Functions description and explanation

Ruled Based Classification

Description

Classification of incoming requests according to clear and simple programmable rules. Based e.g. on labels in your current ticketing system.

Purpose

The function enables classifying requests according to clear, simple, pre-programmed rules. In this way, the system, after classifying the task, can take the appropriate actions it calls for.

Machine Learning

Description

A module based on the analysis of large amounts of data. Enables detecting patterns, similarities, common characteristics which lead to an increase in the effectiveness and efficiency of operations in Mr Wolf. The result’s effectiveness is calculated as a percentage.

Purpose

This function enables smart classification of requests based on their common properties, such as subject, content and labels, and provides estimated effectiveness of the calculations. Machine Learning can be used in the same scope as Rule-Based Classification, described above.

Self Learning

Description

One of the most important functions – the process of continuous improvement of the model created by Machine Learning. Created once, the model will accurately represent the cases on the basis of which it was built.

Purpose

With the Self Learning module, unrecognized cases are manually classified by an employee, then are regularly placed in the training set and a new, more effective model is generated automatically at determined intervals, taking into account the new cases from the set.

Predict

Description

A module for performing tests and calculating the effectiveness of a Machine Learning model.

Purpose

The module enables ongoing control of the operation and effectiveness of the Machine Learning module at any moment, so that it can be verified and tweaked. You can e.g. type in a particular message and see how the system will understand it and what action it will decide to take.

Multi Language detection and support

Description

If a process includes multi-language correspondence, you need a separate Machine Learning model for each of them. For this purpose, we created a special module which detects over 35 languages. 1-2 sentences of the text are enough to make it very effective.

Purpose

It enables detecting the language of the text on which Mr Wolf is working, and as a result, using ML functions in the appropriate language model. With this function and the number of handled languages, you don’t have to worry whether the system will work in other markets.

Sentiment Analysis

Description

A function enabling interpretation of the sender’s emotions based on content and expressions in their message.

Purpose

Using this function, it’s possible to detect positive/negative emotions in correspondence. It helps in analyzing customer service quality, can be used to change the service attendant, calculate statistics and the line’s working effectiveness.

Intelligent Information Extractor NLP

Description

In order to effectively respond to plain text requests, it is essential not only to categorize them correctly, but also to appropriately extract data from the text.

Purpose

The module enables extracting any information from a text. It can be done using simple patterns or advanced models trained with neural networks. With it, your employees can e.g. receive key information from an e-mail with any content possible, not losing any time reading long or complex messages.

Savings Report

Description

Monitoring benefits from automation after implementing it provides valuable information. You can do it using the reporting module, where a dashboard is used to present savings calculated on the basis of a statistic of automated tasks.

Purpose

It enables ongoing monitoring of savings resulting from automation using a simple dashboard, presenting various metrics and trends. The savings can be analysed per day or per task. The savings can be analysed per day or per task.

Human Guided Form

Description

In the process of automation, there are numerous extremely time-consuming tasks, which at least at the beginning of implementation require human intervention. It might be verification of classification accuracy or extraction of the correct data from a request. It may also be model training – manual classification and teaching the algorithm to the model. For this reason, we created a semi-automatic module, where the machine performs a maximum of automatic tasks, but its work is manually verified by an employee.

Purpose

The module enables seamless implementation of automation without fearing that a case will be handled incorrectly – a human controls the process, can correct the data and cancel incorrect tasks.

Advanced Reporting Module

Description

When you need very advanced statistics on the work of automated processes, you can use the advanced reporting module. It is possible to create any kind of metrics using the collected data and present them visually. The module can be viewed in the GUI or send regular summaries to e-mail addresses.

Purpose

It is possible to create customized reports on processes, present them on the screen or e-mail them as PDF files.

Drag&Drop New Processes Upload

Description

Creating processes in Mr Wolf is very easy. It consists only in preparing an appropriate definition in Python, and then dragging&dropping it into the admin panel.

Purpose

The module enables easy creation and changing of process definitions using drag&drop.

Adding own scripts

Description

It’s a function of Mr Wolf which enables you to create definitions of new processes by yourself. We implement the platform and possibly a process for you. You can add the next processes by yourself. We can help you with that by organising training.

Purpose

Adding new and changing existing process definitions by yourself.