Adopting new technology and innovations is important for companies to succeed in today’s competitive environment. One technology that has received enormous attention in recent years is Artificial Intelligence (AI). Although more and more companies are already using AI and creating value with it, many are still struggling with its adoption. Where can it be used? What capabilities are needed? How will the AI solutions address real business needs? How can we realize the value from AI implementations? These were also the main questions in my Master’s Thesis, which analyzed AI adoption in six Finnish companies at the forefront of the AI and analytics adoption.
The main findings from the thesis suggest that companies should also put effort into post-implementation activities to realize the benefits from AI. The value and benefits aren’t realized unless the organization’s roles and processes are changed to support the interaction between humans and machines, the end users make data-driven decision and they actually use the implemented tools. On the other hand, many things already earlier in the process also affect getting the organization to change the behavior. Let’s take an example. Concentrating on the end users and their tools during the implementation will make the subsequent deployment easier. Therefore, it is important for organizations to take into account post-implementation activities but also at an earlier stage concentrate on the mechanisms that make the change easier.
The level and rate of AI adoption differ between industries and companies. Therefore, it is difficult to give unequivocal instructions to companies on how to approach and plan AI adoption activities. However, the findings from my thesis provide a foundation for some general guidelines.
1 – Ensure management’s priority and support
Management’s prioritization and support of AI adoption activities is a prerequisite for successful AI adoption. It came out in the thesis that unless management understands and trusts AI’s potential and benefits, the probability of success is minimal.
2 – Secure access to required capabilities for AI adoption
To enable AI adoption, a company needs understanding of AI, technological capabilities, strategic and business related capabilities, and digital capabilities. These can be developed internally or acquired outside of the company. However, it is advisable that an understanding of AI and strategic and business related capabilities should reside inside the organization.
3 – Analyze the digital maturity of the company
It is important to analyze the company’s current maturity regarding digitalization. Higher digital maturity makes the AI adoption easier in two ways. First, the digital processes enable collection of larger amounts of high-quality data. Second, the earlier digitalization activities may have resulted in a positive mindset towards implementing new digital solutions.
4 – Address real business problems with implementation activities
Companies should not adopt AI because of AI itself. More important is to see AI tools as a way to solve business problems. There are three ways companies can ensure that implementations address real business problems. (1) Implementation projects agile and continuous execution, (2) involving end users and focusing on their tools as well as (3) enabling smooth collaboration between technology and business people.
5 – Preparing the organization for deployment
Change management practices can be used to prepare employees to make changes in their behavior and deploy the implemented AI tools. Continuous and concrete communication about the reasons for change, the impact on different roles, as well as achieved successes, creates a positive attitude toward AI. This positive attitude speeds up the change itself.
6 – Focusing on behavioral changes in the organization
The benefits of AI adoption aren’t realized until the organization has changed the behavior and adopted new ways of working. On a higher-level this requires managers to plan changes in roles and processes, ensure that data-driven decisions are made, and prepare end users to really use the implemented tools. However, this is not only the responsibility of managers but it is advisable to have everyone onboard from the very beginning.
In conclusion, successful AI adoption requires more than fancy technology platforms and data scientists. If you recognize that perhaps your company has not concentrated on post-implementation activities enough nor equipped end users with the right tools and ensured that they know how to use them, now it is time to act.