The ecosystem for Intelligent Assistance is definitely maturing. As a result, executives from a broad spectrum of business units are involved in the decision to incorporate virtual agents, chatbots and other varieties of IAs into digital marketing and customer support strategies. Savvy decision makers, benefiting from the experience of peers who have made the leap to IA, are asking the most pragmatic of questions surrounding implementation:

• “What works out of the box?”
• “How much customization do I have to do on an ongoing basis?”
• And (perhaps most importantly) “What’s in that [black] box, anyway?”

In response, Spokane-based Next IT has a new service offering that takes the lid off of the “black box” by making its library of 90,000 business intents and 165,00 unique actions garnered for 11 vertical industries commercially available.

Next IT offers businesses a chance to hire and deploy virtual agents or bots that are informed by a the collective intelligence and experience derived over more than a decade of IA deployments and assembled into a set of “Alme Libraries.” Over the years, those libraries have evolved to include more than 20 million “labeled user questions”, encompassing more than 750,000 unique terms that map to the above mentioned 90,000 intents. This means that today’s IA developers can streamline initial development and shorten the time it takes for a new assistant to understand and respond to what individuals say or text to an IA.

Two sets of tools are made available directly to enterprise executives and developers under the Alme brand.

Alme Conversational Intelligence has elements that perform three core functions.

• “Cue AI” ingests massive amounts of data (unstructured and structured) to determine where a bot or virtual agent will create the largest return on investment (ROI).
• “Prompt AI” then steps in to tag or categorize large amounts of natural language input to a limited number of user intents.
• Then “Trace AI” applies continuous machine learning techniques to identify risky utterances and trends in order to improve responses over time.

Alme Conversational Platform is the run time environment that supports Conversational Intelligence. It features the company’s flagship software, Context IQ Engine. The individual modules integrated into the Context IQ Engine include:.

• Next IT’s Prompt Predict Engine – which uses machine learning to suggest prompts and is recommended “for implementations that do not require full personalization and contextual integration”
• Author AI – a content management system that enables business users to access and manage digital interactions and responses across multiple endpoints and contexts
• Conversation Designer – a visual tool for designing conversations.
• Alme Lab – which is an integrated development environment (IDE) that Next IT’s personnel have used to model the precise language a bot or intelligent assistant will use to understand utterances. This includes requesting clarifications through multi-step mixed initiative dialogue or incorporating context and third-party data to personalize understanding and experience.

Opus Research has long known that taking an “open” approach to Intelligent Assistance is a must in order to promote high levels of accuracy and efficiency when trying to understand natural language input. The more data – in the form of previous utterances and contextual data – that is brought into conversations, the higher the probability that an IA will recognize and intent and respond appropriately. Next IT is making those data elements available to enterprise personnel and providing them with tools to integrate them into Intelligent Assistant offerings.

As if to validate Next IT’s business-unit-friendly approach, cloud-software giant Oracle highlighted its own set of “Artificial Intelligence Software as a Service.” at the annual Oracle Open World gathering. As cloud-based deployments grow more extensive and complez, Oracle sees more vital functions for virtual assistants. In a couple of keynotes, company founder Larry Ellison observed that only machine-based intelligence has the capability to detect patterns of fraudulent activity and attendant security risks at scale.

To prevent fraud loss and loss of company reputation, companies will have incentive to embrace machine learning and artificial intelligence to fight fraud. While they are at it, don’t be surprised if business unit executives show how CRM systems and Contact Centers, which are seeing complex business processes and huge sets of data of their own, will argue that they have a growing number of incidences that necessitate the use of algorithms.

As company president of product development, Thomas Kurion, explained to analysts, “We’re not replacing business rules or people. We are augmenting them.”

Categories: Conversational Intelligence, Intelligent Assistants