Decision-MakersÕ Guide to Enterprise Intelligent

| ReportDecision-Makers’ Guide to Enterprise Intelligent Assistants

Report

»

»August 2015Dan Miller, Lead Analyst & Founder, Opus ResearchDerek Top, Research Director, Opus Research

Opus Research, Inc.350 Brannan St., Suite 340San Francisco, CA 94107www.opusresearch.net

Published August 2015 © Opus Research, Inc. All rights reserved.

2

Report »

»

Enterprise Intelligent Assistants (EIAs) offer human-like, automat-ed services as a natural way for individuals to carry out commerce through companies’ voice response systems, Web sites and mobile apps. In this report, Opus Research defines “Eight Characteristics of Highly Successful EIAs" and identifies 13 firms whose platforms best embody those traits.

Decision-Makers’ Guide to Enterprise Intelligent Assistants


Table of Contents»

Defining “Intelligent Assistance” . . . . . . . . . . . . . . . . . . . . 4Success With Mobile Personal Assistance Drives Growth . . . . . . . . . . . . 4An Amalgam of Many Technologies . . . . . . . . . . . . . . . . . . . 6Eight Characteristics of Highly Intelligent Assistants . . . . . . . . . . . . . 7Accommodating “Mobile First” Support Strategies . . . . . . . . . . . . . . 8Consistency Breeds Loyalty and Vice Versa . . . . . . . . . . . . . . . . 8Responses Based on Personally Identifiable Information . . . . . . . . . . . . 8Reporting Tools Are Crucial for Proving Value . . . . . . . . . . . . . . . 9Integrating with Existing Platforms . . . . . . . . . . . . . . . . . . . 9Making the Most of Machine Learning . . . . . . . . . . . . . . . . . . 9Real World, Real Time Experience is the Best Teacher . . . . . . . . . . . . 10Solutions Must Scale and Add Domains . . . . . . . . . . . . . . . . . 10Path to Maturity: IAs Get Predictive to Prove Business Value . . . . . . . . . . . 12Next Steps: Scale, Live Agent Integration and AI . . . . . . . . . . . . . . 12Vendor Profiles . . . . . . . . . . . . . . . . . . . . . . . . . 13

[24]7 . . . . . . . . . . . . . . . . . . . . . . . . . . . 13aivo (agentbot) . . . . . . . . . . . . . . . . . . . . . . . . 15Artificial Solutions . . . . . . . . . . . . . . . . . . . . . . . 17Creative Virtual . . . . . . . . . . . . . . . . . . . . . . . . 20IBM Watson Group . . . . . . . . . . . . . . . . . . . . . . 23Inbenta . . . . . . . . . . . . . . . . . . . . . . . . . . 25Interactions LLC . . . . . . . . . . . . . . . . . . . . . . . 27Kasisto . . . . . . . . . . . . . . . . . . . . . . . . . . 29Next IT . . . . . . . . . . . . . . . . . . . . . . . . . . 31noHold . . . . . . . . . . . . . . . . . . . . . . . . . . 33Nuance . . . . . . . . . . . . . . . . . . . . . . . . . . 36SmartAction . . . . . . . . . . . . . . . . . . . . . . . . . 39Verbio. . . . . . . . . . . . . . . . . . . . . . . . . . . 41

Table of FiguresFigure 1: The Enterprise Intelligent Assistant Roster . . . . . . . . . . . . . 5Figure 2: Enterprise IA Revenue Forecast . . . . . . . . . . . . . . . . . 5Figure 3: Vendor Assessments . . . . . . . . . . . . . . . . . . . . 11Figure 4: IA-to-AI Continuum . . . . . . . . . . . . . . . . . . . . . 12

»

3

4


Defining “Intelligent Assistance” Opus Research coined the term “Intelligent Assistance” to describe how smartphones and other intelligent end-points combine with wireless networks, The Cloud and Big Data to recognize people and help them accomplish specific goals. Upon further review, it is important to parse the overall concept into its two basic terms.

• Intelligent – Most directly maps to the idea of “Artificial Intelligence” which generally refers to the ability ofcomputer systems to carry out tasks as if they were human. Pattern detection, speech recognition, machinetranslation and recommendation engines are its pale precursors. For the purposes of this document, intelligencerefers to the ability of computers to observe and make sense of our behavior and understand our intent (what wemean when we say something or key words into a text-box).

The term “intelligence”, as in “Central Intelligence Agency” or “Machine Intelligence”, is also relevant here. It conjures up images of observation of individuals’ activities and words and aggregation of data from a multiplicity of sources. This brings the enterprise databases that support customer relationship management (CRM), workforce optimization (WFO), and responses to frequently asked questions (FAQs) into the mix. Third-party data aggregators, such as credit reporting bureaus, are also major sources of information that supports automated decision-making.

• Assistance – synonym for “help” or “aid.” The term “automated assistance” describes how machineintelligence is applied to assist or augment human endeavors. For example, recognizing patterns in past behavioror input can enable an enterprise to recognize the purpose of an interaction more quickly or even predict what acustomer or prospect will do next. This maps directly to creating a better customer experience for individualstrying to accomplish specific tasks.

Just as important, it can apply to making live customer care agents more efficient by providing pertinent, relevant and current information about the purpose of an individual’s call and proposing or prompting the agent determine the next best action.

Success With Mobile Personal Assistance Drives GrowthThe I A phenomenon is expanding geometrically. Usage is at an all-time high as individuals turn to personal assistants on their smartphones or virtual assistants on websites to answer queries and get things done. Expectations are even higher as machine learning and the experience curve kicks in and IAs grow ever more successful at responding correctly. When Apple introduced the iPhone 4s in 2012, its marketing prowess and billion dollar advertising budget convinced millions of people to talk to and expect results from Siri. In the ensuring years, the ranks of mobile personal assistants have grown to include Nuance’s Nina, Google Now (“okay Google!”), Microsoft Cortana, Artificial Solution’s Indigo, Get Abby and, most recently Hound (from SoundHound).

Increased comfort and successful experience with mobile personal assistants has had an impact on expectations and use of Enterprise Intelligent Assistants. In this document, Opus Research has identified firms, listed in Figure 1 (below), with technology platforms that foster Enterprise IAs. They support automated, natural language handling of customer queries through webchat or speech-enabled IVR (interactive voice response) systems.

»

5

Figure 1: The Enterprise Intelligent Assistant Roster

While many of these companies are privately held, Opus Research estimates that they generated approximately $230 million in revenues in software licensing and services revenue in 2014. Enterprise spending on their solutions is growing at roughly 28% annually and should exceed $500 by 2017 on its way to being a billion dollar opportunity in 2020.

Figure 2: Enterprise Spending on IA

SOURCE: OPUS RESEARCH (2015)

Decision-Makers’ Guide to Enterprise Intelligent Assistants »

Company Platform[24]7-IntelliResponse IntelliResponse Virtual Agent PlatformAivo AgentBotArtificial Solutions TeneoCreativeVirtual V-PersonIBM WatsonInbenta Inbenta PlatformInteractions Adaptive Understanding Technology™ Kasisto KasistoNextIT AlmeNoHold NoHold PlatformNuance NINASmartAction IVA

Verbio CX Platform


$1,200

$1,000

$800

$600

$400

$200

$02014 2015 2016 2017 2018 2019 2020

6

An Amalgam of Many TechnologiesIntelligent Assistant technologies leverage investment that enterprises make in communications, computing and cloud-based resources that support multi-channel customer care, as well as Web-based and mobile e-commerce or digital commerce. This investment is just a modest share of consensus estimates of annual enterprise spending on platform as a service (PaaS), big data and analytics or even contact center and customer care infrastructure.

