, product, or service depicted in the Forrester Wave. Information is based on best available resources. Opinions reflect judgment at the time and are subject to change.

© 2017 Coveo Solutions Inc - Proprietary and Confidential13

Coveo is used by customers across industry sectors

Software & Technology

Financial Services

Manufacture

Healthcare

Others

Partners

© 2017 Coveo Solutions Inc - Proprietary and Confidential

Daniel CadochPartner Manager at Coveo

dcadoch@coveo.com

dancadoch

@dancadoch

14

Thank you. Any questions?

15

Product Update

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

AI-Powered Search from Coveo

Secure, Unified Indexing

Secure connectivity to reach and unify across the entire enterprise

Content

In-App Search Journey

Integration in the course of work or digital interactions

Context+

Usage Analytics

Wisdom from the interactions’ cues likely intent

Intent+

Machine Learning

Data science drives “what you need next”

Relevance=

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

Coveo Cloud V1

➔ Designed for Coveo for Salesforce

➔ ML & Usage Analytics

➔ Hundreds of live customers

Coveo Cloud V2

➔ Supports all use cases

➔ Designed for developers

➔ Advanced security

A Brief Summary

2013 2014 2015 2016 2017

Coveo Enterprise Search

➔ On-Prem.

➔ Microsoft-centric

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

18

Today’s Developer Requirements

Easy Scaling Flexible Ingestion Model

Extensible Indexing Pipeline

Sophisticated Query Language

Ready to Use Search UI

Analytics and ML Features

Coveo Platform

:35

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

New architecture

Scale & Resiliency at core

Extensive use of micro-services

Push API Infrastructure

Indexing Pipeline Extensions

Crawling Modules

HIPAA Environment

New APIs

20

Coveo Cloud V2

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

21

Coveo Cloud V2: Security

Customer IsolationPrivate customer data isolated and single-tenant

Strong Data CentersHosted on Amazon Web Services

Leveraging High Availability Options

ComplianceReviewed by an independent security firm

SOC 2 Compliance hosted on AWS (25+certifications)

EncryptionEncryption at rest and in-transit

Dedicated key

24/7 MonitoringContinuous monitoring for security, availability and performance

HIPAAHIPAA compliant environment

BAA ready

OAuth & SAMLCustom providers for admin console authentication

Disaster RecoveryComplete DR plan in place

Penetration TestingYearly penetration and vulnerability assessments by 3rd party

Content and PermissionsDoc-Level Permissions within Coveo Index

Option to select objects and fields to index

Access ManagementAudited and reviewed fine-grained accesses

Personnel Security | Change ControlSecurity check performed by an independent firm

Strict segregation of duties and constant monitoring of changes

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

22

Indexing Pipeline Extensions (IPE) – Leverage remote services

Indexing Pipeline

Google Translation APITranslation…

AWS RekognitionDeep Learning, Image recognition

meaningcloudText Analytics

Extensions – Examples (tested)

Index

Connectors

Salesforce EinsteinImage recognition, NLP

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

23

Coveo Query Language

https://developers.coveo.com/display/public/SearchREST/Anatomy+of+a+Complex+Query+Using+Extensions

{{accounts=$type(name:'Account') @systitle="{!>Name}"}}{{contacts=$join(fromResultSet: {{accounts}}, toResultSet: $type(name: 'Contact'), field: '@sfaccountid')}}{{emailAddresses=$valuesOfField(field: '@sfemail', resultSet: {{contacts}})}}$emailsExchangedWithThoseAddresses(addresses: {{emailAddresses}})NOT @sysisattachment$sort(criteria: 'datedescending')

(NOT @syssfid=={!>id})(NOT @syssfcaseid=={!>id})

/*--- Extract related cases ---*/{{relatedCases = @objecttype=="Case" $some(keywords: '{!>Subject}', match:1) @sfparentcasenumber$correlateResultSet(resultSet: @syssfid=={!Id}, field: '@sysconcepts', maximumValues: '25', modifier: '1000')}}

/*--- Extract parent cases of related cases ---*/{{parentCases = $valuesToResultSet(values: $valuesOfField(field: '@sfparentcasenumber', resultSet: {{relatedCases}}), field: '@sfcasenumber' )}}

/*--- Find some relevant knowledge ---*/$some(keywords:$removeStopWords(keywords:$anyKeywords(keywords:$valuesOfField(field: '@systitle', resultSet: {{parentCases}}))))

https://developers.coveo.com/display/public/SearchREST/Query+Extension+Samples

General Extensions$q$qf$qre$qrf$weight$sort$fold$loadParent

Extensions for Lists of Values$valuesOfField$keepMatchingValues$removeMatchingValues$removeEmptyValues$removeDuplicateValues$replaceInValues$mergeValues$joinValues$splitValues

Join Extensions$valuesToResultSet$joinOnValues$join$filterJoin

Correlation Extensions$correlateListOfValues$correlateResultSet$correlateUsingIdf$some$removeStopWords$noStemming

Query ExtensionsExample 1: filter query on all email messages with contacts from selected accounts

