Businesses Take on #AI

I have recently joined a program on the application of Artificial Intelligence (AI) for businesses. The main obstacles, as I have learned through the course, are that most businesses do not first and foremost understand the technology and secondly, can not, in most of the cases, justify the investment if their internal processes are profitable.

We have to also accept that many small to medium size companies would be wary of additional costs involved in developing or creating a platform that is driven by the various AI technologies that are there; hence many would not have funds set aside for research and development or innovations.

To support businesses take on AI, I would suggest they get some basic understanding of the technology that will subsequently help in building a business case to justify the investments.

Understanding AI

AI or Artificial Intelligence comprises a vast set of technologies that can be an abstract unless the business models are fully outlined and patterns can be identified. In the program we have differentiated between what is known as "narrow AI" for machines to address a defined and set problems, and "general AI" which, in theory, is meant to solve many of the problems that we throw at it and most of the hype is around it, although the technology is yet to fully mature.

Another aspect to understand is that the applications of AI are unlike other technologies in a sense that they need to be developed vs applied directly. For example, if a company is going through a digital transformation to be more agile. They would maybe choose to transfer their infrastructure onto the Cloud and invest in a wide set of applications to enable their business to run from remote locations. With AI, the problem to solve is to be defined and a model would need to be built and a choice of which area of AI can be used must be decided upon and these areas include but not limited to:

• Machine Learning
• Natural Language Processing
• Robotics

If the business problem is automation, robotics or software can be programmed with a defined set of rules to support with the process. If the rules are all known, this form of AI would fall under the "narrow AI" field since the machine does not need to learn beyond what the program is set to do.

On the other hand, if the business problem is more complex and requires the machine to pick up patterns and run more analysis; Machine Learning and Deep Learning can be leveraged with algorithms to tackle the data sets provided and run a system that evolves with usage. An example would be a security system for facial recognition. The patterns to identify the discrepancies in human faces and learn from them is essential since photos for every human being on the planet cannot be obtained, obviously.

Another model that requires learning, is the use of autonomous robotics. The usage of these is heavily subjected to workplace safety regulations as the need to safe guard humans from the technology is vital.

As more business problems take on AI; the market would mature and applications that can be configured, relatively easy, might become available and require less development and more configurations to lower the time and cost involved in the process. For robotics this is true of many of the models being built by vendors who have that in mind. For example, this company builds robotics that can be used in any warehouse.

Invest in the Future

In response to the financial difficulties with applying AI; we need a business model / case that measures the monetary gain behind increased efficiency and throughput per business, naturally. With most advancements in technologies, the benefits focus on increasing the speed of how tasks are run and subsequently increasing production or service level in a non-product focused model. The challenge comes from the iterative cycle and the experimental nature with AI as it is still a new technology, that might translate to prolonged and delayed ROI (Return on Investment). Setting the expectations of the stake holders and the team involved in the process is key to ensure successful pursuit of a working AI model.

On the other hand, companies should not get distracted with the urge to adapt AI for the sake of it. Tech. giants like Facebook, Google, IBM, and Microsoft among others who are heavily investing in AI must do so, as their entire business model is based on innovation. Smaller companies can chose fewer initiatives and might benefit from existing models that are there. For example, a local retail shop might leverage a smart mobile App to help shoppers and enrich their experience with relevant details about the products on display.

More mature companies can cap the time allocated to get an AI model working and can also run with it as a parallel process to their existing working models.

In all cases, we can't deny the impact AI has on improving the brand image and it will always be best for any business to future proof its working models!

Have your say; Does your company have a working AI model? or are they still considering their options?