Tens of millions of real problems automatically solved and the unique ability to answer online users’ issues, stand behind the InteliWISE industry leading natural language processing (NLP) technology.
The recent advances of our Natural Language Understanding enable us to deliver instant, automated response solutions, incl. Virtual Assistants and Virtual Agents, that engage users in meaningful virtual conversations, without involving human assistance, efecting in solved problems and optimized customer service operations.
Our Virtual Assistants revolutionize the way that millions of individuals interact with on-line banks, e-shops or public administration portals and systems. They engage in natural, virtual conversations, through emulation of human knowledge and behavior.
How our technology difer:
- Quickly and accurately determines the problem and intent of an issue, based on users’ everyday language, either typed or spoken
- Provides instant, actionable answer or guides user to a solution with emulation of a human dialog
- Puts through to a live chat agent at any given moment, whenever problem is really complex
- Provides constant quality of a support across channels, incl. built-in mobile / tablet applications
- Has been proof – tested with tens of millions of virtual (man – machine) interactions
InteliWISE provides a unique combination of an Automated Response System, powered by a Natural Language Understanding, with escalation to a proactive Live Chat support. See how it works:
Additionally, the system ofers an unbeatable set of quick set-up tools, making it easy to deploy for any type of customer.
The key components are:
- The Natural Language Processing Engine
- The End-User Presentation Tool
- The Knowledge Base & Script Management
- The Content Authoring and Administration Panel
- The Dynamic Data Exchange & Integration Module
- The Reporting Tool
- The Proactive Live Chat Module
Natural Language Processing
Best-in-Class Industry Technology, powering the natural man – machine interaction, combining a number of automated response methods, both for text and spoken interactions.
InteliWISE NLP Engine is a statistical classifcation with text correction and sense matching. However, it also utilizes syntactical rules reduction, which means it is a hybrid (statistical & rule based) approach to language processing.
The Engine use compound classifers and spanning – meaning weights and text distances. It is based on large text corporas (eg. British National Text Corpora) with conversational transcripts and ad hoc rules. Finally it uses semantic relations such as synonyms, antonyms, etc. InteliWISE NLP Engine is also able to handle misspellings and grammar errors.
We utilize proprietary correction rules, automated text corrections based on text corporas and also semantic sense matching.
The paradigm of choosing the best pieces of solutions without gaining worse re- sults and the accurately predicted boost of semantic nets‘ popularity as well as their feedback to the structure of Internet, let us precisely identify the optimal paths of further research. A lot of time was spent on consequently joining chosen pieces into one hybrid system. Below we present the bene ts gained from choosing the best parts of existing solutions. Our system is not build upon separate pieces of solutions – our system is a unique hybrid with characteristics similar to other known approaches. – said Marek Trojanowicz, CTO.
The key features of our NLP Technology enabled delivering the following:
Intent problem understanding
After quickly and accurately determining the subject and intent of an issue, the NLP-powered virtual assistant guides consumers to right answers. It also:
- Understands slang, misspellings, acronyms, and grammatically incorrect words or phrases
- Deploys the right social protocol for human – like behavior incl. greetings and farewells, criticisms, compliments
- Delivers social engagement capabilities beyond business centric dialog
Handling spelling/grammatical mistakes
At input level (i.e. customer typing a question), data is preprocessed with spell correction. 2 methods are being applied:
- Using external systems e.g. Google search and its ‘suggestions’ feature
- Using our stored data (text corpuses based on proper dat a/e.g. BBC text corpuses or client provided text corpuses).
When using stored text corpuses, the Engine compares user input with stored structure to check the spelling and correctness of text and suggest corrections for further investigation of internal engines (as alternative variants of user input).
Synonyms and prefixes
Since there are many ways of asking questions, after text pre-processing, the Engine composes alternative variants using synonyms and pre xes. Synonyms and antonyms give our solution the power of generalization over many language constructs – and do not change the meaning of user input (positive and negative weighting).
After creating many language constructs we remove as many parts of a sentence as we can without changing the meaning of it. We have observed that many language constructs are very redundant and do not change the meaning and the quality of matching process (not removing them could make the process of result factorization impossible or asymmetric).
Finally our technology analyses all statistically matched rules to choose the best one.
