Understand why this detection method is called "You Only Look Once"
Creating an Intelligent Enterprise
Algorithms are the secrets of the most valuable digital companies
AI algorithms are already part of our daily lives. They have lived with us for a long time. In the early days of computing, batch processes demanded sorting algorithms (sorts) in order to place files in a certain order, better suited to their processing. With the Internet and the explosion of data generation, we have seen emblematic examples of the use of algorithms, which are the success of numerous companies, such as Google's search algorithms, recommended by Amazon and Netflix, or those that select the content to be used. shown in users' timelines on Facebook. Algorithms are also an essential part of new disruptive technologies, such as autonomous vehicles. An autonomous vehicle is a car filled with predictive algorithms and models, which make decisions at every moment. Looking at this, what do we observe? What are the best kept secrets of Amazon, Google and Facebook? Yes, your algorithms. These are the ones that make these companies so valuable.
Efficient algorithms can be a source of monetization for companies. A logistics company will be able to license its just-in-time delivery algorithms to a non-competing company, such as to a manufacturer of commercial refrigerators that, in partnership with a retail chain, automatically replenishes inventory, based on customers' consumption predictions. Futurology? No, it is a matter of looking at the logistics operation as a competitive differential. With the advent of the Internet of Things and 3D printers, the options and complexity of the logistics operation are considerably expanded, but on the other hand, new and disruptive business opportunities are opening up. This means that some operations usually seen as commoditized and placed under ERPs and other off-the-shelf software, can make a difference, if replaced by predictive algorithms. Is an HR system just payroll or can it include a sophisticated talent recruitment, assessment and retention algorithm that could be offered to the market?
It is an evolution from simple operational control, where all companies that use the same software operate in a similar way, for an analytical and predictive process. A process that is now commoditized can become a competitive differentiator. One more step in the trend that sooner or later, every company will become a technology company. Products will increasingly be valued for the sophistication of their algorithms. In fact, products will be embedded within services.
With the spread of the concept of "algorithm or AI economy" it will not be difficult to imagine a future where we will have "algorithms stores", composed of thousands of software components (algorithms) that can be combined to create more complex ones. An example is Algorithmia, which is a library of algorithms available for use.
As an example, let's look at a recommendation algorithm. He is constantly balancing what he knows with what he does not know. I explain: as a customer makes purchases, the algorithm begins to understand their buying habits and seeks to make recommendations that are of interest to the customer. On the other hand, he cannot just consider what he already knows, because sooner or later, the number of offers will tend to become less attractive. The algorithm has to make new explorations, try to offer something new and validate whether the customer likes it or not. The idea is to expand the set of offers, with attempts at mistakes and successes. The care is not to make a lot of mistakes, and to insistently offer things that don't matter, and not to stick to previous habits, only offering what is already running out. It is this balance that is the secret of algorithms like Amazon's. The level of efficiency of the algorithm is to achieve the optimum balance between known preferences and the potential for new opportunities in product categories that can expand the set of customer preferences. To do this, it is necessary to collect and analyze data on customer interactions with the company in all channels. The algorithm has to process this data and it self-adjusts itself to each client, so that the set of preferences of one is not necessarily applicable to another client. It is a 1-to-1 interaction.
I ask: how many e-commerce sites in Brazil adopt recommendation algorithms? I am amazed to see big stores at most listing products by type or price, ignoring buying habits and not exploring the potential for new sales.
But, faced with a scenario of frequent disruptions, we can go further with the algorithms. The digital disruption is making it clear that organizations that stick to models and rigid organizations are at serious risk of survival. Why not imagine an organization that self-adjusts its processes and business models based on algorithmic decisions? Well, we are still a long way from an algorithm that tells us exactly when and how to create a new business model. But, how about applying the same principles that self-adjust the algorithms for the management of organizations? A vast majority of a company's administrative processes are highly repetitive and do not require creativity. In practice, they do not add value to the business and many were designed for a world of 20 or 30 years ago.
We will then provoke a discussion: adopt algorithmic principles in the management of organizations. Looking at a company we see that at its highest level we have the vision (direction and purpose of the company), followed by the business model and supported by the processes, systems and organizational structure. The traditional practice has been to freeze the vision and the business model (created at the foundation of the company, perhaps tens of years ago) and only to incrementally innovate processes, systems and organizational structure. It worked very well when the scenario was more static and less disruptive, but does it work today? As for an Airbnb, Uber, WhatsApp generates angry reactions from the affected sectors and companies, is that the current, rigid model did not let them make the changes in a timely manner. Maybe they didn't even let them know that the world was changing around them.
Applying self-adjusting algorithmic principles, not only does the support layer (processes, systems and organizational structure) change, but the vision and business models also adjust themselves to market dynamics. The company is no longer a rigid structure with top down decisions going downhill across the organization.
It is not science fiction. I recommend reading an exciting article, “The Self-Tuning Enterprise”, at https://hbr.org/2015/06/the-self-tuning-enterprise, which shows a real case, the Chinese Alibaba, applying these concepts in its day by day.
Further details can be seen in the book “Smart Business: what Alibaba´s sucess reveals about the future of strategy”, by Ming Zeng, chairman of the Alibaba Group and member of its Academic Council.
At DataH we have a project called Ziggy Stardust that proposes to automate at least 70% of all administrative processes, creating an almost autonomous company, managed by algorithms.
Companies born in the Internet world are forced to constantly reinvent themselves as they are at the epicenter of the digital transformation vortex. But the changes don't just affect them. All organizations, sooner or later, will be sucked into this vortex. Thus, it is worth looking at the future from another perspective, considering that the organizational model we use today begins to lose validity in the face of a scenario of accelerated transformations. The future will be for self-adjusting companies. The stray will not survive.
About the author
I2AI makes up the Advisory Board of the largest Artificial Intelligence Innovation Network in the country
Twelve business institutions come together to define the strategy and guidelines for the performance of the innovation ecosystem. There is R $ 70 million available for the business sector to innovate with disruptive technologies
Why you need to know what this new and disruptive milestone in AI means for humanity
The secrets of Artificial Intelligence projects
Every day companies of all sizes and segments face the challenge of implementing their first project using Artificial Intelligence. However, this fascinating technology with infinite possibilities brings some requirements for