The failure rates of projects related to Artificial Intelligence, Big Data, Machine Learning, Automation etc. is startling.
A recent study by some of the top publications including Harvard Business Review concluded that there was more than a 75% chance that projects revolving around Artificial Intelligence, Machine Learning, Deep Learning etc. will fail or not even take off.
Through this article, we wanted to share our experience and what we believe those points of failure are.
Having worked with some of the leading logistics, technology and oil & gas companies around the world, we could not agree more with the article by Harvard Business Review. We have been involved in projects that use next generation sensors, tracking devices, IoT devices etc. Our focus has been to help organizations improve productivity, eliminate waste and optimize the way they work. However, what we also found was that organizations begin initiatives focused on AI, Big Data, Machine Learning, etc. due to the pressure of having to be there. Even worse it is driven as a sales and marketing initiative without a clear vision, understanding or realization of the investment, requirements, goals and stakes involved. These projects are meant to result in the elimination of waste, improve productivity and create cost savings but often result in a waste of time, money and resources and a giant white elephant that can neither be disposed of nor be carried.
There are a few things we need to do to make sure that these projects succeed and go beyond the buzz that these words such as AI have created:
Key Performance Indicators and Goals 1. For any project, sit with the stakeholders and define clear goals of what we are trying to achieve. Think very practically where the organization is from business, financial and competitive point of view. Set goals that are measurable and achievable. 2. Focus on specific areas for improvement – If the initiative is a big data project, what kind of insights and how much of insights and information do we want? Ensure that the return on investment is beneficial by taking into account all dependencies. 3. If the initiative is related to automation, define what areas we expect to see as savings, elimination of waste, improvements in productivity etc. as the goal of the project and properly account for resources and infrastructure spending?
Viability and Feasibility 1. A lot of these initiatives are led by innovation teams. The biggest challenge with innovation is getting carried away and not focusing on specific goals to be achieved. 2. Clearly define long term and short term objectives for the project. The short term objectives must be tangible outcomes that can be achieved with realistic constraints and within the bounds the technology can deliver. 3. Any insights, objectives, etc. must properly be convertible and flow into the requirements backlog for the project so as to create a tangible outcome in the form of a new feature, enhancement or a product change.
Understanding requirements
- For any successful digital technology outcome, it is necessary to understand the importance of proper requirements. It is very important to understand how requirements flow from a high level, from the sponsors point of view, understand the user to define the user goals and finally the most important and detailed aspect, the system requirements.
- A large percentage of project failures can be attributed to not understanding the system requirements. These requirements fall into two categories – Functional and Non Functional requirements.
- Functional requirements specify the functionality that the developers must build into the product to enable users to accomplish their tasks, thereby satisfying the business requirements. In simpler words, functional requirements state what the system must do. Non Functional requirements are the constraints or standards that the system must have or comply with. Non-functional requirements define the system’s quality characteristics.
Technology Challenges and Collaborative Teams 1. Taking a very practical approach to innovation is important for success. It is important to understand the constraints and boundaries for any project. 2. There are two kinds of boundaries – those to which we can elevate human thinking and those boundaries within which new thinking, ideas and innovation can be really bought to life. 3. While the human mind provides limitless possibilities for innovation, we must consider the impact of ever changing requirements, constraints in technology, limitations in infrastructure and most importantly affordability and deadlines will all have on what can be done practically. 4. These constraints often come in the way of providing the best experiences and innovation as the outcome, but properly understanding the functional and non-functional requirements and examining those requirements from the point of view of desirability and viability will provide the best outcomes. A collaborative approach between the technology and design teams is critical to ensure that the right experiences can be so they don’t remain a concept on paper only. A lack of understanding between desirability, viability and feasibility often result in experiences that are below standard and innovation that is not effective. Hence a Human Centered approach that incorporates proper research, user insights, stakeholder requirements rapid and iterative ideation, prototyping and proof of concept is highly recommened.