In December 2014 IDC, for instance, noted that global spending on a “third platform” f information and communications technologies would reach $3.8 trillion. In June 2015, Gartner observed worldwide IT spending at $3.5 billion, noting that a 5% decline was precipitated by the rising strength of the U.S. dollar, among other considerations. By contrast, Opus Research has observed double-digit growth in spending on relevant customer-facing IT technologies, including:

3 Contact Center and Interactive Voice Response (estimated $6.5 billion in spending 2015)

o Voice processingo Call Processingo Recording and monitoringo Cloud-based contact centers

3 E-commerce Web Infrastructure (estimated $25 billion)o Chato FAQso E-commerce and transaction processing

3 Knowledge Management & Analytics (over $100 billion)o Big Data & Analyticso Sentiment Analysiso Business Intelligence

As keepers of the longest-standing customer-facing resources in the enterprise, many Contact Center and IVR specialists have deep experience with natural language processing and intelligent call handling. More recently they have been working with their counterparts who have built the foundation for both live agent and automated Web chat. The ideal IA platform is designed to leverage investment and experience with speech-enabled IVRs, automated FAQs, Web chat and real-time analytics. This makes it an important component of “Marketing Technology,” which has been in the spotlight since IDC famously forecasted that chief marketing officers (CMOs) would drive more IT spending than chief information officers (CIOs) within two years, taking responsibility for over $32 billion in annual spending.


THE FIRMS UNDER STUDY IN THIS REPORT BRING A

CUSTOMER-FACING LAYER TO THE SOLUTION STACK THAT ENABLES

ENTERPRISES TO CARRY ON COMMERCIAL CONVERSATIONS WITH

LOYAL CUSTOMERS WHO WANT TO GET THEIR ISSUES RESOLVED

EFFORTLESSLY OR TRANSACTIONS COMPLETED QUICKLY.

7

Eight Characteristics of Highly Intelligent AssistantsEnterprise Intelligent Assistants bring user control and a natural user interface to both automated and human-assisted self-service. They are instantiated in two principal ways:

• Q&A – Supporting automated Webchat, conversational voice response and agent assistance to resolvecustomer queries and frequently asked questions based on the “best answers” previous interactions.

• Conversational E-Commerce – Offering human-like responses to natural language queries, requests orinstructions. These are more likely to involve avatars, persona or branded virtual agents that are associated withthe company’s brand or marketing initiatives.

Opus Research has developed a short list of the characteristics that can be associated with the most successful deployments of Intelligent Assistance technologies in the enterprise. The nature of the list may surprise you with the most important attributes as follows:

3 Support of a “mobile first” customer experience – Conversations with customers most often start on a mobile device or through a browser of the company’s mobile app.

3 Consistent responses and experience across multiple devices and channels – Maintaining the context of interactions across multiple channels and devices, at varying times and methods, improves customer experience.

3 Responses based on personal knowledge – Verifying and authenticating an individual allows the Intelligent Assistant to recognize the purpose of contact and predict the next best action for the company to take.

3 Support of specific “key performance indicators” – Offering reporting platforms and “dashboards” that support enterprise strategies, tactics or goals for customer satisfaction, automation rates (e.g. call deflection), upselling and, ultimately, return on investment.

3 Leverage of existing CX investment – ugmenting the technology and staff that supports both phone and eb-based natural conversations through both automated resources and live agents, including mobile apps, Web chat, IVR and contact centers.

3 Ability to learn – Meaning that it can refine and improve responses based on learning user behavior or on direct user input.

3 Successful track record - Benefitting from the experience, data aggregation and “learnings” from numerous deployments across multiple verticals.

3 Future Readiness – Providing for the ability to handle a growing number of vertical markets and knowledge domains at scale. Also referring to the ability to move from search and needs assessment to transaction completion.

Below is a deeper dive into each of these characteristics, culminating in a tabular display of how the 13 firms under investigation in this document fill these criteria.


8

Accommodating “Mobile First” Support StrategiesThe proliferation of smartphones is driving high expectations for Intelligent Assistance as customers take greater control of the digital resources that support their everyday activities. Around the world there are 5 billion mobile phones in service and 2 billion are smartphones, according to eMarketer. In the U.S., nearly two-thirds of the population have smartphones and 87% of Millennials have devices that “never leave their side,” according to Zigby Analytics.

Smartphone owners use their devices as tools for socializing, planning, communicating, shopping and navigating both the digital and physical world. In a survey conducted for Bank of America by Braun Research, 24% of respondents characterized their phone as a “Personal Assistant” while 11% termed it a “Daily Planner,” followed by smaller percentages using terms like “life line”, “wing man/woman,” “coach,” or “concierge.”

Speech-enabled, natural language resources – like Siri for iOS, GoogleNow, Cortana, Nina Mobile, Hound and others – fulfill the need for a natural interface in a broad variety of functions. They enable smartphone users to cut through complexities by talking in their own terms. Enterprise-based virtual assistants complement the capabilities of smartphone-based personal assistants, but also take into account several other sources of real-time data and metadata in order to answer questions quickly, complete transactions or make an intelligent transfer to a live agent.

Consistency Breeds Loyalty and Vice VersaWhether it is an avatar-based “virtual assistant” on a ebsite or simply an automated chat “bot” supporting Q&A, the magic of an enterprise IA is its ability to provide a consistent, accurate response to a customer’s query, regardless of channel, device or communications mode. As customers grow accustomed to human-like automated assistance, they will be grateful to get the same, accurate answer regardless of how and when they try to reach a company.

But consistency is just the beginning. A number of the firms included in this study have software platforms that support communications across a number of channels. Others maintain the context of the call throughout the customer’s “journey,” which may be self-initiated or be the result of an outbound “alert” informing a customer that he or she must take immediate action. Of course, in order to receive outbound alerts and to garner the advantages of other company-provided perks, individuals are counseled to enroll in a company’s loyalty program or at the very least, download and register on a company’s mobile app.

Responses Based on Personally Identifiable InformationOnce enrolled and authenticated, customers will find that the services they receive are personalized, speedy and informed by past activity and indications of preferences. Smartphone owners who have experience with Google Now or the new Apple Proactive Assistance will already expect as much. The providers of Intelligent Assistance latforms should provide enterprise customers with sufficient predictive analytics capabilities to recognize customers at the point of contact in order to anticipate the purpose of their call or contact. Authenticating users early can help enable the speediest way to resolve their queries or complete their desired transactions.

The notion of using “personally identifiable information” (PII) to support a commercial transaction sets off alarm bells among privacy advocates as well as casual shoppers. On the other hand, multiple customer satisfaction surveys indicate that people do not like to repeat themselves when they are transferred from one customer service rep to another. They also resent having to answer challenge questions in order to assert their identity when talking to an agent of any kind. Thus, there is a recognized benefit to using PII, specifically the content of past conversations


9

and assertion of identity when moving from one modality to another. On the other hand, businesses have to take measures to protect any personal information that they have collected and stored from unwanted access.

Reporting Tools Are Crucial for Meeting Enterprise GoalsEnterprise executives justify procurement and implementation of IA platforms based on their ability to fulfill specific operational objectives. In addition to basic “service level agreements” surrounding uptime and time-to-respond to a trouble ticket, they have quantifiable key performance indicators (KPIs) surrounding customer satisfaction and call deflection, plus ROI objectives based on cost savings and revenue enhancements.

When selecting an IA platform provider, it is important to look at their integrated reporting capabilities. As one vendor described it, a system should capture “KPIs from every transaction and serve up closing survey and track it so that you can tell the company about success in their own metrics.” If a company is installing an IA to provide automated handling of 80% of customer queries about a specific product, the system should report that. If the objective is to increase the number of calls involving live agents that culminate in a “close,” that metric should be reported as well.

In addition, the system should take so-called “Voice of the Customer” (VoC) surveys into account. Like any workforce optimization (WFO) resource, the system should tally the number of satisfactory responses and correlate it with the actions of the IA.

Integrating with Existing PlatformsIAs are often deployed as intelligent front-ends or integration points for existing customer care and support resources. Speech-enabled IVRs, especially those with an open-ended, “How May I Help You” greeting, are already carrying out some of the IA promise. The technical staff that developed and maintain such resources have already made strides into the IA world. As smartphone-based mobile apps move from the first generation of stand-alone functionality to offer more integrated, “in-app” capabilities, IAs have a role to play as the conversational resource that understands what individuals are asking for and responding correctly or routing them to the right resource.

