After the COVID-19 pandemic and the harsh economic impact of the disease, the steady decline in business profitability across multiple industries continues to threaten the erosion of investing in innovation. Fortunately, at DFX5 we have a novel solution: use artificial intelligence to generate new sources of revenue. Yes, just like artists use machine learning to draw a painting or compose a symphony, so can your business develop new sources of revenue by using artificial intelligence.

The idea called ‘DDD’, pioneered by our Director of Artificial Intelligence Charlie Rosado, is simple in its beauty: Data-Drives-Decisions. In a ‘DDD’ model, multiple machine learning algorithms are used to drive decisions within controlled environments without human intervention. The models decide how the app ‘looks and feels’, the items offered, the prices, the personalized coupons sent to each unique user, and the overall customer engagement strategy and channels. All decisions are driven by data as understood by artificial intelligence models.

Why does this revolutionary idea yield new sources of revenue? Because the machine learning algorithms have personalities and unique capabilities that go beyond human power. The diverse algorithms focus on their own unique vision of the information and they individually identify patterns, tendencies, and potential opportunities. When combined, all these different perspectives create new sources of revenue.

At DFX-5 we use a multi-disciplinary approach to create your unique roadmap for success. The new sources of revenue are highly dependent on having a full understanding of the current business environments. To gain such advantages we have a three-step process recipe to develop the new sources of revenue:

• Optimize Business Critical Processes
• Discovery
• Trust in the Process

Optimize Business Critical Processes

It all begins with a proper understanding of current business operations. To obtain such an understanding a firm must already be fully integrated in cloud-enabled solutions and environments. Proper data governance and guardrails must be in place, with full integration of systems in a scheduled and well documented format. These pre-existing requirements are consumed by the machine learning algorithm models. Each model will focus on different sections of the information and each is designed to look for unique types of patterns within the data. The models use diverse mathematical equations to produce their own unique ‘understanding’ of the business operations. These understandings or perspectives are then used as the building block to find ways to optimize or improve what is already there. In our experience, by first understanding the status and optimizing the operations, highly valuable insights are gained and at times financial yields are obtained in savings. While the objective of the step is to gain a full understanding of the business and seek to optimize current operations, many times secondary gains are obtained.

Discovery

Discovery begins once the machine learning models have developed their own unique perspectives on the business. The models, fully automated and acting on their own volition, will attempt to discover new patterns to predict future outcomes. The predictions are then validated and evaluated. The results, whether accurate or disparate, are then fed back to the model. The model will integrate the new results into their perspectives and will now weigh in that result to impact the next prediction. The iterative process is repeated ad infinitum; thus the models are in a constant discovery mode, in where each is forecasting and learning from its results. Mathematical outcomes tend to yield an infinite approach towards the desired outcome, whether it is increased sales, diminished wait times, or any other key metric.

Trust the Process

Most of our clients tend to be highly intrigued with using machine learning tools to generate new sources of revenue but are afraid to implement. The most common fear is uncertainty of what may happen given the youth of the technology. This is where the power of using certified AWS solutions architects and engineers is important. Our team ensures that the process is done within secure and highly controlled environments. We implement guardrails that protect the business, limit the ability of the algorithms to perform actions, budget expenditures, control allocations, and safely protect from any disaster recovery. The process begins with training in environments that are safe (imagine training wheels, protective gear, large helmet, and lots of cushion), then we begin to start creating a completely real business environment. Once the model is training in a real-time business environment, then we deploy it in a controlled format, sometimes referred to A/B testing. Once confident the model can perform, we begin diverting some minimal traffic and gauge performance. Once the model has trained in a safety, then deployed in a real-time controlled business environment, and then processed real-traffic at a small scale, we begin to divert more traffic, until the point that all traffic events are handled by the model autonomously. This is trusting the process, and knowing that at the end, given we have understood the business properly, discovered new hidden patterns, and then deployed in a robust and safe manner, a winning formula that discovers new sources of revenue is created.