The first step in developing an AI solution is to discover and finalize a creative idea at the core of your solution. You may have devised several ideas, therefore you need to have a process for short listing the most viable idea and honing in on it. We recommend carrying out customer interviews and experiments that validate your assumptions and prevent you from delving into areas that are not relevant. We often use the method of Google Venture’s Sprints method which brings together a team of 5-8 people to find answers to big problems in just 5 days, and validates the answers with real customers. By doing this in only 5 days, it ensures the team focuses only on the important aspects, and ensures nobody becomes too attached to the solution before it is validated or discarded. If you are using a cloud-based AI service, you can also add developers to your SPRINT team, which will help you to develop a real and functioning prototype. Running these such experiments is critical to making fast progress.
Now you need to develop a strategy around the creative AI idea. Since AI is a new technology, you may find that for most companies, multiple iterations of strategic reviews and management approval are needed to get a go ahead. This is not only because AI is new in nature, but also management may not have a clear idea on the impact on the bottom line. Jeff Bezos, CEO of Amazon, says that many benefits resulting from AI are internal to the company, meaning a reduction of costs may not be immediately obvious. The value to customers can sometimes be translated in the form of significantly better personalization, higher accuracy, and ease of use.
As with any project, an AI project must start with absolute clarity on two areas defining the strategy:
- Value Proposition (VP): What new value are you offering to your customer segments? Which customer segments are you addressing and what are their needs? Will your new value offering better satisfy customer needs?
- Business Model (BM): How will your business benefit from the new offering? How will you monetize this value proposition? This involves a clear understanding of the sources of your revenue, costs, partners, and interactions involved.
These two areas describe the total value of your AI project to customers and to your business overall. You will find that these two areas are non-trivial and often need the highest quality of attention and time. The rest is a matter of implementation.
Articulating a VP for AI based projects requires an excellent understanding of the new capabilities that are enabled by AI today and in the near future. The low hanging fruits of AI capabilities are speech input, computer vision input, people and face recognition, image and object recognition and scene description. New capabilities are evolving every day. The core VP for your business will be centered around a creative AI idea that leverages one or more new capabilities of AI, in order to deliver new and unique value to your customers. The BM converts the VP into a profitable business. Development and approval of a solid VP and BM for AI solutions is the first and most important step for AI projects in most businesses. Since AI is a new technology with benefits that are often not well understood, this often takes much longer than expected. The availability of AI services and platforms help in implementation of the solution. The AI&U canvas is useful in capturing all critical elements of your solution, and allowing you and other decision makers to discuss, and form a consensus.