Example 2: Extract unique keyword list from all related cases subject

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

24

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

25

Query Pipeline: Featured Results

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

26

Relevance Management - Query Pipeline

Ranking Expressions

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

27

Relevance Management - Query Pipeline

Ranking Weights

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

28

Relevance Management - Query Pipeline

Triggers

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

29

A/B Testing: Compare two Pipelines

A/B Testing

Machine Learning & Relevance

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

32

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

33

Optimizesfor future users

Analyzesuser behavior

Learnsbased on success

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

1. Intelligent Query Suggest

2. Automatic Relevance Tuning

3. Recommendations

4. Intelligent Term Detection

5. Personalization

Key Use Cases

Optimizesfor future users

Analyzesuser behavior

Learnsbased on success

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

Personalized per cluster of users

User clusters➔ Users are categorized by their

search behavior

➔ Generated automatically

Salesforce HT Portal➔ 200 user clusters generated and

balanced automatically

35

Personalized Query Suggest

User = Anonymous

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

Personalized per cluster of users

User clusters➔ Users are categorized by their

search behavior

➔ Generated automatically

Salesforce HT Portal➔ 200 user clusters generated and

balanced automatically

36

Personalized Query Suggest

User = Richard Tessier

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

37

RecommendationsAppExchange – The future

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

Standard Relevance

Most popular items automatically promoted

MXtm 5500 match

cuay59 is part number of the usb receiver

With Machine Learning

2. Automated Relevance TuningLogitech Self-Service

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

39

How to setup a server as master –slave? Like 1 master – 2 slave 2 slave shaold have same index content so that we can make load balace url for 2 slave. Please tell how to install.

How to setup a server as master –slave? Like 1 master – 2 slave 2 slave shaold have same indexcontent so that we can make load balace url for 2 slave. Please tell how to install.

4. Intelligent Term DetectionCoveo Self-Service Community

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

41

Machine Learning configuration

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

For all featuresAll features use “Hyperparameter optimization” consisting of a mix of local search, simulated annealing and genetic algorithms.

“Hyperparameter optimization” is used to optimize/personalize the model parameters for each client use case

Automated Relevance Tuning (ART)Bayesian Classification of queries to documents, with time decay

Entire visit evaluated when learning the model. Models queried before the index.

RecommendationsCooccurrence-based algorithm using a log likelihood ratio.

Query SuggestQuery Performance score combined with a customized distance metric based on Smith-Waterman and Monge-Elkan approaches

https://www.google.com/patents/WO2016187705A1?cl=en

42

Some additional background…

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

43

Future: Streaming Data Platform as a provider to other apps

RelevanceTuning

Usage Analytics

Machine Learning

RankingSuggestions

UI &Query Engine

Ranking fromModels

BehavioralData

Other data sources

StreamingData Platform

FilteredEvents

OtherApps / EnginesFiltered

Events

Events EventsRanking Rules

44

Index Agnostic Platform

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

Coveo Platform – Enterprise (Current)

Indexing

Query & Relevance

45

Connectivity

UsageAnalytics

MachineLearning

JS UI QueryEngine

EnterpriseConnectors

CloudConnectors

Coveo Index

Entire stack managed by Coveo

End-to-end Compliance & Security

Advanced Features

HIPAA option

Proven reliability and maturity

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

Indexing

Query & Relevance

46

Connectivity

UsageAnalytics

MachineLearning

JS UI QueryEngine

CloudConnectors

Elasticsearch

Coveo Platform – Index-Agnostic (Future) Tap into a large dev. community

Leverages existing Elasticsearch install-base

Ease data-governance challenge

Opens to Azure and other clouds

First step into a series of “index-agnostic” options

Pilot: 2017-Q4

EnterpriseConnectors

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

47

Coveo – Proprietary Index

Coveo Index

Search Region >>

Coveo Intelligence EngineSelf-Learning Relevance, RecommendationsPowered by Machine Learning

Search OptimizationRelevance, Thesaurus…

Coveo UIUI ConfigurationLightning Components

Deflection AnalyticsPowered by Coveo Usage Analytics

Coveo ConnectorsUnified Index, Federated Search

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

48

Coveo Intelligence EngineSelf-Learning Relevance, RecommendationsPowered by Machine Learning

Search OptimizationRelevance, Thesaurus…

Coveo UIUI ConfigurationLightning Components

Search Region >>

Deflection AnalyticsPowered by Coveo Usage Analytics

Customer owned index. Hosted anywhere

Coveo – Index Agnostic

Coveo ConnectorsUnified Index, Federated Search

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

49

Coveo on Elasticsearch Differentiators

Use your own Elasticsearch cluster. No Coveo VM/Blackbox

Keep leveraging your investment, developments, knowledge

Connectivity: cloud connectors, push-API

Item-level permissions

Indexing Pipeline Extensions

OOTB Relevance based on ”Coveo’s Secret Sauce”

Usage Analytics focused on “user experiences”

Machine Learning services (relevance, recommendation, etc.)

Flexible UI Framework compatible with all field-level features and types

“From signup to search in 5 minutes”

50

Resources

Cop

yrig

ht ©

201

7 C

oveo

Sol

utio

ns In

c.

All

right

s res

erve

d. P

ropr

ieta

ry a

nd C

ON

FID

ENTI

AL

JavaScript UI Framework

https://github.com/coveo/search-ui

Coveo Swagger (API)

https://platform.cloud.coveo.com/docs

Documentation

https://search.coveo.com

Coveo-Labs

https://github.com/coveo-labs

Q&A

http://answers.coveo.com/51

Developer focused resources

Thank you!