Handling ambiguities – clari cations and scenarios
When user input is not clear enough or client needs to see the process separated in multiple steps (questions) – the Engine may suggest scenarios. User can get the clarifying questions or he can be locked on the path or provided the “jump-out” possibility to leave the process at any stage. User is presented with suggested next steps and “feels” that system is driving him in the right direction. There can be as many scenarios built on the same data, as we need. We can reuse once entered data to design many processes – thus not replicating stored knowledge and making it possible to fast correct and edit.
Broad coverage refers to having a large range of domains covered by a system, for example attempting to cover all possible search topics. Using Contextual Auto Complete we maximize the success of a rst hit response.
Use of context
Different contexts can be analyzed before the answer is delivered, incl.:
- History of the current conversation
- History of prior conversations by current user (useful for personaliza- tion/preference setting over time)
- User interaction data – collection of interactions across many users
- How and Where the interaction started on a page
By emulating the fow of human conversation, InteliWISE solution quickly and accurately determines the problem, and guides to solution or puts through to a live person.
It’s just like mapping the live agent script and Q&A and applying them onto a friendly, Virtual Assistant.
InteliWISE Dialog enables to emulate the fow of human conversation, in both tone and content. Technically, it is a set of rules spread over the part of Knowledge Base.
As for the dialogue management, main technology platform and elements enable:
- Intelligent search, getting into the intent and the meaning of the questions, powered by AI, Natural Language Search engine;
- Interactive chat, with scripted dialogue, that narrows down search options, guiding user to the problem solution, not only displaying a set of answers;
- The use of an user-friendly interface, with or without avatar, enabling customer to type in the question in their own language and easily track down the response;
The dialogue is designed in a directional graph approach with the ability of assigning weights and locking paths to ensure step-by-step dialogue. It is also context-based and use conversational history for that purpose, so the next step can depend on previous questions. Moreover, the system is able to provide user pro ling based on login id, access level or any other feature that is sending at the beginning of conversation. Finally it is automatically integra- ted with the website so the dialogue may also depend on the page or section that the user currently visits.
Handling “don’t knows” and errors
First of all – we do not let the user too much possibilities to be left without knowledge and our service – while typing the is presented with “auto completed” sentences which suggest one of best matched questions the system knows about in regards of typed words. It gives the opportunity to stop typing as soon as possible and choose from “pre- constructed” questions and access the knowledge article as well.
If the user constructs the questions that confuse our system, we have many ways to cope with. We have multi-level of “don’t know” templates, tracking how many times user confused the system to change the system reaction. We have possibility to automatically switch the user to the real person live chat session to help him better than automa- tic system can operate. When using data driven scenarios we can react di erently (in pre-designed manner) at each stage of clari cation process.
User is never left without any suggestions even if we use “don’t’ know” template as a system answer. We always pre- sent best matched suggestions based on user input – even if they are far from the user asked, they conform to some parts of user input in syntactical or statistical manner.
However, if the system or knowledge base is not capable to provide a right answer, our system has a built-in live chat feature, that can escalate any chat to a knowledgeable live operator.
Depth of conversational capability
The system’s benchmark is based on number of elements, incl.:
- Human-like system
- Converses only when necessary, to help user get answers as soon as possible
- The ability to provide variable response to the same question, in a conversation with the same user.
- Capability for more complex dialogs than simple Q/A (i.e. change topic then return to original topic, system has it’s own goals projected by the scenario)
- Contextual Auto Complete – provides prompts for answers
Customized Knowledge Base
Providing over 95% of right, fully automated responses, requires dedicated Knowledge Base, populated from Q&A, help and other existing content.
Knowledge Base contains information that is used by the engine to present the best answer to the user.
Knowledge in InteliWISE System is structured in separable knowledge bases with given weights. We can deploy new solutions as mixture of existing data properly weighted (trained). We can make some language constructs more important than other by intelligent weighting (by statistical learning or by weighting by people). We separate ‘pears’ and ‘apples’ – we do not store all data in the same manner.
We organize data in “usage clusters”:
|Domain or client specifc Knowledge
|Spell correction data
||Synonyms and antonyms
Seamless integration with live chat
In a numerous, successful web self-service projects, any virtual conversation can be escalated to a Live Operator, via an assisting, seamlessly integrated live chat functionality.
