Multiple-case design

A case study can be established around a single case or multiple cases (Eisenhardt 1989;

Yin 2003). Yin (2003) sees that when one has the choice, multiple-case designs may be preferred over single case designs if there are enough resources. According to Herriot and Firestone (1983) the evidence from multiple cases is often considered more convincing and the overall study may be seen stronger. Yin (2003) describes that in single-case design researcher ''puts all eggs in one basket'' and this can make the study vulnerable. The analytic benefits from having multiple cases may be more extensive. Additionally, Yin (2003) sees that the context of the two or more cases are likely to differ. If one can still arrive at common conclusions under these varied circumstances, the researcher has expanded the external generalizability of the findings compared to a single case alone. While the benefits of multiple-case design seem evident, it is important to choose each case carefully and intentionally, so that each case should serve a specific purpose (Runeson & Höst 2008;

also Benbasat et al. 1987; Yin 2003). The cases should serve in a similar way as multiple experiments do, with either similar results (cases give support to each other) or contrasting results (cases cover different theoretical conditions). In the latter case, different results might be expected but for predictable reasons (Yin 2003).

The multiple-case design is seen to complement this research for few different reasons.

Firstly, the technology is relatively new. It is expected that different sources have different viewpoints on the subject. Putting eggs in one basket as Yin express it, might lead to results that do not capture the phenomena. There is a handful of purchases in the public sector

concerning the topic and different purchasing methods have been used among the purchasing authorities. This alone implies that there are different angles to the subject. This confirms why multiple-case design supports the research. It is also expected that suppliers and purchasing authorities would have different perspectives to the subject. Observing the matter from one side will most likely lead to a one-sided result. Even though it is expected that there are differences among the cases, the interest of this research is also to look for the similarities and even try to point out a best practice if it is possible.

The cases that were chosen to this research are very shortly introduced here. The focus is on RPA as cases of intelligent automation were hard to find at this point. The central purchasing unit of the government is regarded as one of the cases. Three cases of public authorities were chosen as cases too, because they already had made RPA purchases and they have results concerning the use RPA.

4.2.1 Case 1 View from the suppliers and research field

There is already a good number of companies offering RPA solutions and services in Finland. This case includes the views of a couple of them. Finnish Tax Administration (Case 2) and HUS (Case 4) have the same supplier and this supplier was selected to be interviewed. Also, the preferred supplier of Hansel Ltd. (Case 5) dynamic purchasing system of IT consultation was interviewed as they offer RPA related services. Because these suppliers have been awarded contracts or are on the list of preferred suppliers, it is anticipated that they can be considered suppliers of highest quality and experience. The supplier of the Finnish government Shared Services Centre for Finance (Case 3) was left out from this study as it turned out that the initial supplier that was awarded the contract sold their RPA business to another company later. The new supplier was not interviewed as one of the goals was to receive supplier views from the public procurement and tendering procedures.

An interview with a Senior Scientist from VTT Technical Research Centre of Finland was interviewed as well as he had been part of RoboÄly project focusing on ''Public-sector development needs for robotic process automation and artificial intelligence – development needs of robotics and artificial intelligence from the perspective of information infrastructure''. As part of their research they had interviewed RPA suppliers and public authorities.

4.2.2 Case 2 The Finnish Tax Administration

The role of the Finnish Tax Administration collects more than 50 billion euros every year in tax revenue, which is used for maintaining and developing public services. They employ about 5,000 professionals and serve customers i.e. the tax payers. These numbers suggest operations should be effective and systems are reliable. The organization is doing continuous development work to improve the services. The Finnish tax administration was chosen as a case because they have strong knowledge about robotic automation. The organization has used dynamic purchasing process to tender RPA.

4.2.3 Case 3: The Finnish Government Shared Services Centre for Finance and HR

The Finnish Government Shared Services Centre for Finance and HR (hereinafter referred as ''Palkeet'') provide finance and HR services for central government agencies, departments and funds, as well as state-owned businesses and fully state-owned limited companies that provide services for the Finnish government. They were one of the first government authorities to tender robotic process automation and the organization sees that digitalization, process automation and the use of data in management will have an immense effect on the development of finance and HR administration in the near future. Palkeet has built own competence around RPA and they have developed an RPA deployment service to be offered to governmental agencies. Palkeet was chosen as a case because they have experience about RPA and conducted RPA procurement using a negation procedure in the tendering. Furthermore, Hansel Ltd. acted in the RPA procurement in a consultative role.

4.2.4 Case 4 HUS - The Hospital District of Helsinki and Uusimaa

The Hospital District of Helsinki and Uusimaa (hereinafter referred as ''HUS'') is a Joint Authority formed by 24 municipalities. The aim is to offer specialized medical care to all member municipalities and their 1,6 million inhabitants. HUS has more than 22,000 employees and is one of the largest employers in Finland. HUS Logistics acts as a joint purchasing unit of the municipal federation and provides for the logistics services. HUS Logistics is one of the largest public procurement units in Finland and competes with the procurement of medical equipment and supplies, furnishings and various care or support services in the hospital. HUS has also an IT management as a support service which has been involved in the procurement of RPA. HUS has set up a framework agreement of

robotic process automation service in purchasing collaboration with other contracting entities. Another reason why this made an excellent case for this research was that HUS Logistics exercises joint purchasing with the similar objective as Hansel Ltd. (Case 5) i.e.

saving resources from the members or inner customers.

4.2.5 Case 5 Hansel Ltd.

Finnish government uses centralized purchasing unit, Hansel Ltd., to reduce public expenditure by increasing productivity in government purchasing. The central purchasing agency negotiates framework agreements or sets up dynamic purchasing systems in selected product and service categories for the use of its customers. The aim is to reach lower prices through volume discounts, provide appropriate quality and reduce overlapping work among the government organizations. In addition, the agency offers tendering services to support government organizations in specific and individual purchases that cannot be acquired through framework agreements. Purchasing agency does not have a framework agreement or a dynamic purchasing system specifically for knowledge work automation technologies such as robotic process automation or intelligent automation. Importantly, Hansel Ltd. was chosen as a case because it could possibly generate synergy benefits.