The IA platform providers evaluated in this document follow engagement models that start with an assessment of existing resources. The best-in-class offer multiple integration points between those resources and their core platforms. This approach lends itself to seamless navigation with the IA as an initial point of contact who can then transfer or escalate the conversation to other automated resources or live agents.

Making the Most of Machine LearningPerformance analytics should not stop with executive reports. Assessing the success or failure of the IA is at the root of “machine learning” that leads to constant refinement and improvement of the IA-supported self-service experience. In brief, IAs need to learn from their successes and failures, and be dynamic enough to refine their responses over time to improve their ability to respond accurately to a customer query and resolve their issues.

The firms included in this report have, to varying degrees, made machine learning a fundamental part of their solution sets. Engagement models often start with discovery of the best answers to frequently asked questions. Source material can be existing FAQs on a ebsite or transcripts of successful conversations with the “best” live agent. The corpus of correct answers, in the ideal case, is a dynamic database that evolves as the platform discovers and aggregates the multiplicity of ways that questions are answered and matches them with answers that correlate with the best outcome.


10


Real World, Real Time Experience is the Best TeacherWhen building corpora of responses to customer queries, nothing beats real world experience for providing context. Think of Web-based search. It is clear to users and competitors of Google that there is a real advantage to culling through and categorizing the entire World Wide Web in order to build the best algorithms for speedy, accurate search. When enterprises seek to provide the same levels of speed and convenience to their customers, they know that they benefit from the sort of “deep understanding” that results from handling high volumes of domain-specific information.

Many of the solution providers described in this document have a number of successful deployments in specific vertical markets, including financial services, telecommunications, travel and hospitality, healthcare and insurance, technology and retail. While they protect the confidences of their clients, much like Google, they find that all of their implementations are informed by their experiences with prior clients, customers and verticals.

Solutions Must Scale and Add DomainsFinally, we have to take each platform provider’s ability to accommodate both high levels of traffic and the addition of new knowledge bases and domains of expertise. Another key selection criteria is the ability of an IA to be predictive, meaning that it anticipates the purpose of a contact before it takes place and manages queries and requests quickly.

In Figure 3 (Next Page), we present a quick evaluation of offerings from the firms described in this document according to five of the criteria described above.

MANY OF THE SOLUTION PROVIDERS DESCRIBED IN THIS DOCUMENT HAVE A NUMBER OF SUCCESSFUL DEPLOYMENTS IN SPECIFIC VERTICAL MARKETS, INCLUDING FINANCIAL SERVICES, TELECOMMUNICATIONS, TRAVEL AND HOSPITALITY, HEALTHCARE AND INSURANCE, TECHNOLOGY AND RETAIL.

11

Figure 3: Vendor Assessments


COMPANY MOBILE FIRSTTRACK

CONTEXT PERSONALSUPPORTS

KPIS REPORTS

LEVERAGE EXISTING CX

INV’TLEARNS TRACK RECORD FUTURE

READY

[24]7-IntelliResponse

Aivo

Artificial Solutions

Creative Virtual

IBM Watson

Inbenta

Interactions

Kasisto

Next IT

NoHold

Nuance

SmartAction

Verbio

Point of emphasis, up-and-running – firm has made this attributecentral to its platform offering.

Mature offering, gaining traction – “in the field” with one or more enterprise customers.

Fully productized, not widely implemented.

In planning, but not developed.

Not on product roadmap


Figure 3: Key to Vendor Assessments 5

12

Path to Maturity: IAs Get Predictive to Prove Business ValueThe prime directive of an intelligent assistant is to be a consistently supportive “natural interface” for customers or prospects across multiple channels and modalities to communicate and carry out business with companies of their choosing. In addition to the criteria illustrated in Figure 3, enterprise executives should be mindful of a third-dimension of attributes that IAs must have in order to lower customer effort and deliver desired business results.

Next Steps: Scale, Live Agent Integration and AILooking into the future, IA platforms must incorporate “artificial intelligence” (AI). At a minimum, they should interact like a company’s best agent or representative. They must “know” the customer and quickly recognize that customer’s intent so that they can help complete a task or transaction. Then they must retain information (with the customer’s permission of course) in order to perform tasks successfully the next time that customer contacts the company through the IA.

The first step is to perform the pattern matching and understanding functions “at scale.” For large enterprises, this means more than just handling millions of queries, transactions or conversations in real-time or near real-time. It means providing companies with the tools or services required to add areas of expertise and deep knowledge quickly. That is key to supporting specific business objectives. They must also do more than merely mimic the “best agents” and add business rules and application logic to perform “warm transfers” or escalations to live agents in seamless and pleasing ways.

Figure 4: IA-to-AI Continuum

Maturity

From the point of view of searchers, browsers or shoppers, Enterprise IAs are the go-to resources that they encounter every time they contact the companies with which they want to carry out business. They operate on the individual’s behalf while, at the same time, collecting data, rendering reports and, basically, learning how to do a better job the next time.

Predicting the next action of individual shoppers, recognizing their intent and accelerating the time it takes to complete a task or transaction benefits both the customer and the company. Within the next three years, as the concept of an efficient IA matures and the ecosystem of service and solution providers expands, Enterprise IAs will be at the primary point of contact supporting real world commerce in the digital realm.


Static Info:Q&ABased on limited info sourcesOne-way

Dynamic/Complex"Big Data"UnstructuredReal-time ( .g. location)User provided

TransactionalMore complex ( .g. product selection)Knows payment preferencesUser contro led PIIConstantly improving

13

[24]7[24]7US Headquarters:910 E. Hamilton Ave., (Suite 240)Campbell, CA 95008Tel: +1-650-385-2247

Founded: 2003Employees: 5,000+Revenues: $255+ million (estimate from TechCrunch)Market Cap: private company

Company Highlights:Internal development of knowledge management and predictive analytics software to support live agent-based chat has been augmented through the acquisition of technology and personnel from Voxify, Tellme and IntelliResponse to create enterprise software to help large enterprises anticipate what their customers or prospects want, simplify interactions, and learn from those interactions so that future experiences constantly improves.

Key Executives:PV Kannan, Co-founder and CEOShammugam Nagarajan, Co-founder and Chief People OfficerKathy Juve, Chief Marketing OfficerPatrick Nguyen, Chief Technology OfficerDavid Lloyd, President-IntelliResponse Division

Enterprise Intelligent Assistance Products[24]7 acquired IA specialist IntelliResponse in November 2014. IntelliResponse (founded in 2000) offered virtualagent technology for a roster of clients that represent 450 deployments and over 150 brands. The combinationof IntelliResponse’s Virtual Agent Platform with [24]7’s Predictive Experience Platform is designed to support apredictive, omnichannel experience that reduces customer effort needed to get questions answered or to completetransactions. The Virtual Agent-based Webchat also has the ability to perform “smart escalation,” when it determinesthat a live agent is required to respond correctly or make a sale.


Vendor Profiles

(Continues on next page)

14

Revenue models and engagement strategy[24]7 has long adhered to an outcome-driven or success-based revenue model. Clients pay an agreed upon rate foraccomplishing stated objectives.

Customers/Verticals:Featured implementations span financial services (CIBC, CapitalOne, RBC Royal Bank of Canada, Bank of Internet USA, Atlantic Credit Union, ECCU), airlines (CopaAirlines, VirginPulse, AA Cargo), utilities and telcos (EPCOR, Union Gas, Enbridge, Optus) and higher education (Yale University, Penn State, Cornell)

Differentiators:

3 Combined Virtual Agent and Predictive Analytics supports complex use cases at scale

3 “Works like the best agents” consistently across multiple channels

3 Tools for dministration and performance measurement

3 Multiple successful deployments in key verticals


15


aivo (agentbot) US office:181 Fremont StreetSan Francisco, CA 94105USATel: +1-888-882-3418Founded: 2011

Headquarters:Ituzaingó 129, P4 - Of.ACórdoba – 5000Argentina Tel: +54 351 526 9250

BrazilAvenida das Nacões Unidas`3° andar, Saõ PauloBrazil 12.495Tel: +55 11 3958 7427

Employees: 30+Revenues: private companyMarket Cap: private company

Company Highlights:Aivo created and markets AgentBot, a platform for automated multi-channel customer care that has a broad customer base in Latin America, including Sky, Visa, AT&T, GM and Telefónica. It has expanded into the United States by opening an office in San Francisco. As of mid 2015, Aivo agents were in operation in 10 countries and handled over 100 million conversations each month across English, Spanish and Portuguese.

