M42 Intelligence 

The AI for Service Desk that keeps your data local

Introduction

M42 Intelligence is a product family that brings Artificial Intelligence (AI) capabilities to M42 Service Management solutions to boost agent efficiency and reduce manual effort in handling tickets and exchanging information.

M42 Intelligence uses secure generative AI and Natural Language Processing (NLP) technologies to empower Service Desk agents and help end-users solve issues quickly and at a lower service cost. With a choice of Matrix42-only or supported 3rd party AI models and private or public cloud data residency, M42 AI gives compliance-conscious European organizations the best of both worlds: productivity-boosting generative AI capabilities with complete control over the privacy and security of their data. 

The service desk has been left behind in the race to digitalize and automate, costing time and money. A massive 83% of end-user tickets are still handled manually by agents, even for low-level and repetitive issues. In addition to hampering productivity, this keeps service costs high.

At the same time, users are increasingly happy to solve issues with the help of an intelligent chatbot. It's now the #5 most preferred method out of 12, up from #10 in 2020, reflecting growing levels of comfort with digital self-service and familiarity with AI virtual assistants.

There is a strong case for using generative AI to help users solve issues and agents to be more productive. However, some organizations are concerned about entrusting sensitive data to a “black-box” cloud-based AI tool that might be stored overseas and used to train large language models (LLMs).

M42 Intelligence product family includes AI for End-users and AI for Agents. 

• AI for End-users empowers users to solve simple issues on their own, quickly, 24x7
• AI for Agents helps agents to manage increasing complexity and workloads efficiently.

M42 Intelligence reduces ticket volumes and cost per ticket, helps to get faster resolution times, and improves the overall user experience, both for end-users and agents.

AI for End-users

M42 Intelligence allows end-users to solve different issues independently and easily find what they are looking for 24x7.

We describe AI Chatbot in this solution description. For information related to rule-based chatbots and Live Chat, please check Efecte Chat for Service Management documentation.

AI Chatbot and AI Knowledge (GenAI, NLP, live chat, rule-based scenario) are secured with the latest TLS 1.3 protocol, which provides stronger encryption and faster connection establishment. This ensures that conversations remain confidential and all transmitted data is properly protected.

AI Chatbot

AI Chatbot is capable of understanding user messages and answering with predefined, curated responses. AI Chatbot uses Natural Language Processing, guaranteeing that every user is able to get support fast. Model can interpret user questions and provide useful answers, guide the user to additional resources, or help escalate the issue to live support agents. (NOTE: NLP models can communicate only in one of the specified languages: EN, GE, FI, SE, ES, PL)

AI Knowledge Discovery

AI Knowledge enables fast retrieval and generation of responses to user queries using Retrieval-Augmented Generation (RAG) technology. The tool integrates with various platforms, such as SharePoint, Confluence, and M42 knowledge bases, consolidating information from multiple sources. Advanced generative models enhance functionality within a conversational interface, simplifying access to knowledge, automating simple queries, and improving support team workflows. AI Knowledge operates 24/7, is scalable and secure, and provides responses exclusively based on internal organizational data.

Difference Between AI Chatbot and AI Knowledge

AI Chatbot

• Functionality: AI Chatbot serves as a conversational tool powered by Natural Language Processing (NLP). Its primary function is to respond to user messages with predefined, curated answers.
• Scope: The chatbot focuses on interpreting simple queries and delivering fast responses. It can also guide users to additional resources or escalate issues to live support agents.
• Use Case: Ideal for handling repetitive queries and supporting basic end-user support processes.

AI Knowledge

• Functionality: AI Knowledge is an advanced tool for retrieving and generating responses to user queries using Retrieval-Augmented Generation (RAG) technology. It leverages generative models to provide more advanced and comprehensive answers.
• Scope: Instead of relying on predefined answers, AI Knowledge dynamically searches through various knowledge repositories (e.g., SharePoint, Confluence, Efecte knowledge bases), consolidating data from multiple sources.
• Use Case: This tool is designed to handle more complex queries that require access to large volumes of information from various organizational systems. It acts as an intelligent search and response generation system.