Knowledge Base is utilized by users for self-service. It can address most common and less complex questions. However there are situations which require human interaction. Thus Inteliwise System also provides seamless switch between Virtual Agent (automated solution) and Live Chat operator (human being). The conversation can be switched between Knowledge Base and LiveChat operator many times, as well as transfered between multiple operators.
The Live Chat operator is also able to search through the Knowledge Base for faster responses. The system even enables the operator to manually switch the conversation back to the Virtual Agent (Knowledge Base) if the topic is well covered by the Automated Solution.
Integrations with External Knowledge Sources & Data
Seamless integration with existing systems enables delivering dynamic responses based on the part of day, time or specific conditions such as signed- in or authenticated status. The response can be also dynamically loaded from any external source such as customer Knowledge Base, XML file or any other API-enabled system.
End User – Presentation Module
The interface’s main component is the InteliWISE Presentation Module. It is a client side application works similarly to simple virtual machine. A set of instructions was defined which can be used for building custom applications’ logic. Each command implements some basic functionality and is executed in sequences by abstract form of processor.
This approach makes the new module the most flexible and complete solution for integrating our Video Avatars, NLP engine and other web applications with any kind of site. Moreover, it can be used for managing static web pages and adding new interactive functionalities to them.
From user’s perspective the system provides rich HTML5 communication module. A text input field for asking manual questions, list of most popular questions and also related topics. It handles misspellings and grammar errors as well. If it is unable to answer the question it provides a “don’t know” message and gives you the ability to rephrase your question or connect to the Live Chat operator for further assistance.
Content Authoring Tools
Comprehensive administration tools, in the form of self-service Panel, enable our partners and clients the full range of content authoring changes and system modifications, incl. the look and feel and the precise moment of interaction.
Advanced Content Authoring Tool
InteliWISE provides with a robust tool for updating and optimizing the knowledge base – InteliWISE Knowledge Engineer. It is a desktop app, it combines features that enable Editors to create, secure, and manage every Knowledge Bases designed by or for the client. Our tool enables to edit, add, or delete Q&As, scenarios, and scripts on number of different levels.
Web-based Administration Panel
The unique, InteliWISE Panel enables administration and customization of almost every element of the InteliWISE Virtual Agent Solution.
Any authorized project team member can run manage changes in:
- The Look & Feel of the solution
- The moment of engagement (i.e. when Virtual Assistant shows up to a user)
- The fi rst – welcome message
- The customized script
And many others.
Real-time analytics and performance dashboards
Our performance reporting tools provide critical customer information, including the full transcript on exactly what your customers are talking about – and how.
Our reporting tool will report on information like:
- What exactly did customers ask for?
- How they formulated questions in their words?
- For which questions they did not find answers on site?
All the reporting is presented in the web-based Control Panel. The access to the Panel and Dashboard is given to our customers, or InteliWISE team sends the reports in the form of e-mails. The performance monitoring can also be integrated with Contact Center Solution, or other systems, that would enable monitoring of important customer information from Virtual Agent without leaving company’s preferred analytic systems.
Upon the selection of a system, when a company’s team uses an analytic system to monitor sales or customer relationship information, there is now invaluable available for access. Once InteliWISE Virtual Agent is installed on the website site, the Reporting on its performance is immediately available through the company’s preferred analytic system.
Conversation history and genuine users’ language
A very special reporting that InteliWISE system provides is the review of any of the past conversations. Although it provides a lot of useful information, going through all of the conversations may be a very time-consuming task. Therefore it is recommended to use the Filtering option. The tool lets you adjust the Chat History section so that it shows you the information that is essential to your analysis. You can search by Questions, Answers or even Host/IP addresses.
Our Natural Language Understanding and other elements of the system are built on a scalable technology platform, enabling both on – premise and full clould – based deployment.
Mobile device solution
InteliWISE technology is also available on mobile devices (e.g. iPad) via native application. They may also include speech-enabled systems.
Whenever InteliWISE system is integrated with ASR systems (designed to recognize speech), it leads to a complete access to knowledge on mobile devices or in call center. Using our scenario driving we can build complete non- -linear IVR solutions.
InteliWISE solutions are patent pending in the United States # BM Dkt No. 67214529-001100, that protects systems and methods for generating and implementing a computer generated human – a machine web interface based on natural language processing and avatar virtual agent based character