Key Executives:Martín S. Frascaroli, Founder and CEONicolas Ramos, Regional ManagerDaniela Sargiotto, Global Expansion ManagerLeandro Lopez Mazzarini, COOMarcelo Toranzos, CTO

16

Enterprise Intelligent Assistance ProductsAgentBot is Aivo’s core product. It operates atop the company’s internally developed natural language processing technology and can understand customer questions and recognize intent in order to resolve the customer needs using the company’s knowledge and customer data.

How AgentBot Works:Understands the Intention: Semantic engine with own dictionary, minings and regionalism database;

Long-Term Memory for natural conversations; Integration with third-party solutions (company systems)

Interacts and Resolves: Responsive interaction window with multimedia complements; Facebook and Twitter integration; Integration with SMS, Telegram and other messaging services; iOS and Android App; Integration with Salesforce, Zendesk and others; Open API

Learns and Evolves: Manages all the interaction content; Qualitative, Topics Analytics Tool; Learning tool

Revenue odels and ngagement t rategyEngagements typically start with a kickoff phase, during which goals and scope of the project are set and a minimum number of key issues are defined. Clients can use AgentBot’s tools to begin creating their own interaction database. The AgentBot integrates with other customer support tools, such as CRM, ERP, live chat and ticketing systems. Learning and improvement is based on baked-in analytical insights that identify missing answers and improves the customer experience.

Pricing models (as well as implementation options) are flexible to adjust to the company needs. Pricing is driven by the number of conversations per month carried out by AgentBot. Many current customers are large organizations, but Aivo has pricing plans for small and medium businesses.

Customers/Verticals:Featured customers and implementations include: Telefonica, Movistar, Visa, AT&T, Fiat, GM, AIG, Sony, Netshoes, Sky, Whirlpool, City of Buenos Aires, Anhanguera, MagazineLuiza.

Aivo focuses on Telecoms, Banks, Online Services, Utilities, Government and cable operators.

Differentiators:

3 Natural Language Understanding coupled with a constantly evolving dictionary and machine learning. It also uses “built-in analytics and statistics both to generate reports for business line managers and to improve the customer experience”.

Focus on the customers’ intention and experience. Offers human-like interaction (Empathy) and personalized responses, detects when to transfer to a live agent using integrations with Salesforce, Zendesk, Olark, LiveChat, Twilio and others.

Supports multiple digital channels and integrates with the company information.


3

3

17

Artificial Solutions Headquarters:Artificial Solutions Scandinavia ABÖsterlånggatan 43, 4 fl111 31 Stockholm+46 8 663 54 50

UK OfficeArtificial Solutions UK, Ltd 3 Riverside House Mill Lane Newbury, Berkshire (UK) RG14 5QSTel: +44 (0)1635 523267

Founded 1999Employees: 20+Revenues: private companyMarket Cap: private company

Company Highlights:Artificial Solutions coined the term Natural Language Interaction (NLI) as its area of specialization leading specialist in NLI. It has developed patented technology that enables people to hold two-way meaningful conversations with applications and services running on computers, mobile technology and other electronic devices in a humanlike, intelligent manner. Company has engagements in 26 countries with 200 clients.

Animation that defines service: http://www.artificial-solutions.com/campaigns/teneo/artificial-solutions-teneo/

Key Executives:Lawrence Flynn, CEOPeter Roost, COOAndreas Wieweg, CTOChris Bushnell, CFOAndy Peart, CMO

Enterprise Intelligent Assistance ProductsTeneo is the core software platform for building Intelligent Assistants as well as any form of natural language-powered service in multiple languages, across multiple devices and platforms. Creation can be done by the company’s own professional services team or, ultimately by clients themselves. The company has also generic mobile personal assistant called Indigo, which can be seen as a showcase for the capabilities of Teneo on mobile platforms (iOS, Android, Windows), though it can also communicate across multiple platforms and devices. The latest release of Teneo (version 4) focuses on natural language data analytics capability.


18

Four components to Teneo:

• Teneo Studio – enables professional services or enterprise personal to build the natural language app. It is highlygraphical and also allows drill-down into the logic to enable a business person to build an app.

• Teneo Interaction Engine (run time) – enables individuals to key in, speak or type a query and interprets it in 35languages (with the release of Teneo version 4) to determine the most appropriate response

• Teneo Analytics Suite – includes a “dashboard” that illustrates key performance indicators (KPIs) and providesinsights; Teneo Analytics Reporting which is a tool that analyzes pre-tagged data to optimize solutions andprovide insight; Teneo Analytics Discovery, which enables analysis of unstructured data (e.g. chat logs,trans ripts, social media, forms ; Teneo Analytics Informer, providing collection analysis and insight from hugeamounts of data; Teneo Analytics Real-time, which performs real-time analysis to support personalizedresponses based on machine learning.

• Teneo Language Resource – takes a statistics-based approach based on multiple years of experience andmillions of conversations to derive meaning from natural language input in multiple languages across a variety ofknowledge domains. TLRs enable super-fast implementation of new languages onto an existing implementation,by providing ready-made building blocks of native-proficiency language resources

Artificial Solutions also distinguished itself by being the first to build a mechanism for Personal Assistants (without deep subject knowledge) to facilitate communications with Intelligent Assistants or “specialists” which tend to operate as isolated solutions within an enterprise or search utility. Called the Teneo Network of Knowledge. If you think of MPAs (including Indigo) as ubiquitous resources that support millions of users, the Teneo N K can listen to a query and work out where the expert IA is for responding. For example, if you say “I’d like to book a flight” and your Personal Assistant knows that you have a loyalty program with American Airlines, it will put you in touch with American’s IA.

Revenue models and engagement strategyLicensing of Teneo Platform software varies by features employed and volume rofessional ervices are involved in initial deployment, identifying and aggregating information from a number of sources and building “understanding,” and for ongoing support of machine learning.

Customers/Verticals:With 200 implementations in 26 countries, Artificial Solutions has showcase customers in the following:

• Financial – Widiba Bank (Widdy), Co-operative Bank (Mia, internal aide to contact center staff),Banca Sella (Stella)

• Retail – IKEA (Ask Anna), Skruvat (auto parts supplier incorporates Teneo into Web-based interactions)

• Telecoms – Vodafone (Ask Holly), Telenor (Emma), NTT (HiKari Nishino), Kabel Deuthschland (Julia)

• Travel and Leisure – Connexxion Netherlands (Lisa), Statens Vegvesen Norway (IDA)

• Utilities – Yello Strom Germany (Eve), Deutsche Post (Jana)


19

Differentiators:

3 Unique in its ability to deliver solutions across multiple platforms, languages and devices from a single instance. Thus a single solutions can be expanded without requiring a new build each time.

3 Integrated Teneo Platform supports development, run-time, analytics, reports and refinement in anticipation of Natural Language Interactions via ebsites, mobile devices, automotive and home automation

3 Teneo Network of Knowledge supports communications between mobile personal assistants (PAs) and the proper Intelligent Assistant (IA)


20

Creative Virtual UK Headquarters:Creative Virtual Ltd165-167 Cannon Workshops5 Cannon DriveLondonE14 4ASTel: +44 (0)20 7719 8332

US OfficeCreative Virtual USA68 Southfield Avenue2 Stamford LandingSuite 210Stamford, CT 06902Tel: +1 203 276 0396

Other offices in Mumbai, Amsterdam, New South Wales, Singapore and Hong Kong

Founded 2004Employees: 140+ Revenues: private companyMarket Cap: private company

Company Highlights:

The company positions itself as follows:• Be everywhere your customers are at any time

o Help, contact us, product, sales, call center, live chato Mobile, tablet, web, Social, SMS and IVRo Enable users to Talk, Type or Tap

• Offer intelligent, personalized, contextual help – when neededo Utilize data (cookies, API) to know the customer and/or prospect

• First Interaction Resolutiono Right channel and smart escalation based on customer journey, escalation to Live Chat, IVR, Call Back

• Constantly improve based on voice of the customer and outcome tracking


21

Key Executives:Chris Ezekiel, Founder & CEOPeter Behrend, CTORichard Simons, CEO, Creative Virtual USA

Enterprise Intelligent Assistance ProductsV-Person, Creative Virtual’s software platform to support intelligent assistance. Its core capabilities are as follows:

• Knowledge Management: V-Portal

• Natural Language Engine: Virtual assistants (based on V-Person) give customers the sensation ofcommunicating with a “real” person who can hold entire conversations.

