The Role of AI in Business Processes

(English translation by ChatGPT)

Abstract

This article explores the role of artificial intelligence (AI) in business processes, with a particular focus on its potential in small and medium-sized enterprises (SMEs). The article emphasizes the benefits of using AI, such as improved efficiency, customer satisfaction, and innovation. It discusses the importance of awareness and investment in AI knowledge and expertise for SMEs. The Business Process Framework (eTOM) of TMForum is used as a framework to analyze business processes. The article presents scores and conclusions based on AI categories, including machine learning, natural language processing, and computer vision. Examples of business processes, such as product deviation management, market and sales accounting management, and security and privacy management, are discussed to illustrate the potential benefits of AI. The article highlights that careful planning, implementation, and consideration of ethical, legal, and security aspects are necessary when using AI in business processes.

Babelfish.nl offers AI services to small and medium-sized enterprises.

• Pum Walters (PhD) was TOGAF-certified information architect, associate professor in software development, cybersecurity expert, telecom architect, and consultant.
• Jacintha Walters (MSc) is an AI and cybersecurity expert. She has worked on multiple AI projects.

Introduction

Business processes define how tasks are performed and how information is processed. They form the backbone of many organizations. Optimizing business processes is crucial to achieve efficiency, effectiveness, and competitive advantage.

In today's dynamic world, artificial intelligence (AI) is an emerging technology with tremendous potential to transform business operations. AI enables systems and machines to learn, reason, and make decisions based on complex data. This opens the door to automated and intelligent solutions that can enhance routine tasks, accelerate decision-making, and generate valuable insights.

Large companies have already invested significantly in the potential of AI. They have AI experts on board and have developed strategies to effectively apply AI in various aspects of their operations. They have taken the lead in identifying possibilities and leveraging the benefits.

However, for small and medium-sized enterprises (SMEs), the story is different. Many of these companies may not have seriously considered how AI can be applied to their business processes, and for some, AI is still an abstract concept, and its potential benefits have not been mapped out. Other companies have taken initial steps but are still at the beginning of their journey in exploring and implementing AI solutions.

It is crucial for SMEs to become aware of the possibilities and impact of AI on their business processes. By harnessing AI, they can remain competitive in an increasingly digitized world and promote efficiency, customer satisfaction, and innovation. It is advisable for these companies to invest in acquiring knowledge and expertise in AI and to start identifying areas where AI can add value.

An important framework that can aid in improving business processes is the Open Digital Framework (ODF) of TMForum, which includes several models. TMForum is a collaborative organization in the telecom domain. One part of the ODF is the Business Process Framework, also known as eTOM. This framework provides a detailed, multi-layered map of essential business processes, originally designed for the telecom industry but applicable to other industries as well.

In this article, we focus on exploring which business processes, in general, can benefit from the use of artificial intelligence. Though the answer is brief – many, if not most processes – we delve deeper into the potential benefits of AI in different levels of business processes. We have used Google Bard to estimate the potential of AI based on machine learning, natural language processing, and computer vision. We have assigned scores to indicate the extent to which AI can play a role in each process on a scale from 1 to 5. However, it is worth noting that the value of these estimates may be limited (for instance, during the research, Bard was observed to assign significantly different scores on different days).

While the estimates and conclusions are tentative, they provide valuable insights into the potential applications of AI in business processes and can serve as a starting point for further exploration. The article also presents aggregated scores for different levels of business processes and identifies some preliminary conclusions regarding the areas where AI may have the most impactful potential.

It is essential to note that while AI shows promise, its use requires careful planning and implementation. The complexity of business processes and the diverse needs of various industries necessitate tailoring AI solutions and considering ethical, legal, and security aspects. Therefore, a solid foundation of business and information architecture is vital to successfully integrate and harness AI.

While this article is not a definitive guide, it hopes to provide readers with a foundational understanding of the possibilities and challenges of using AI in business processes. Further research and in-depth analysis are necessary to explore the specific applications and benefits of AI in different industries. The evolution of technology and the ongoing development of AI will undoubtedly create new opportunities and further enhance its potential in business processes.

