In today’s rapidly evolving technological landscape, companies face critical decisions when integrating artificial intelligence (AI) into their operations. There are two primary strategies for deploying AI within an organization: (1) the Big Bang IT approach and (2) the gradual integration method. Each strategy has its own set of advantages and challenges, and choosing the right one can significantly impact the success of your digital transformation.
This article explores these two battle plans, providing insights into their processes, potential pitfalls, and the lessons learned from real-world implementations. By understanding the nuances of both approaches, businesses can make informed decisions to drive productivity, enhance profitability, and achieve sustainable growth in the era of AI.
The Big Bang IT
A Big Bang IT involves undertaking a colossal digital transformation project. The goal is not to integrate new tech with the existing tech but to start from scratch.
Here is how the process unfolds: A consulting firm will work on your project for several months to develop a Transport Management System (TMS). However, during this time, your vision, organization, and operational scope will evolve. Consequently, you will end up with a result that is completely misaligned with your current needs and initial requirements. It’s important to understand that the cycle of a Big Bang IT is a V-cycle, with little room for maneuver and agility.
A Big Bang IT is an extremely long and sometimes counterproductive project for three reasons:
- It mobilizes all teams, leading to a loss of bandwidth.
- The ROI is not felt until the end, which means months or even years. Teams, seeing no medium-term benefits, may become demotivated.
- Companies always tackle the core of their organization to prepare for the future. To do this, they restart from technologies without truly understanding their needs and use cases. However, starting from scratch to build your own solution is not necessarily the best idea because there are already solutions that allow adapting the existing ones, thus saving time and money! In recent years, at Kardinal, we have witnessed a lot of Big Bang IT projects turning into nightmares.
Here are the top three biggest Big Bang IT failures:
A large group spent over 5000 man-days developing a TMS, which ultimately could not integrate route optimization. A Big Bang IT can really be a hassle!
A solution was to be implemented for optimization in a company, but once on-site, it was explained that the information system was being overhauled, so we had to wait. This overhaul was supposed to last 6 months and ended up taking 2 years. Why? New people with a different vision had joined the company and wanted to modify certain aspects of the project. It is crucial to understand that any minor change in an organization can complicate a digital transformation process.
Lastly, a major carrier worked for over a year with an optimization solution provider to integrate all the business constraints. On the D-day, when the solution had to be deployed, the operators were so overwhelmed by the rules that they could not use the solution! The solution brought more complexity than solutions…
Gradual tech integration
To avoid the burdens and failures of a Big Bang IT, we advise integrating tech gradually into your company. This battle plan relies on a much more agile and iterative ROI-based approach than Big Bang IT.
The curve of this solution evolves steadily because the solution is built with the client. The scope expands as learnings are made, gradually increasing productivity and profitability. Initially, the goal is not perfection but a Minimum Viable Product (the most minimal yet functional version of a product) that can be improved.
Also, this battle plan will only work if it is well initiated by your management and understood by your operators. It must be a genuine project, a strategic priority for your company. Digitalization should not serve non-central and scattered projects. Additionally, a dedicated team should handle this battle plan. Ideally, it should include a project manager, a person from the logistics field, and another from IT. Avoid assigning the project to a newcomer or solely to the IT department. Regular meetings between project leaders and your management should be organized to keep your teams motivated and demonstrate the importance of their work.
Here are the main stage of this methodology:
Defining Pain Points First, pain points are defined. In a logistics organization, there is always a bottleneck. To identify the pain points, you need to ask: what challenges need to be resolved? What is the company’s project? To properly analyze your pain points, you need structured data. Data helps to accurately target problematic use cases. For example, in delivery time slots, if 5% of the total delivery volume is on appointment, it is manageable by an operator. But if this figure increases significantly, for example due to the growth of the B2C market, it will create extremely complex organizational issues in tour planning. Implementing a route optimization solution is a good response to this problem. You can also rely on your operators to clearly define your pain points. Thanks to their process expertise, they know where there are gains and where to dig for value.
Defining the Roadmap When the project is clear and the data well-structured, data is used to find improvement keys. A roadmap is established with all the use cases that need to be addressed. The focus is on a high-value use case. The project is then split as much as possible to have clear stages. Next, KPIs (ideally 5) for project progress and success are defined. Common KPIs include productivity, number of tasks per driver, cost (OPEX), kilometers traveled, and service quality.
Iterative Implementation This approach is divided into five steps:
- Implementing a pilot with an initial simulation phase on a restricted and isolated perimeter,
- Validating the pilot by operators based on a one-day dataset. Gradually, the perimeter is expanded, ensuring value at each step.
- Field testing, for instance, with two systems in parallel production. Ultimately, users will choose one of the two systems. The test can also be done on a perimeter in a depot, with a planner working on a small group of drivers. The idea is to have very qualitative outputs.
- Validating KPIs,
- Deploying the project on a broader scope.
With this well-managed chronology, operators will trust the solution. This is crucial because the algorithm does not think the same way as a human. AI has a more macro vision than operators, which can unsettle them. Trust is built through simulations, tests, exchanges, and user training. As iterations progress, the project will become more ambitious and grow in scope. This battle plan should be seen as a kind of spiral that grows and exponentially generates growth.
In conclusion, the path to successfully integrating AI into your business requires careful consideration of your organization’s specific needs, goals, and capacities. The Big Bang IT approach offers a comprehensive overhaul but comes with significant risks and potential for misalignment with evolving business needs. On the other hand, a gradual integration strategy allows for more flexibility, continuous improvement, and alignment with real-time feedback from users. While each method has its pros and cons, the key to a successful digital transformation lies in a well-defined strategy, strong leadership, and a committed, collaborative team. By learning from past experiences and adapting the best practices, businesses can harness the power of AI to drive innovation, efficiency, and long-term success.