Artificial Intelligence | Lesson 4.2

Business Driven AI Innovation: The AI vision

We plan to explain Business-driven AI innovation with FutureHouse, the fictional real-estate platform enabling users to buy and sell homes online, which we discussed in section 1. We came up with fancy AI features and discussed how these features could help FutureHouse to attract more users. But how did FutureHouse come up with the idea of those features? How did they prioritize their investment? Unless you work for a tech giant, your organization will also face these questions.

Therefore, we decided to travel back in time to the beginning of FutureHouse’s AI journey and analyze the conversations that probably happened in its meeting rooms. These conversations will give you an idea of the type of conversations you may experience while developing an AI module. The conversations in this section are taken from (Valigi & Mauro, 2020). Let us move on to the kickoff meeting:

CEO: AI is all over the news these days; it seems like it’s going to play a big role in any business. We definitely need to start looking at it before our competition or some tech startup does it first.

CTO: You’re right; it’s moving so fast, and we need to get ready. I suggest we start investing in infrastructure so we can process all our data. Then we can train our engineers or hire new data scientists and start building AI algorithms.

Marketing manager: Honestly, guys, I think it’s a fad. Now it looks cool to say “We use AI,” but it will eventually fade out.

AI leader: I agree that some people are jumping on the bandwagon too quickly, but I believe it would be shortsighted to ignore AI altogether. Of course, it’s not a silver bullet; we need a strategy to make sure we can get value out of the things AI is great at.

CEO: Mmm, but I’m not sure I’m fully aware of what things AI is great at. There’s so much confusion in the media that I struggle to understand what’s real and what’s hype.

AI leader: This is the first thing we need to fix, in my opinion. How about organizing training for the executive team so we’re all more aware of what AI can really do, and then we plan the next steps?

CTO: I think that’s a good idea; it would also help the technical team and the business guys learn to work together.

Marketing manager: Okay, let’s do this. Maybe it’ll help me change my mind about this technology. But I think it’s crucial that we start with a strategy. Without some form of AI vision, we’ll end up losing time and money.

AI leader: Glad you’re all excited about doing AI training, and I definitely agree with you that we do need to have a vision. On the other hand, I think we need more experience to build an AI vision. Training alone won’t be enough: we need to start getting our hands dirty before spending time on long-term planning. I suggest that after our training, we go looking for some AI projects to start experimenting with. They’ll guide us in defining our AI vision.

CEO: I like where this is going. Do we all agree on getting some AI training and starting to experiment with this technology? I want to have a pilot running four months from now; then we can stop, look at what we learned, and start talking about our overall strategy and vision. How does that sound?

There is a lot to uncover here. This conversation has many elements of an organization thinking of using AI in their business for the first time. Indeed, parts of this script will often play out in actual meetings. First, let us unpack its more important parts. We want you to focus on the people. This is a conversation, not a monologue or forced strategy by the managers. You need a team to make any meaningful changes. You cannot do this alone! Here we briefly discuss the types of people we had in this meeting.

• The initiator: Someone who starts the discussion. It can be someone who has heard about AI on the news, has the experience and sees the opportunity, or wants to initiate a new project to advance their career. In our case, we imagined this person was the company’s chief executive officer (CEO). This is an ideal scenario, but the initiator might also be a manager.
• The tech specialist: A technologist who has a technology-oriented approach toward AI. In our fictional conversation, it is the chief technology officer (CTO). This person usually tends to look at AI from the technical perspective and forgets the business side. The perspective of the tech specialist must be balanced with the perspective of the more business-oriented people.
• The skeptic: This is someone who does not believe there is any value in implementing AI. Often, these people argue that AI is either a fad or just “statistics”. Sometimes these people will gain interest in AI after learning more about it. Sometimes they need to see results. In our case, the skeptic is a marketing manager, but it could be anyone else in the organization.
• The AI leader: This person understands AI, hopefully has some experience with it, and makes sure that the technology is used in a way that brings value to the business. They have a deep understanding of the business and know the core principles of AI to find the sweet spot between the two. Hopefully, this person can be you after finishing this course.

The cornerstone of this conversation is the AI leader. In a couple of moments, this conversation could have taken the wrong turn and moved towards a potential dead-end in its effort to define an AI module. Luckily for our fictional company, the AI leader intervened and proposed good strategies to prevent this from happening.

The first problem in this conversation is that none of the participants in this conversation (the initiator, the skeptic, and the tech specialist) has enough knowledge of AI to make decisions. In these types of situations, the organization often follows the loudest voice. To make matters worse, the skeptic might raise some issues which are due to a lack of knowledge and are not easily addressable by someone without the required knowledge. This is a recipe for disaster!

Luckily, the AI leader proposed to run non-technical training for all the decision-makers. Education is the second key point in this conversation. Making sure that the “business people” (CEO, CTO, etc.) understand AI methods is essential for a successful implementation of AI in an organization. As long as you are not the hard-core technology development, the “business people” are at the helm. Hence, you want them to understand how AI works and what benefits it can bring to the organization.

After building the knowledge foundation of the business team, the skeptical rightfully reminds everyone that AI must be used as a tool that serves the business goals, and it is important to start with a clear AI vision. A clear path defines what AI can do for the business and how it will help the organization work more efficiently and serve its customers better.

Having a vision about how AI can help the organization is essential to integrating AI into its business model, and every organization must find it on its own. Building an AI vision precisely is an iterative process of design, experimentation, and redesign of the vision based on the experimentation’s success or failure. Neglecting the experimentation and focusing on the long-term strategy tends to end in major dead ends and failures. If you eliminate experimentation and redesign from the process, there is no way to ensure the organization is on the right track.

A good initial step to build a long-term AI vision is to start by focusing on AI projects that are well scoped. You can use AI to build new products and features or even optimize existing processes. By running AI projects, you will gain the needed experience and learn what you do not yet know about how to form an AI vision suitable for your organization. Figure 1 illustrates this process of moving from AI projects to building a meaningful AI vision.

Figure 4.1 | Your AI vision will be the product of several AI projects. That vision will rarely follow a straight path: you may have to build several AI projects that go in different directions and observe the result and learn.
Attribution: Zero to AI, Figure 7.1. Nicolò Valigi and Gianluca Mauro. Link to source. All rights reserved.

Next, we dive into how to design new AI projects, test them, prioritize them, and choose where to focus your energy.