The Business Process Framework (eTOM)

The Business Process Framework (eTOM) is an essential framework specifically developed for telecom service providers but has broad applicability in other service industries as well. The framework provides a structured and comprehensive "map" of business processes in the telecom sector.

eTOM is built on different levels, where each higher level addresses more detailed categories of the underlying levels. The layered nature of eTOM allows for viewing and analyzing business processes at different levels of abstraction, providing in-depth insight. There are over 100 level-2 processes and more than 600 level-3 processes. The processes at level 4 and higher are more specific to telecom and less relevant to this article, which focuses on an overview. Therefore, we limit ourselves to levels 2 and 3.

The framework offers a wide range of processes, such as customer management, sales, supply chain management, and interaction management. While eTOM was originally developed for the telecom sector, it is highly applicable to other service industries as well. For example, the level 2 process 'supply chain management' is relevant in most sectors. Moreover, the model, despite its obvious limitations, is also useful in other industries.

Due to the broad applicability of eTOM, it has become a valuable tool for business architects. It provides a common language and reference point for understanding and optimizing business processes in various industries. It enables a structured approach to examining the different functions, processes, and steps involved in delivering services to customers.

Analyzing Key Business Processes

Identifying business processes that can benefit from artificial intelligence (AI) is of great importance in today's business environment. In this section, we will delve into the different categories of AI and the methodology used to assess the role of AI in each business process.

Three categories of AI were considered: machine learning, natural language processing, and computer vision. Machine learning uses algorithms to identify patterns and insights in large amounts of data; natural language processing focuses on understanding and analyzing human language, while computer vision focuses on interpreting and understanding visual data.

To assess the role of AI in each business process, a methodology was followed. A five-point scale was used to estimate the extent to which AI can play a role in each (eTOM) level 3 process. Scores range from 1, indicating that AI is unlikely to play a significant role in that process, to 5, indicating that the process can be fully performed by AI.

To generate these estimates, Google Bard, a tool that provides AI-based analyses, was used. While it is essential to note that the estimates from Google Bard are not entirely reliable, they have served as a basis for analyzing the role of AI in different business processes.

In the following sections of the article, we will present the aggregated scores for the different levels of business processes and draw some preliminary conclusions based on this data. The goal is to provide insights into the potential of AI in various aspects of business operations and to encourage further discussion and exploration.

Exploring the Role of AI

We discuss the role of AI, taking into account three categories: machine learning (ML), natural language processing (NLP), and computer vision (CV).

Level 2 processes provide a broad overview of the various functions within an organization. These processes include essential activities such as marketing and sales, product management, customer management, service management, resource management, collaboration with business partners, and the overall business operations of the organization. The diagram shows the aggregated scores for these processes.

A more detailed exploration takes place at Level 3 processes, which offer in-depth insights into the specific activities within each function. Here, we can further examine the potential role of AI in each process. Example Level 3 processes could include 'Predicting customer needs based on historical data', 'Automatically classifying customer feedback for sentiment analysis', and 'Automating invoice processing using optical character recognition (OCR)'. These examples illustrate how AI can contribute to improving operational efficiency, gaining customer insights, and streamlining business processes.

To better understand the involvement of AI in Level 3 processes, we have assigned scores to each process on the five-point scale. These scores are based on Google Bard's estimates. For example, a score of 3 would indicate that AI can play a significant role in that specific process but may not fully automate it.

By analyzing the aggregated scores for Level 3 processes, we can gain insights into the broader implications for Level 2 processes. Examples of Level 2 processes that can benefit from AI include 'Automated lead generation and qualification' in the domain of marketing and sales, 'Optimized product launch and lifecycle management' in the domain of product management, and 'Automated customer support and issue resolution' in the domain of customer management.

It is important to note that AI's involvement in each process depends on various factors, such as data availability, task complexity, and specific business contexts. Not all processes will be equally suitable for AI applications, and careful consideration is necessary to weigh the potential benefits against implementation challenges.

By analyzing the scores of Level 3 processes and translating them into Level 2 processes, we can get a better understanding of the broader opportunities for AI in each functional area. This allows organizations to focus their efforts on identifying and implementing AI solutions that have the potential to improve efficiency, customer satisfaction, and competitive positioning.

