Tarmo Koppel

Services

This page describes work done on order for companies and other organizations.

A student conducts research in a company as part of their thesis

A student doing a thesis must collect data in both bachelor’s and master’s studies, for example, send out a questionnaire to which n 100 respondents respond OR n conduct a 45-minute interview with 7 people. They are called data subjects. Therefore, the development of every AI solution should be related to its testing, e.g. by test clients, test employees, test users or other persons who can express an opinion on whether it works for them and is beneficial. Therefore, obtaining data is also one of the biggest challenges of a thesis. In addition to the AI ​​focus of the thesis, one should also figure out who the data subjects are, then the topic of the thesis can also be proposed to students.

General services

Below is a complete workflow (pipeline) for the implementation of an AI-based business process. The described workflow is adaptable to both companies and public sector institutions.

Realization of AI-based Business Process
|
+-- 1. Process Modeling
|
+-- 2. Process Analysis
| |
| +-- 2a. Technical Feasibility Study
| +-- 2b. Operational Feasibility Study
| +-- 2c. Economic Feasibility Study
| +-- 2d. Legal Feasibility Study
|
+-- 3. Process Design
| |
| +-- 3a. Process Improvement
| +-- 3b. Technology Development Work
| +-- 3c. Technology Configuration
|
+-- 4. Process Implementation
| |
| +-- 4a. Piloting
| +-- 4b. Staff Training
| +-- 4c. Change Management
|
+-- 5. Process Measurement
|
+-- 5a. Efficiency Study

Figure 1. Realization of AI-based Business Process, Workflow (Koppel 2024)

PIPELINE

1. Process Modeling
Process modeling involves mapping business workflows and processes to understand how they operate and relate. Modeling helps visualize processes and identify bottlenecks.

2. Process Analysis
Process analysis involves evaluating existing processes from the aspects of efficiency, costs, and quality. The analysis identifies bottlenecks and defines where AI can make work cheaper, faster, or more efficient in other ways. Process analysis clarifies the company’s actual digital needs. This may involve several types of feasibility studies.

  • 2a. Technical Feasibility Studies
    Technical feasibility studies evaluate the technical resources required and already available to the company. Technical feasibility indicates whether the company is technically ready to implement an AI system and integrate it with all necessary processes. It includes the capability of technical resources as well as the competence of the technical team.
  • 2b. Operational Feasibility Studies
    Operational feasibility studies evaluate the benefits the company gains by implementing AI into business processes. It shows whether and to what extent the intervention (installation of AI) meets the company’s operational requirements and expectations. It may also clarify what AI technology can currently achieve—what works and what doesn’t; learning from successful use cases in the given context.
  • 2c. Economic Feasibility Studies
    Economic feasibility studies address the financial benefits of the proposed AI system. This includes cost and revenue analysis, determining whether the costs are justifiable and whether the benefits outweigh any potential harm.
  • 2d. Legal Feasibility Studies
    Legal feasibility studies examine potential conflicts of the proposed AI system with legal requirements, such as data protection and other laws.

3. Process Design
Process design involves creating new or improved processes that meet the organization’s goals and needs. It includes simplifying and automating processes, including the implementation of AI solutions.

  • 3a. Process Improvement
    This includes creating new procedure rules, guidelines, and documentation.
  • 3b. Technology Development Work
    Developing new solutions if necessary.
  • 3c. Technology Configuration
    Involves tailoring existing technology solutions and installing/configuring market solutions within the company.

4. Process Implementation
Process implementation involves the practical application of designed and developed processes. It typically includes the following:

  • 4a. Piloting
    Piloting is the introduction of new technologies. The intervention study aims to determine whether the AI solution works under real-world conditions. Companies can use intervention studies to test new working methods before full implementation.
  • 4b. Staff Training
    Training employees on new systems or methods.
  • 4c. Change Management
    Managing organizational changes during the implementation of new processes or technologies.

5. Process Measurement
Process measurement involves assessing the performance of digitalized/automated processes. In certain cases, this enables real-time data collection and analysis to evaluate the effectiveness of improved processes and draw conclusions for further improvements.

  • 5a. Efficiency Study
    Efficiency studies determine whether the new AI solution positively impacts business performance. The efficiency study is used to assess how effective the existing (non-AI) business process is and to see if there is room for improvement with AI. It also identifies any shortcomings in the implementation and draws conclusions for further improvements. The efficiency study may also address customers to assess whether their satisfaction with certain products or services is sufficiently high.