• Reporting & Analytics: V-Portal reports provide insights into customer behavior and interactions, includingVoice of the Customer (VoC) polls. Transcripts of customer conversations yield views of what customers aresaying, thinking and feeling. The analytics engine can also identify questions that have gone unanswered to helptailor future responses.

• Professional Services and Hosting: V-Portal can be managed by Creative Virtual’s professional servicesorganization in conjunction with internal staff. As a managed service, Creative Virtual can handle implementationand ongoing management using the V-Portal suite of tools, everything can be administered by internal staff orjointly in a “hybrid” arrangement.

Revenue models and engagement strategyLicensing of V-Person is based on conversations count or an unlimited conversation model. provide term or perpetual Licenses. V-Person can be client hosted or cloud. nstallation and setup services and an ongoing managed service if required.


22

Customers/Verticals:HSBC, Verizon, TWC, IHG, Lloyds Banking Group, E.ON, Autodesk, E*TRADE, Tesco, National Rail Enquiries, Allianz Australia, NIBC Direct, Citibank, RSPCA and more.

Differentiators:

3 V-Portal: Knowledge management tool drove purchase of V-Person by TWC, IHG, Chase andE*TRADE. Created based on experience working with major corporations, V-Portal can be used by clients or managed end-to-end by Creative Virtual.

3 V-Person: Creative Virtual claims more integrated/personalized solutions deployed than any othervendor. Available in desktop, tablet and mobile. Supports FaceBook, Twitter, Lithium and SMS implementations.

3 Professional Services Team: Highly knowledgeable in relevant vertical markets and with template installations to reduce setup time and improve outcomes.

3 Showcase deployments: B2B (AutoDesk), Telecoms (Verizon, TWC), Finance (Chase, E*Trade), Travel/Hospitality (IHG, NRE)


23


IBM Watson GroupHeadquarters:51 Astor PlaceNew York, NY 10003-7139Tel: 1-800-IBM-4YOU(1-800-426-4968) General Inquiries.

Founded 2014Employees: over 2,000Revenues: operating unit of IBM Corp (with over $24 billion annually)Market Cap: IBM Corp = $161 billion

Company Highlights:IBM Watson Group and its Cognitive Computing activities have the stated objective of enabling “a new partnership between people and computers that enhances, scales and accelerates human expertise.” IBM has steadily built the Watson Ecosystem, a set of efforts designed to attract large firms as well as independent developers with expertise in specific verticals, including healthcare, travel, finance, food services, etc. In April 2015, Watson Health Cloud will provide a secure and open platform for physicians, researchers, insurers and companies focused on health and wellness solutions. The HIPAA-enabled Watson Health Cloud will enable secure access to individualized insights and a more complete picture of the many factors that can affect people’s health.

Key Executives:Michael D. Rhodin, Sr. Vice President, GM IBM Watson GroupRobert High, CTO IBM Watson Group, IBM FellowStephen Gold, Vice President, CMO of IBM Watson GroupAlexa Swainson-Barreveld, Vice President, Watson Products and Solutions

Enterprise Intelligent Assistance ProductsIBM Watson Group transforms a set of cloud-based technologies into a grounded, funded, go-to resource for developing enterprise-based intelligent virtual assistants.

Watson Engagement Advisor - Empowers users to get questions answered.

Watson Discovery Advisor accelerates the discovery process by detecting patterns in large, unstructured and structured data sets. Can be trained to learn over time and ingest large, domain-specific corpora.

Watson Analytics – applies advanced analytics, available from the cloud to guide data exploration, automate predictive analytics and enable a presentation of results in a dashboard on automatically created infographic

24

Watson Explorer Promotes adoption of “Cognitive Computing” rather than “Intelligent Assistance” by simplifying an enterprise’s ability to deploy IBM’s foundational Big Data and Analytics solutions.

In terms of integration with enterprise customer care resources, it initiated a partnership with Genesys in June 2014 to integrate the Watson Engagement Advisor with the core Genesys Customer Experience Platform. The design implies that Watson could serve as a source of answers or recommendations for contact center agents.

Besides Genesys, other IBM Watson partners include SoftBank, Mubadala, The Cleveland Clinic, Elsevier Publishing and WayBlazer. They serve as a great indicator of the vertical direction IBM and partners are develop “Cognitive Solutions” for a number of vertical markets, Watson Health Cloud, Watson Travel Cloud and Watson Retail Cloud. The “Watson Developer Cloud” offers 29 cognitive services (at the time of publication) which enables both enterprise IT and independent developers to create intelligent assistants.

Revenue models and engagement strategyBased on experience, IBM defined a deployment model that started with a -week period during which a cloud-based instantiation of Watson would use input from product literature, training documents, Web pages and subject matter experts to build a core of Q&A Pairs that became the beginning of its cognitive development. Without mentioning price or deployment goals, IBM claimed that a positive ROI could be reached in roughly months, presumably based on the deflection of calls to live customer support agents.

Customers/Verticals:Watson addresses multiple verticals with strengths in financial services, healthcare and travel. Named customers include ANZ Bank (Australia), USAA and Deakin University.

Differentiators:

3 Extensive resources for deep understanding and analytics

3 IBM’s high profile promotion and funding, as well as the base of professional services personnel to support broad implementation

3 IBM Watson Developer Cloud and Ecosystem provides services and tools that enable app developers to deploy cognitive computing without building deep, core tehnology skills


3 IBM's Bluemix cloud portfolio can help developers build Watson-based solutions

25

Headquarters: 777 Mariners Island Blvd., Suite 220San Mateo, CA 94404Tel: +1 408-213-8771 (offices in France, Spain, Brazil and Chile)

Founded: 2005Employees: 50+Revenues: (private company)

Company Highlights:Founded in Spain, and moved to the Silicon Valley in 2011. Global agreements with Tickmaster, Groupon were major wins for the company. $2 million investment in April 2014 InverSur Capital. Inbenta takes a cloud-based approach to supplying AI-powered intelligent search to support customer care interactions and e-commerce transactions.

Key Executives:Jordi Torras, CEO and FounderFerran Surina, COOJordi Prats, CTOFerran Casellas, CFOMeeSun Boice, Country Manager, U.S.

Enterprise Intelligent Assistance ProductsCloud-based Platform: pplies artificial intelligence to support accurate responses to customer search queries. It markets primarily to large customer-oriented companies that have high levels of traffic on their ebsites, high-volumes of e-mail or calls to contact center agents. Its computational linguists identify the topics that are primary subject of queries and build a constantly updated database (or lexicon) of questions and answers that become the basis for self-service and assisted self-service responses. Q&A can take place through a search box on the company website, Web chat, email, or through Veronica, the avatar described below.

Inbenta Avatar: Called Veronica, employs AI-based semantic search technologies, speech processing and natural language understanding in near real time to provide a natural interface for customers or as prompts for live agents.

Inbenta Backstage: A reporting utility that lets functional executives observe how well they are meeting their business objectives. Backstage lets them observe how successful the system is at answering incoming questions, identify cross-selling and upselling opportunities, how responses are clustering or other mechanisms that reveal the most relevant search topics or sales opportunities.


26

Revenue models and engagement strategyInbenta charges an initial fee for professional services involved in initial setup of the service by the assigned conversational linguist. Then it follows a pay-as-you-go or SaaS model for use of the core technology, including access to the backstage and to the linguists for constant refinement.