In the next sections, we will delve deeper into specific examples of Level 3 processes and the reasoning behind the assigned scores. We will also provide concrete examples of Level 2 processes that can benefit from AI. Through this research, we hope to provide insight into the potential benefits of AI in business processes and inspire organizations that are still in the early stages of their AI journey or want to explore further how to integrate AI into their existing business architecture. The diagram shows that AI can be effectively applied in most processes, with some processes being partially or almost entirely performed by AI.

However, it is essential to note that while AI offers many possibilities, it is not a universal solution for all business processes. It requires in-depth analysis, evaluation, and alignment with an organization's specific goals and needs. With a strategic approach, businesses can leverage the benefits of AI and gain a competitive advantage in the modern business landscape.

The raw data underlying this article can be found here

Examples

Product Deviation Management

The application of AI, particularly machine learning, can offer several benefits in product deviation management processes. ML algorithms can identify patterns, detect and predict deviations early, handle the triage and classification of anomalies automatically, and continuously learn and improve. This enables organizations to proactively identify potential product issues, improve product quality and reliability, and provide customers with more certainty.

In this process, AI, especially ML, can be effectively utilized as follows:

1. Pattern Recognition:
   ML algorithms excel at detecting patterns and deviations in large volumes of data. In the case of product deviation management, ML can analyze historical data related to product behavior, performance, and user feedback to establish normal behavior. By comparing real-time incoming data with these patterns, ML algorithms can identify deviations or anomalies that may indicate product issues.
2. Early Detection and Prediction:
   ML models can be trained to predict and detect anomalies before they develop into serious problems. By continuously monitoring various data sources, such as product usage data, sensor values, or customer feedback, ML algorithms learn the patterns associated with normal product behavior and identify anomalies that might indicate a deviation. Early detection enables proactive actions such as initiating a detailed assessment or implementing measures before anomalies escalate into critical issues.
3. Automated Triage and Classification:
   ML can automate the triage and classification of product deviations based on their characteristics and impact. By analyzing historical data and learning from past cases, ML models can categorize anomalies, prioritize them based on severity or impact, and assign appropriate actions or workflows for resolution. This automation reduces the manual effort required for analyzing and managing a large number of anomalies, leading to better response times and more efficient resource allocation.
4. Continuous Learning and Improvement:
   ML models can continuously learn from new data and feedback, enhancing their ability to detect anomalies over time. As product deviation management processes encounter new types of anomalies or evolve with changing product characteristics, ML algorithms can adapt and refine their detection mechanisms. This iterative learning process makes the system more accurate and effective in identifying and managing anomalies.

By harnessing ML in predicting, detecting, mitigating, and managing product deviations, organizations can enhance their ability to proactively identify and address product issues.

It is important to emphasize that further research and exploration are needed to realize the full potential of AI in product deviation management. Technology continues to evolve, and there are opportunities for further optimization and application of AI algorithms in this context. By applying AI intelligently, organizations can gain a competitive advantage and continue to innovate in a rapidly changing market.

Market and Sales Accounting Management

The accounting process determines how an organization's financial condition is monitored based on regularly compiled financial statements. The marketing process is responsible for managing and developing an organization's sales. The accounting process is linked to the marketing processes to monitor market trends and manage the effectiveness of sales actions initiated by the marketing processes.

NLP offers the ability to analyze textual data, interpret financial statements, perform comparative analysis, and support decision-making. It can generate valuable insights about market trends, the effectiveness of sales actions, and potential deviations between marketing campaigns and financial outcomes. Moreover, NLP can aid in identifying risks, highlighting key information, and suggesting improvements in sales promotions and financial management.

1. Text Analysis:
   NLP techniques can be used to analyze textual data about market trends, sales actions, financial statements, and other relevant documents. By applying techniques such as text classification, sentiment analysis, and information extraction, NLP algorithms can extract valuable insights from unstructured text data. This can aid in monitoring market trends, assessing the effectiveness of sales actions, and identifying any deviations between marketing campaigns and financial results.
2. Financial Statement Analysis:
   NLP can assist in analyzing financial statements by extracting and interpreting essential information such as revenue, costs, profit, and expenses. NLP algorithms can automatically process and extract relevant data from financial reports, enabling swift analyses of financial performance indicators. By applying NLP techniques, the accounting department can gain valuable insights into the organization's financial situation and identify any deviations or inefficiencies in sales actions or marketing campaigns.
3. Comparative Analysis:
   NLP enables comparative analysis by comparing textual data from marketing reports, sales actions, and financial statements. By analyzing language used in marketing campaigns and correlating it with the corresponding financial outcomes, NLP algorithms can identify potential deviations between the expected impact of marketing efforts and the actual financial results. For instance, NLP can identify situations where a marketing campaign generated high sales but had a disproportionate impact on the organization's costs.
4. Decision Support:
   NLP can support decision-making by automatically identifying key insights and emphasizing essential information in textual data. By analyzing the language used in marketing and financial documents, NLP algorithms can flag potential risks, highlight areas of concern, or suggest improvements in sales actions or financial management. This can assist decision-makers in the market and sales accounting process in making informed decisions based on a thorough understanding of market trends, financial implications, and the efficiency of sales actions.