Key Difference

AI Chatbot is based on predefined, selected answers, which makes it effective in handling simple and repetitive queries (note: NLP models can only support one language: EN, FI, GE, ES). On the other hand, AI Knowledge is a more advanced solution that is able to dynamically retrieve and generate answers based on extensive organizational knowledge bases. GenAI models on which it is based support most languages ​​(local GenAI models support: EN, FI, GE, SE, PL). This makes AI Knowledge ideal for managing more complex queries that require broader context and access to multiple sources of information.

AI for Agents

M42 Intelligence for Agents helps support agents to manage increasing complexity and workloads efficiently. It includes the following capabilities:

Email Assist

• Email message generation
• Email message auto-correct
• Email message completion

Configurable AI (AI Actions)

• Content summarization
• Configurable, dynamic content generation based on context
• Translations

AI Ticket (ViCo)

• Ticket classification
• Find similar content
• Content reuse by copying from similar items

Email Assist

Email Assist allows agents to communicate with end users efficiently, with less manual effort. Email Assist can be easily accessed in the email editor, by clicking the button in the bottom.

Email Assist currently has three key features:

• Generate: This allows agents to create a new message from scratch. Agents can get ready-made, generated responses by providing contextual data such as ticket subject or description as selections before they request the message to be generated. Additionally, the agent can define the language in which they would like to receive the answer from a list of supported languages.
• Correct: Agents can also use Email Assist for spell checking and correct any typing errors. First, they must select text in the email editor and then request the correction in the Email Assist window.
• Complete: Agents can finalize any incomplete message with the Complete feature. Like with the Correct feature, the agent needs to select the text they wish to complete.

Configurable AI (AI Actions)

Configurable AI enables support agents and service managers to extract key information and transform it to useful content with a single click. With configurable actions, AI Actions can be used in various use cases: from ticket summarization in a handover situation, to translating and generating content utilizing the data and hand, such as Knowledge base articles. AI Actions includes the possibility to customize prompts, so that it can be adjusted to unlimited amount of use cases.

AI Actions is only available only in the new UI of Efecte Service Management tool, and has to be set up based on your configuration to be made available.

AI Ticket (ViCo)

AI Ticket (formerly Virtual Coach) helps service desk agents find relevant information and work more efficiently utilizing existing data. AI Ticket is only available in the Classic UI of ESM.

When enabled, AI Ticket can be easily accessed in the top-right corner while working on a data card.

Currently, AI Ticket can help agents with the following features:

• Similar content provides ESM users with a list of items that AI considers to be relevant for the open data card. AI Ticket uses Natural Language Processing to find items in the platform that look similar to the open data card. This allows agents to find useful information, such as previous root causes or resolutions easily.

• Suggestions enable ESM users to classify data cards with, for example, category. Model builds a statistical model from the existing data in order to suggest the most likely values for fast classification. By clicking a suggested value, you can set the value for the configured classification attribute. This allows you to utilize suggestions with e.g. ticket categories, support groups, related services, as long as there is enough historical data to derive the suggestions from.

• Use Copy to fill data cards faster by copying multiple values from a data card found in Similar content. This allows agents to reduce the manual effort needed to fill in details for repeating issues.

For full guide on configuring AI Ticket, please see the configuration guide.

Matrix42 GenAI

Matrix 42 GenAI enables you to easily get started with AI-powered customer service, without having to have the technical capabilities to manage generic third-party services. Instead, you can leverage our purpose-built large language models designed to help with ITSM use cases while supporting multiple languages. 

Matrix42 GenAI currently has the following capabilities:

• Support for Email Assist
• Support for Configurable AI (AI Actions)
• Support for AI Knowledge Discovery

Language support:

• English
• Finnish
• Polish
• Swedish
• German