Customers/Verticals:Inbenta market focus is horizontal in nature, addressing high volume query-and-response opportunities. Customers include Groupon, Ticketmaster, Schlage, Act-On, coupa, Franklin Planner, LiveNation, Trane, and Banco Santander, among others.

Differentiators:

3 Focus on Q&A and e-commerce requests

3 Matching each customer with a “conversational linguist”

3 Speed to implement and ability to scale


27

Interactions LLC Headquarters:31 Hayward Street-Suite EFranklin, MA 02038U.S.A.Tel: 866-637-9049

Founded: 2004Employees: over 1,000 (230+ full-time and 800 “intent analysts”)Revenues: private company

Company Highlights:Interactions develops award-winning speech and natural language technology solutions used by Fortune 500 companies worldwide. Its patented Adaptive-Understanding™ technology seamlessly integrates automated speech recognition (ASR) platform with human intelligence to provide understanding capabilities. Interactions provides conversational virtual assistant solutions to top brands and delivers speech and multi-modal interface solutions to power any device, service or application.

The company’s solutions are delivered across any device with type, touch or talk capabilities, and have handled more than one billion transactions to date. Rooted in natural speech recognition and built to adapt to human conversation, Interactions’ solutions have delivered cost-savings and increased efficiencies in sales and support for some of the largest companies in the world. Interactions LLC was founded in 2004 and is headquartered in Franklin, Massachusetts with additional offices in Indiana, New Jersey, New York and Texas.

Key Executives:Mike Iacobucci, President and CEODave Parkinson, COO and EVP, SalesYoryos Yeracaris, CTOPhil Gray, EVP Business DevelopmentJay Wilpon, SVP Natural Language Research

Enterprise Intelligent Assistance ProductsInteractions provides a full range of text-based virtual assistant solutions, fully hosted Chat and SMS solutions as well as mobile experiences that combine talk, touch and text. Core offering is human-like, IVR-based interactions supported by natural speech recognition and built to adapt to human conversation.

Interactions’ primary product offering, Adaptive-UnderstandingTM, uses the company’s patented iProxy technology to combine automated speech recognition with Human Assisted Understanding. This technology supplements speech recognition and improves language understanding, yielding over 95% accuracy in customer interactions. The company now handles 450 million calls per year across 47 customers.


28

Speech and Language Technologies (SLT), which include:

ASR: Automatic Speech Recognition - Interactions ASR provides speech to text conversion using uniquely generated acoustic models that predict how words sound in a given environment, e.g., on a mobile phone. These acoustic models are combined with language models and pronunciations for exceptional accuracy. Interactions ASR solutions are adaptable to specific domains, environments, and languages.

TTS: Text to Speech - Interactions TTS starts with a database of high-quality recorded speech produced under optimum conditions. The individual sounds in speech (phonemes) are labeled to select the best set of sounds for natural sounding spoken words. Voices can be male or female. Both North American English and Latin American Spanish are supported.

VB: Voice Biometrics - Interactions Voice Biometrics uses the human voice to provide speaker recognition and authentication. Using the unique characteristics of each voice, Interactions Voice Biometrics is used to verify a speaker’s claimed identity (Verification) or identify a speaker from a known group of people (Identification).

LU: Language Understanding - Interactions LU turns data into information. By flagging the people, locations, topics, intentions, etc., found in any segment speech or text, Interactions LU provides the basis for actionable intelligence applications and business processes.

Combining Speech and Language Offerings - The integrated Interact Speech and Language Platform makes it simple to combine services to optimize applications. The common code and development tools allow the easy integration of several different modules to provide sophisticated and differentiated applications.

Revenue models and engagement strategyInteractions follows a “success-based pricing” strategy. Interactions only charges for successful transactions that add business value to clients. A successful transaction is defined as the completion of a discrete task, such as authenticating a caller or completing a reservation. In this model, there is a shared interest for continuous improvement. Whereas most traditional voice self-service systems charge on a per-minute basis regardless of success, Interactions only charges when delivering value to clients.

Customers/Verticals:Interaction ’ featured clients and customers include Hyatt, Best Western, Humana, TXU Energy, Lifelock, Asurion, EyeMed, AllConnect, and others.

Differentiators:

3 Human-assisted understanding, based on “adaptive understanding technology” that allows for the right understanding technology to be used in the course of a conversational interaction

3 Patented way to balance automated services, human understanding and intelligent transfer (or escalation) to live agents

3 An integrated technology stack merging internally developed adaptive understanding with full suite of speech processing, natural language understanding and voice biometrics, formerly AT&T Watson.

3 Expanding capabilities to automotive, mobile and “smart” homes.


29

KasistoUS Headquarters:110 5th Avenue 5th Floor New York, NY 10011Tel:

Founded: 2013 (spun-off from SRI International in 2014)Employees: privateRevenues: private

Company Highlights:

Key Executives:Zor Gorelov, CEOSasha Caskey, VP of EngineeringDror Oren, VP of ProductRuth Brown, VP of Operations

Enterprise Intelligent Assistance ProductsKasisto offers an SDK that enables banks to offer services through an automated “Virtual Specialist” as a capability of their branded smartphone apps. It is a cloud-based platform with “deep knowledge” of banking services, so consumers can use human-like conversational interactions, through voice or text, to access financial information and perform complex transactions on their mobile banking application as if they were working directly with a human financial advisor, customer care agent or bank teller.

Revenue models and engagement strategyThe initial engagement model is a paid proof-of-concept or pilot followed by an annual platform subscription. Size of contracts is based on an “Active User Fee” (defined by the prior 90 days). Success-based “e-commerce fees” may also be added, based on successful upsell or redeemed offers.

Customers/Verticals:Kasisto’s initial focus is banks and financial institutions where it has developed deep knowledge and understanding. The initial intellectual property was developed as part of a deployment by BBVA. Undisclosed live pilots in the U.S. and Asia.


30

Differentiators:

3 Deep domain knowledge: acting as a “banking specialist with over 75 use cases (initially)

3 Conversational UI/Reasoning: requiring no coding by banking implementers

3 Multi-modal input and output: not a “voice banking product” supports voice, text, touch through smartphone app

3 Customizable & cross channel: designed for banks to private label with their own brand


31

Next ITHeadquarters:421 W. Riverside Ave., Suite 1150Spokane, Washington 99201Tel: +1 509-242-0767(office in São Paulo, Brazil)

Founded: 2002Employees: 150Revenues: (private company)

Company Highlights:Founded in 2002, Next IT is one of the largest and longest standing providers of virtual assistant technology. Alme, its natural language platform, has demonstrated the value of AI as virtual experts in specific subject matter domains, especially healthcare, travel and financial services.

Key Executives:Fred Brown, CEO and FounderRick Collins, President EnterpriseMitch Lawrence, President, HealthcareTracy Malingo, EVP, Product & DevelopmentJennifer Snell, VP of Marketing

Enterprise Intelligent Assistance SolutionsThe A Platform: provides domain-specific natural language understanding based on input from a number of sources. The Alme Administrative Tools (embedded or integrated in the Intelligent Assistance Platform) includes:

• Alme Response Manager: nables enterprise personnel to review, edit and instantly publish responsechanges.

• Alme Review: rovides a utility that can rate conversations to assess user experience

• Alme Explorer: nables users to search conversation logs and expert parts that match specific criteria

• Alme Reports: nables user to evaluate aggregated metrics in near-real time to gain insights aboutusers

The Alme UI has many options enabling users to talk, tap or type input into smartphones, PCs, Webchat and various messaging platforms. The architecture is designed to solve the overall problem of providing consistent responses over time and across multiple devices.


32

Revenue models and engagement strategyNext IT offers subscription, transactional, or perpetual licenses for Alme and supports cloud and on-premise deployments

Customers/Verticals:Verticals are healthcare, travel, finance, communications and governmentCustomers with vitual assistants:

• Aetna – Ask Ann engages with 50% of new visitors, leading to a decrease of 29% in calls to contact center withincreased user satisfaction due to better personalization.