By leveraging NLP in the market and sales accounting process, organizations can gain deeper insights into the relationship between marketing efforts, financial results, and the overall efficiency of actions. This can lead to improved financial management, enhanced decision-making, and more effective alignment between marketing strategies and financial objectives. It serves as a valuable tool to improve the performance of market and sales accounting and lay a solid foundation for successful and profitable business operations.

Security and Privacy Management

The business process of security and privacy management identifies the organization's assets, including information, and involves stakeholders through activities that evaluate, develop, document, and implement policies, procedures, and practices to protect the organization's assets and meet the privacy requirements of customers, business partners, and all other stakeholders in the context of laws and regulations.

By applying computer vision (CV) in the Security & Privacy Management process, organizations can enhance security, ensure data protection, detect unauthorized activities, and enable rapid incident response. This contributes to better protection of assets, data privacy, and overall safety within the organization.

1. Access Control:
   CV techniques can be used to analyze visual data and detect unauthorized access. For instance, CV algorithms can utilize security cameras to recognize and verify faces, allowing only authorized personnel access to secure areas. This enhances security and helps protect assets and sensitive data.
2. Data Protection:
   CV can also be employed to enhance data protection. By using CV algorithms, images and videos can be analyzed to identify and anonymize personally identifiable information (PII). This enables organizations to comply with privacy regulations and safeguard data confidentiality.
3. Detection of Unauthorized Activities:
   CV can assist in identifying unauthorized activities or security breaches. By utilizing security cameras and CV algorithms, suspicious behavioral patterns, such as unauthorized access, vandalism, or theft, can be detected. This enables organizations to respond quickly and take appropriate measures to maintain safety.
4. Incident Response:
   CV can also be applied in the incident response process. By monitoring visual data, such as surveillance footage, CV algorithms can detect suspicious activities or incidents and automatically notify relevant parties. This accelerates response time and supports effective incident response and forensic investigation.

Conclusions:

In this article, we have analyzed which business processes can benefit from the use of artificial intelligence (AI). By examining and categorizing the Business Process Framework (eTOM) based on machine learning, natural language processing, and computer vision, we have gained insights into the potential role of AI at various levels of business processes.

Our analysis of Level 2 processes reveals that there are several domains where AI can be applied to optimize processes. Particularly in the areas of marketing & sales and product management, there are numerous opportunities for machine learning. Natural language processing is slightly less prominently applicable, while computer vision is mainly relevant in service and resource management.

In the more detailed Level 3 processes, we have investigated and analyzed specific examples. We assigned scores to the extent to which AI can play a role in each process. We observed that machine learning and natural language processing often score similarly, either high or low. On the other hand, computer vision has its specific areas of application.

The findings demonstrate that AI has broad potential to enhance and optimize various business processes. However, implementing AI solutions requires careful considerations and evaluations, taking into account data availability, task complexity, and business context.

This study provides an initial insight into the possibilities of AI in business processes. It emphasizes the need for continuous exploration and research in this area, encouraging organizations to explore and harness the potential benefits of AI to gain a competitive advantage in the rapidly evolving business landscape.

While this research lays a valuable foundation, it is essential to note that it is not a definitive guide. Implementing AI solutions requires customization and ongoing adaptation to the specific needs and goals of each organization. Therefore, further research is crucial to continually monitor the evolution of AI and its impact on business processes. By continuously innovating and exploring, organizations can fully leverage the opportunities AI offers to strengthen their competitive position and create value for their stakeholders.