3 Steps to successfully implement AI in your organisation

Today, it is necessary to implement AI in the organisation to have a competitive edge in the market. But, organisations face difficulty when it comes to tackle the new technology. In this post, we talk about three steps which can be followed to successfully implement AI in your organisation.

In the last 6 months, I have demonstrated conversations with a total of 63 companies of all sizes and formations. These businesses are primarily in the automotive, manufacturing, and energy industries.

My research reveals that around 25% of these companies start with a few AI initiatives, such as AI formations or AI use case identification workshops. In contrast, 47% of these companies have not started any AI initiatives yet. In some circumstances, they do not regard it as part of their strategy and vision. This means 72% of the companies I’ve spoken with have never developed an AI solution. This critical observation shows that 19% of them have 5000+ employees, and an additional 19% have 40000+ employees.

One can conclude a few things out of this conclusion – it is clear that the AI race is still at its origins. AI presents a massive amount of opportunities still to be exploited, which will improve both customer satisfaction and process efficiency. Additionally, I am concerned that those companies that are not considering AI are playing with fire in their overall strategy. AI is a source of competitive advantage, and like with many companies during the .com revolution, not paying attention to it could mean upheaval. Finally, there is a noticeable amount of AI illiteracy to be covered, and accordingly, there is the need for guidance to have a sensible approach to the science.

In this post, I am talking about the three main high-level steps an organization must follow to tackle artificial intelligence efficiently. These are – define an AI vision, identify AI use cases, and develop the critical needs to implement them.

Step 1: No AI vision is a synonym of failure: That’s how straightforward it is. Like a business without vision, approaching artificial intelligence without an AI vision is also condemned to collapse. Suppose an organization called “Sharp Charging” with a business model is to sell energy to Electric Vehicles (EV). Their vision is to “be the most reliable energy supplier for EV.” The CEO has decided to implement AI for an unrelated reason. The following proposals were initiated from this implementation:

  • The HR department developed a complex AI model to automatically sort the incoming applications.
  • A product owner wishes to develop a dynamic pricing to adapt to the different EV users and deliver a better customer experience.
  • A validation engineer is convinced that anomaly detection could reduce the validation costs of chargers by 15%.
  • The IT department thought it would be wise to invest in a new IT infrastructure before developing any AI use cases.
  • The development team believes they should first and foremost hire a data scientist and an AI engineer to add expertise.

While all of these points might be necessary at some point in time as the company’s AI journey advances, in 4 out of the 5 cases, the outcome would lead to failure. The fifth case would have been an educated guess.

The reason is that all actions, despite their necessity, are not taken under an agreed thoughtful AI Vision. An AI vision everybody knows and considers whenever an AI-related action is taken. Actions that, ideally, always sum up towards the company’s goals.

Step 2: Time to identify AI use cases: The previous example has defined the following AI vision: “Leverage AI to increase the customer experience of EV owners.” Considering this, would you say the HR proposal will deliver any direct value to the EV owners? What about the idea from the validation engineer? Of course not. The first thing the CEO should add in the company’s AI use case pipeline is the proposal from the product owner. Nothing says that the other two concepts are wrong; it’s simply that the organization’s engagement and the return on investment of their time and money in developing AI solutions will be increased once that use case is aligned with the AI vision.

Now let’s suppose a scenario such as that company, Sharp Charging, has identified 10 use cases in concordance with their AI vision. Our advice for those organizations starting their AI path is to choose the AI use case which seems easier to implement (“quick wins”) and develop it. Internally or externally, go for it. Even if the project fails, an essential amount of knowledge will be gathered, which will be helpful for future development. Remember that you’ll get ahead of 72% of the market by doing this.

Step 3: The AI use cases dictate how to proceed: I know it could be overwhelming to start with the actual development of the AI use cases. There are many components to consider: new team members, IT infrastructure, reorganization, ethics and regulation, having the correct data, etc. However, it is crucial to ask, what is the most important? Where to start? I’ll write an entire post in the future about this topic, but so far, what I can tell you is that the answer is in the use cases!

The identified AI use cases will tell organizations what they require in order to bring them to reality. The idea is to ask questions about them: What is the business goal of the use case? What type of data do we need? Do we have access to it? What kind of data do we have? What are the constraints? What are the assumptions? What infrastructure do I need to bring the use case to production? Answering these questions will define to which degree the components mentioned above need to be developed. This is why the IT department and development team proposals from the previous examples would have led to failure. The investments in resources were based on gut feeling instead of on actual needs. Within our technical services, we use the data mining methodology CRISP-DM to properly approach this step.

Although some companies may not have started their AI journey yet, it is not too late. In my conversations with our customers, I have a clear vision of what is occurring in the market. I’ve seen that the quantity of applications they can cover with some of the most common AI use cases is simply endless. Nonetheless, businesses are slowly realizing and establishing their AI strategies. It is not too late to get started today; however, it will be more challenging at some point in the distant future. My recommendation for businesses that haven’t started yet is straightforward – define an AI vision, identify AI use cases that align to it, select one, and begin developing!

Regardless of your stage with AI development, AI Shepherds is dedicated to helping you.

Albert Pujol
Mechanical Engineer with an MBA and a Master in Big Data and Data Science. Co-founder and CEO of AI Shepherds. Building stuff is what I love most and do best. My vision is a conscious world.

Leave a Comment

Your email address will not be published.

Implementing AI
July 20, 2022

Implementing AI: The 6 critical components

In the ongoing fourth industrial revolution, organisations are trying to implement artificial intelligence for getting a competitive edge in the market. Nonetheless, the lack of information and AI experience leads to common mistakes. In this post, we talk about how to avoid the most common pitfalls by means of the 6 critical components of AI.

View article
Do we really know what Big Data is
May 7, 2022

Do we really know what Big Data is?

In the middle of 2021, no one is surprised by the term “Big Data”. All professional or student of any discipline has heard about it and each one is capable of giving a description, or rather, their interpretation of the concept, its limits or its purpose. However, do we really know what Big Data is?

View article


Subscribe to our newsletter!

Once a month, our CEO Albert writes a newsletter with the most relevant milestones and strategic goals. Don’t miss it!

Scroll to Top