• Alaska Airlines – Ask Jenn, effectiveness enabled company to shut down live chat. Is the “first option” on“Contact Us” page.

• SWBC – Ask Emily supports over 20,000 borrowers per month with an 85% engagement rate and 90% first-timesubmission success rate

• U.S. Army – SGT STAR has answered over 13 million questions to date; eliminated live chat; launched on mobileand Facebook

Differentiators:

3 Vertical-specific, solution oriented commercial deployments of AI

3 Success in custom domain model development

3 Proven integration with “systems of record” and compliance with existing IT policies


33

noHoldHeadquarters:225 S Milpitas BlvdMilpitas, CA 95035Tel: 408 - 946 - 9200 ext. 311

Founded: 1999Employees: PrivateRevenues: Private, claims profitability since 2003

Company HighlightsnoHold concentrates on Web based self-service solutions. Its stated mission is “to deliver real answers to real questions – real fast.” Its product design puts emphasis on making it “simple to use, easy to implement and as close to human as you can get.” noHold turns automated customer support into cognitive customer interactions. With over 15 years of experience in the industry, main focus continues to be to create self-service systems that increase value for its customers.

Key ExecutivesDiego Ventura, CEO and Founder

Products/ServicesSICURA™ is noHold’s Knowledge Management Platform. SICURA™ is cloud-based, making it easily managed from anywhere around the world. The platform also includes authentication to secure company content, while providing an enriched experience for users and employees. At the core of SICURA™ is ArticlesManager™, noHold’s native XML database that allows for the management of content in different formats such as, PDF, Word ocuments, videos and images. One advantage is the content can serve multiple audiences from a single repository of information.

The aggregated knowledge can be delivered in two primary ways:

• Through a Virtual Assistant – an interactive and diagnostic, customer facing application

• Through Search++ - noHold’s Natural Language search engine

Added to this platform are tools that help administers manage content and capture Metrics.

• Metrics – the component of SICURA™ that provides statistics about platform usage. This tool lets administratorsknow metrics on such things as how many people logged in, what questions were asked and what questions thesystem couldn’t answer.

• Content Maintenance – this component allows administrators to identify knowledge gaps, measurequality of content and take appropriate action. Content Maintenance has an easy to use interface, so evennon-programmers can use it.


34

• Content Editor – allows administrators to edit and create new content; even non-programmerswith little experience can use.

• Workflow - describes the process and tools administrators can use to improve the performance of theVirtual Assistant and Search ++.

Another feature added to the platform to enrich user experience and company satisfaction is its capability to integrate with external data sources via an open API set. This open API set enables connection with:

o Customer Relationship Management (CRM) systemso Live Chat platformso Content Management systemso Knowledge Management systemso External Databases

Virtual Assistants are customer facing applications that work similar to live chat, but instead of chatting with a live agent, the end user is interacting with artificial intelligence. The Virtual Assistant is designed to sustain a high volume of concurrent sessions and can communicate in different languages depending on the content stored in its knowledge platform (SIUCRA™). Virtual Assistants are extremely versatile and can be leveraged across various channels to provide users with consistent support anytime, anywhere. For example, through websites, mobile sites, online stores, brick & mortar stores, web chat, call centers, internal help desks, reseller sites and social media networks. Some key features of noHold’s Virtual Assistants include:

• Natural Language Processor (NLP)• Inference Engine• Metrics• Cloud-based• Application Programing Interface (API)• noHold Connect• Multichannel• Customizable• Multilingual

noHold Connect utilizes APIs to allow one company’s Virtual Assistant to communicate with another’s. A real world example would be a user starting a conversation with Bella (Cincinnati Bell’s Virtual Assistant) about a Dell product. Bella would recognize that Dell’s Virtual Assistant would be better suited to answer the question, and passes the conversation on within the same interface. This is particularly useful for companies that have an extensive ecosystem of partners.


35

Revenue and Engagement ModelsnoHold is based on a Software as a Service (SaaS) deployment model. It has structured a monthly recurrent revenue model with pricing that includes a onetime implementation fee, usage fee, and content maintenance provided by professional services team. The engagement model is dynamic. The company had initially worked with companies to develop internal Intelligent Assistants which were largely horizontal in nature. After succeeding with the initial application, prospects showed greater willingness to scale up to customer-facing implementations.

Customer/VerticalsAlienware, Cincinnati Bell, Cisco, Computershare, Dell, Lenovo, McAfee, Toshiba, ViewSonic, Webroot

Differentiators:

3 Simplicity of approach (ROI focus)

3 Support of “open source” approach to code sharing and application development

3 Roster of go-to-market partners: LivePerson, Moxie Software, Salesforce, Microsoft, Lithium


36

NuanceWorldwide Headquarters:1 Wayside RoadBurlington, MA 01803United States

Tel: 781-565-5000

Founded: 1992Employees: 14,000Revenues: approx. $2 billion in FY 2014Market Cap: $5.28 billion

Company Highlights:Nuance enables companies to engage with consumers in a way that’s personalized and conversational by leveraging decades of experience in speech, natural language understanding and artificial intelligence. Nuance delivers this conversation through multichannel experiences that integrate proactive engagement, voice biometrics, conversational IVR, and multichannel virtual assistants. Nina is Nuance’s intelligent multichannel virtual assistant, a digital persona who delivers personalized customer service via a human-like conversational interface. Nina leverages Nuance’s technology leadership and expertise in voice, natural language understanding, conversational dialogue and advanced resolution techniques, to deliver a multichannel, automated customer service experience for the consumer and the enterprise.

Key Executives:Paul Ricci, Chairman and CEORobert Weideman, EVP and GM, Enterprise DivisionVlad Sejnoha, Senior VP and CTORon Kaplan, VP and Distinguished ScientistGregory Pal, VP, Marketing, Strategy & Business Development, Enterprise DivisionTony Lorentzen, GM and VP, Enterprise Cloud Solutions

Enterprise Intelligent Assistance ProductsNina is an intelligent, multichannel virtual assistant, serving as a digital persona who delivers personalized, effortless customer service via a human-like conversational interface via both speech and text. Personas are customizable and serve to support a brand’s identify, as well as provide a natural, conversational interface that understands what customers need and provides answers or navigates them to the right place quickly and accurately.

The hosted Nina Multichannel Platform is a common platform supporting both web and mobile channels, and employs the following components:


37

Nina Virtual Assistant: Technology, hosted by Nuance, that provides the power and intelligence to Nina. Algorithms developed by Nuance capture the intent of the conversations to provide rapid and relevant answers to customers’ queries. Nina can be broken down into:

• Applications & Services (Text-to-Speech, Dictation, Natural Language Understanding, and Dialog Management)• Data (Dialogs, NLU models)• Tooling (IQ Studio – common dialog tool and NES – Nuance Experience Studio, a common grammar tool)

Promoted features include:

• Nina IQ Studio: Builds dialogs quickly and easily; no programming expertise required. NIQS is a common tooland dialog model for web and mobile solutions built on Nina Multichannel. It keeps information offered throughNina up to date, accurate and effective, using machine learning to identify successful interactions and correctunsuccessful responses. It also employs tools and incorporates feedback from “Voice of the Customer” polls tocustomize dashboards and generate dynamic reports that deliver insights within minutes.

• Common architecture: llows customers to leverage Natural Language investments across channels, bothWeb-based and mobile

• Nuance Experience Studio (NES) (a companion to IQ Studio). Nuance launched Nuance Experience Studio toallow customers to quickly deploy Natural Language, a critical component of any Virtual Assistant. It has beenexpanded as follows:

o NES allows non-experts to build multi-slot NLU Grammars quickly and easily; no speech-scienceexpertise required.

o Common NLU grammar tool for web and mobile solutions built on Nina Multichannel.

o NES is tightly integrated with Nuance IQ Studio, with the interoperability allowing a customer to easily publisha Virtual Assistant into production. The combination allows for a “learning loop” capability with changes easilypublished as the Virtual Agent evolves

• Nina PCI certification: Addresses requirements of nterprise customers and expands the use cases Nina cansupport (e.g. Shopping Cart Applications). PCI compliance allows Nina to expand beyond traditional customerservice to verticals like Banking and Retail (e-commerce).

