Process Mining use is growing, as more and more companies recognize the benefits it provides. Studying and monitoring processes can be extremely valuable to increase efficiency by cutting costs, exploiting synergies, recognizing and improving operations which are particularly prone to failure or are particularly repetitive, and therefore there is potential for automation. Process Mining is particularly interesting for larger companies with many processes to keep under control. In this case, it is worthy for those companies to hire full time process mining experts who can operate all the necessary analysis, and give the results. Smaller companies do not generate enough data to justify a full-time employee working on that, so they can only rely on consultancy, which is clearly more expensive on the long run. The research problem this thesis seeks to answer is: can we develop a general Process Mining tool that would give non-expert users natural language answers to key process efficiency questions? This is a broad question, and relies on the definition of a "key process effi- ciency question". In this case we take into account as main goal for the user to find the best way to automate their process. This conjunction between process analysis and automation has been researched previously, concluding that automating a process starting from its worst performing activities is a good way to achieve the best results, and Process Mining techniques can provide the necessary information to automate activities in the correct order. Since this experiment is aimed at communicating this to unexperienced users, the produced tool should be accessible to people who never studied Process Mining, but have some process to keep track of. This study aims then to create a prototype tool which anybody with basis of statistic and graphs can use, to generate some useful results out of process data, leading to efficient application of Intelligent Automation. Ideally this should also stimulate Process Mining awareness and usage, and push for a general improvement in efficiency and automation. This document will describe the tool created, as well as some examples of results obtained, and an evaluation of the results based on a survey with potential users of the system.
L'uso del Process Mining sta crescendo progressivamente, mentre sempre più compagnie riconoscono il suo potenziale. Lo studio e il controllo di processi di business può infatti essere estremamente importante per migliorare l'efficienza degli stessi, tagliando costi, sfruttando sinergie e riconoscendo operazioni particolarmente ripetitive e proni al fallimento, dove l'automazione potrebbe portare benefici. Il Process Mining è particolarmente interessante per grandi compagnie con molti processi da tenere sotto controllo. In questo caso vale la pena, per queste compagnie, di assumere esperti a tempo pieno che possano analizzare i processi e produrre risultati. Compagnie più piccole non generano abbastanza dati da giustificare l'assunzione di esperti a tempo pieno, perciò si possono solo affidare a consulenze, che chiaramente risultano più costose, nel lungo periodo. Il problema su cui questa tesi è incentrata è il seguente: è possibile sviluppare uno strumento generico di Process Mining che possa dare a utenti senza esperienza risposte a problemi chiave di efficienza dei processi con linguaggio naturale? Per rispondere a questa domanda occorre definire cosa si intende per "problemi chiave di efficienza dei processi". Questi possono essere svariati, dipendentemente da cosa l'utente ha intenzione di raggiungere. In questo studio si considera obbiettivo dell'utente quello di automatizzare il processo in questione. La relazione tra Process Mining e Automazione è stata studiata in precedenza, concludendo che automatizzare un processo partendo sempre dall'attività meno efficiente al suo interno è una buona strategia, e che il Process Mining può provvedere le informazioni necessarie per operarla correttamente. Essendo questo esperimento dedicato anche alla comunicazione di questi risultati con utenti senza esperienza in ambito di Process Mining, lo strumento prodotto deve anche tenere essere accessibile a questi utenti. Questo studio ha come obbiettivo di creare uno strumento che chiunque con basi di statistica e lettura di grafici possa usare, per generare risultati, a partire da log di processi, che possano guidare l' efficiente applicazione di Intelligent Automation. Secondariamente, questo studio si propone, di mostrare le potenzialità del Process Mining, e stimolare l'uso di esso e dell'automazione. In questo documento il lettore troverà una descrizione dello strumento creato, alcuni esempi del sui funzionamento, i risultati ottenuti, e una valutazione basata su interviste e questionari con potenziali utenti.
Improving work automation with an accessible business process analyzer
Marabelli, Michele
2019/2020
Abstract
Process Mining use is growing, as more and more companies recognize the benefits it provides. Studying and monitoring processes can be extremely valuable to increase efficiency by cutting costs, exploiting synergies, recognizing and improving operations which are particularly prone to failure or are particularly repetitive, and therefore there is potential for automation. Process Mining is particularly interesting for larger companies with many processes to keep under control. In this case, it is worthy for those companies to hire full time process mining experts who can operate all the necessary analysis, and give the results. Smaller companies do not generate enough data to justify a full-time employee working on that, so they can only rely on consultancy, which is clearly more expensive on the long run. The research problem this thesis seeks to answer is: can we develop a general Process Mining tool that would give non-expert users natural language answers to key process efficiency questions? This is a broad question, and relies on the definition of a "key process effi- ciency question". In this case we take into account as main goal for the user to find the best way to automate their process. This conjunction between process analysis and automation has been researched previously, concluding that automating a process starting from its worst performing activities is a good way to achieve the best results, and Process Mining techniques can provide the necessary information to automate activities in the correct order. Since this experiment is aimed at communicating this to unexperienced users, the produced tool should be accessible to people who never studied Process Mining, but have some process to keep track of. This study aims then to create a prototype tool which anybody with basis of statistic and graphs can use, to generate some useful results out of process data, leading to efficient application of Intelligent Automation. Ideally this should also stimulate Process Mining awareness and usage, and push for a general improvement in efficiency and automation. This document will describe the tool created, as well as some examples of results obtained, and an evaluation of the results based on a survey with potential users of the system.File | Dimensione | Formato | |
---|---|---|---|
polimi_thesis.pdf
solo utenti autorizzati dal 01/04/2022
Descrizione: Tesi magistrale-Michele Marabelli-919995
Dimensione
1.19 MB
Formato
Adobe PDF
|
1.19 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/173715