SunPoint IT Solutions



Robotization of "torn" processes and ML

Perfect and awkward processes Ideal "process" for robotization: • Repetitive • Stable • Consistent • Digital assessment • Labour intensive "Inconvenient" process for robotization • Requires approval • Requires expert opinion • Long waiting times • Partial robotization Robotic Process Automation (RPA) - robotic process automation is a technology for using software robots for low-code business process automation. Basic RPA Features: • Working with the Office suite • Work with various mail protocols • Work with desktop applications and browsers • Working with remote desktop • Computer Vision ML package • Orchestration and scaling • Integration with various systems using standard protocols (web services, databases) • OCR ML capabilities • Development of your own packages • Reuse of algorithms Advanced RPA (ML): RPA and NLP A chat bot is a virtual interlocutor, a program designed to simulate human behavior. Natural Language Processing- general direction of artificial intelligence. Benefits of Robot Automation RPA, Chatbot, Voice Bot and OCR Robot (hands and brain) • Works faster than a person. • Works constantly in the background. • Does not get sick and does not go on vacation. • Performs exactly the job for which • Doesn`t make mistakes. Voice bot (hearing and voice) • Works constantly in the background • Can make calls based on • Bank telephony can be used. • Just one robot Chatbot (communication) • Runs in the background • Adapts to different platforms messages • Easy and quick access to the service Recognition and CV (vision) • Defines document types • Works with images as objects What is ML? Machine learning (ML) is a class of artificial intelligence methods, a characteristic feature of which is not a direct problem solution, but learning in the process of applying solutions to many similar problems. To build such methods, the tools of mathematical statistics, numerical methods, optimization methods, probability theory, graph theory, various techniques for working with data in digital form are used. This is the process of self-identification of signs and patterns by the system (a set of parameters "Weights" and mathematical functions) based on training data (examples) and various training algorithms. What tasks can you solve? • Classification - defining the class of the object • Clustering - grouping and sorting • Regression - give an estimate • Association - find patterns • Dimension reduction - "collect the best outfits from things" • Generative networks - creating "content" • Time series - forecasting • Autoencoders - object recovery • Segmentation - search for objects in images Stages of creating a service based on ML • Analysis of processes, data and IT park of the company, definition of the task • Ensuring the process of accumulating the necessary data and finalizing systems for obtaining missing • Accumulation of data for piloting • Analysis and preparation of data • ML architecture creation and machine model training • Publishing the solution to production Examples of complex processes • Robotization of IT department processes • Robotization of the scoring process for legal entities (59 sources) • Robotization of mass recruitment for retail (search map, search and selection of candidates, survey of candidates, scoring of information security, design, adaptation) • Robotization of the process of collecting and processing loan applications (including approvals and long waiting times) • Robotization of surveys • Providing multi-channel access for employees to the company`s services and knowledge base (mobile application, LC on the portal, messengers) Mass recruitment system TASKS • Automation of the process of search and selection of candidates • Automation of candidate safety checks • Automation of questioning of interest and appointment of interviews • Automation of registration of employees in the company aggregators of candidates. • Automation of the learning process, gamification and adaptation. • The business process is regulated and flexible in management. 9 of the 12 basic steps of the process are automated. days up to 40 minutes. • Employee participation is limited only to compilation of a search map, verification reduced by 3 times. results and conducting face-to-face interviews. • The process of searching for candidates in databases and aggregators of candidates has been automated by 100%. • Labor costs for the implementation of standard procedures have been reduced by 10 times. Providing employees with access to company services (Help Desk, Service Desk) • It is necessary to ensure that employees can use the company`s services through various channels (mobile application of the company, website, portal, messengers, etc.) • Provision of work with various services in a "single window" format. Robotization of the collection and processing of loan applications • It is necessary to ensure that customers can collect requests for obtaining a loan from various channels of receipt in a single portal. format. • Automate the process of primary documents. communicating with the client and obtaining all the necessary information. • Automate the processing of documents received in the form of photographs and scans.