• Integration with leading third party ustomer ervice providers: xpanding Nina’s ability to become a Self-Service Agent for Chat and becoming more intelligent by integrating with knowledge systems

• Multiple Language Support: Increased the number of languages allowing international growth. As a companyacross their Mobile and Enterprise divisions, Nuance supports over 50 languages, and over 100 voices

• Three “out-of-the-box” User Interface alternatives: (embedded, pop-in and beam) for different customerrequirements/objectives


38

Revenue and Engagement ModelsInitial start-up fees and then licensing of core software (varies by size of company, size of customer base or activity. Models vary by customer but in general include a transactional component (tiered session or per session models) and ability to provide a gain share model where the success of the Virtual Assistant is tied to pricing.

Customers/Verticals:Financial ervices, commercial products and restaurants, e.g. USAA, U.S. Bank, Tangerine Bank, Coca-Cola, Domino’s, up2drive, a division of BMW Financial Services, Swedbank, ING NL

Differentiators:

3 History of delivering compelling and conversational self-service globally with proven NLU engine. This includes a global footprint and broad language coverage.

3 Nuance on Demand platform: providing cloud deliver that integrates the latest technologies and processes more than 4 billion interactions per year across three data centers in North America.

3 System integration services: demonstrated ability to integrate VAs to Knowledge Bases, Live Chat, and other Contact Center Technologies.

3 Enterprise Level Offering, emphasizing leading service levels and PCI certification and ability to scale,

3 Multi-channel infrastructure delivers a consistent and persistent experience, across multiple channels and self-service offerings


39

SmartActionHeadquarters:390 N. Sepulveda Blvd., Ste. 2150 El Segundo, CA 90245

Tel: +1 310-776-9200

Founded: 2009Employees: privateRevenues: private

Company Highlights:SmartAction Company, LLC was founded in 2009 by entrepreneur Peter Voss to commercialize the artificial intelligence technology that he had been developing since 2002 at Adaptive A.I., the parent company of SmartAction. Since 2009, SmartAction has steadily grown, from initially attracting only small businesses to recently acquiring many more Fortune 500 companies such as Belkin, Hyundai Motors, and Royal Caribbean Cruise Line. Company goal to handle the most complex calls as possible in order to more accurately allocate agents’ time while simultaneously maintaining a high level of customer satisfaction.

Key Executives:Tom Lewis, CEOSteve Prodger, Senior Vice President, SalesTas Dienes, Chief Technology OfficerMichael Vanca: Senior Vice President, Operations

Enterprise Intelligent Assistance ProductsThe company’s core solution is IVA™, Intelligent Voice Automation. It is a cloud-based artificial intelligence voice self-service platform. The proprietary artificial intelligence engine, or “brain,” was developed by SmartAction’s parent company, Adaptive AI, as a larger R&D mission to build AI prototypes.

IVA is a hosted, cloud-based resource that can handle call volumes for medium-to-large companies. Core speech recognition is licensed from Nuance, but its infrastructure, underlying assumptions and framework are distinctive when compared to other IVR systems. All calls and speech-based interactions are fully-automated. Its combination of automated speech recognition and AI with natural language processing throughout every call make highly complex interactions possible. In addition, IVA has the ability to remember information from the current call (short-term memory) and recall information from previous conversations (long-term memory). As such, conversations and call flows end up more similar to those with a live agent than to those with another IVR. The following other features/benefits are part of solution:

• Extensive segmentation, branding, personalization within call flows• Integration with most ERP and CRM systems• PCI and HIPAA certified/compliant• No end of life – ustomers never have to pay for upgrades or enhancements


40

Customers/Verticals:Offers solutions for 12 industries: financial services, government, healthcare, insurance, media & entertainment, retail, services, shipping & logistics, technology, telecommunications, travel, and utilities. IVA allows for fully customizable solutions, including the conversation flows of different call types as well as the business logic directing each call.

Differentiators:

3 Proprietary, fully-automated handling of complex phone calls, claiming “We make a phone call effortless.”

3 Positions against speech-enabled interactive voice response (IVR) systems. Offering extensive segmentation, branding and personalization within call flows

3 Integration with most ERP and CRM systems

3 Low upfront investment. Short development and deployment cycles.


41

VerbioU.S. Headquarters:2225 E Bayshore Rd, 200Palo Alto, CA 94303

Founded: 2006Employees: privateRevenues: private

Company Highlights:Verbio offers a suite of software and services to support smartphone-based and multimodal use cases. The company was founded in Spain and has a customer base in several European countries and across Latin America. It is focused on delivering a packaged Voice Portal that includes:

• Automated Speech Recognition and Text-to-Speech Synthesis: To support accurate understanding of spokenwords and human-like responses.

• Speech and Text Analytics: For problem detection, fraud loss reduction, workforce optimization and compliance

• Voice Biometrics: To support rapid caller identification and authentication and to detect fraudsters

• Sentiment Analysis: To support empathetic handling of customer interactions by agents or automated systems

• Natural Language Understanding: To identify the purpose of call and the intent of the caller quickly and accurately

Verbio’s customer base spans commercial banks and financial services companies, government agencies, travel and hospitality and large retailers. Its technologies have been integrated into the telecommunications and care fabrics of large telecom service providers, including Telefonica and Teletech. The company’s roster of clients includes BBVA, Agencia Tributaria, Europ-Assistance, the City of Barcelona, Petrobras and the Council of São Paolo..

Key Executives:Co-owner: Carlos PuigjanerCo-owner: Antonio Terradas

Enterprise Intelligent Assistance ProductsThe Verbio Voice Portal is designed to carry on conversations with customers, facilitate smooth transition from self-service to assisted service, detect anomalous occurrences in real-time and rapidly understand and ascertain each caller’s intent as well as sentiment. Its suite of software and services support smartphone-based and multimodal use case


42


Revenue models and engagement strategyVerbio’s platform can be deployed on premises or following a SaaS model, depending on customer needs. There are APIs to make core technologies available to application developers. Verbio has flexible terms for licensing software, based on company size or pay-per-use. Engagements often include professional services to develop initial specifications and then provide ongoing maintenance and fine-tuning

Customers/Verticals:Verbio’s customer base spans commercial banks and financial services companies, government agencies, travel and hospitality and large retailers. Its technologies have been integrated into the telecommunications and care fabrics of large telecom service providers, including Telefonica and Teletech. The company’s roster of clients includes BBVA, Agencia Tributaria, Europ-Assistance and the City of Barcelona.

Differentiators:3 “Empathetic Voice Portal”: includes natural language understanding, conversational voice biometrics,

speech and text analytics, sentiment detection.

3 “4 S’s”: speed, simplicity, security and sensitivity

3 Conversational voice biometrics: provides passive identification of callers

42

About Opus ResearchOpus Research is a research-based advisory firm providing critical insight and analysis of enterprise implementations of software and services that support multimodal customer care and employee mobility strategies. Opus Research calls this market “Conversational Commerce” with tailored coverage and sector analysis that includes: Self-Service & Assisted Self-Service, Voice & Call Processing, Assistance, Mobile Search and Commerce Voice Biometrics. www.opusresearch.net

For sales inquires please e-mail [email protected] or call +1 (415) 904-7666.This report shall be used solely for internal information purposes. Reproduction of this report without prior written permission is forbidden. Access to this report is limited to the license terms agreed to originally and any changes must be agreed upon in writing. The information contained herein has been obtained from sources believe to be reliable. However, Opus Research, Inc. accepts no responsibility whatsoever for the content or legality of the report. Opus Research, Inc. disclaims all warranties as to the accuracy, completeness or adequacy of such information. Further, Opus Research, Inc. shall have no liability for errors, omissions or inad-equacies in the information contained herein or interpretations thereof. The opinions expressed herein may not necessarily coincide with the opinions and viewpoints of Opus Research, Inc. and are subject to change without notice. Published August 2015 © Opus Research, Inc. All rights reserved.


Documents

Decision-MakersÕ Guide to Enterprise Intelligent