Where is all the AI?

The next generation of business analytics

Given all the hype (and fear-mongering) around AI it is a challenge to put your finger on where it is being used, and whether it is worth investing in. In fact, there is probably less AI in use than you imagine. You are far more likely to be interacting with a person when you reach out to customer services, for example, than with an “intelligent” program.

There are some incredible companies out there that are revolutionizing industry performance with the judicious application of AI. Consider those well-designed online stores that cope with spelling errors, and synonyms. Many industries, however, simply have not put AI into use yet, which is why early adoption becomes such an interesting and challenging barrier. Is your business ready? It very likely is and there is no better place to start than with business analytics.

Do you have a burning question about your business? Perhaps you keep thinking “If we only knew … “? If so, you are probably ready to add algorithms to your businesses intelligence arsenal. The challenge, however, that comes with adopting algorithms, machine learning or artificial intelligence (AI), is that there is no one-size-fits-all approach. Much like incorporating any data analytics service, adopting AI means applying a highly personalized service. That means identifying experts that you can trust to have both the expertise required and the professionalism to provide you with a cost-effective solution.

You are not going to be an early adopter until you know what these next-generation tools offer and grasp the essence of what they are and how they work. This is an essential first step to being able to tap into their incredible potential. This article explains the basic concepts and provides some tantalizing insights into the incredible applications of machine-based business analytics, because AI is already here and it is already helping.

The algorithm

At the heart of all machine learning and AI applications lies the simple algorithm; a set of instructions that a computer can execute. Whenever you apply If/then logic with a machine you are running an algorithm. Algorithms belong on a spectrum, from simple to complex. They may be fixed and rules-based, or the set used may evolve and adapt to changing circumstances.

The rules-based algorithm

Rules-based algorithms may clean and analyse data, identify patterns or predict trends, for example, fluctuations in demand for your product, or in the availability of the raw materials you need. Such predictive analysis empowers you to respond before an event, rather than reacting to it.

You may apply algorithms to prescribe business processes. Say you have a massive logistical operation, with labour law constraints dictating the movement and availability of staff. Timetabling such constraints using a traditional spreadsheet can be both time-consuming and error-prone. By extracting the rules that apply to your operations, many different algorithms may be tested and the one that best fits your objective applied. Whether that be maximising staff availability or reducing stock loss etc. Adopting such an approach allows you to test and reject possible solutions before they are ever allowed to impact operations.

Machine learning

When your data set is vast, and even your analysts can not make a guess at the variables that are influencing the outcomes, you may call on machine learning. Machine learning is an excellent tool for analysing massive quantities of data, that a person may dismiss as chaotic. It is the machine learning revolution that is driving the increasing efficiencies in supply chain management as well as the fusion of the world of industrial production and the internet.

Traditional forecasting in sales, for example, would interpret data on: previous sales, season and possibly include some marketing data. The computing power available, often created limitations, and the cleaning and shaping of the data represented a huge time investment. With the power to mine enormous datasets it is no longer necessary to pre-guess the variables that may be influencing your business outcomes. Machine-learning forecasting may include variables such as brand features, returns, the sales channels, and their marketing initiatives. The predictive model can even examine the fine details, such as the wrappings used. Machine learning allows an almost unlimited data-set to be examined, taking full advantage of the computing power available to us.

By continuing to analyse your data, and continuously demanding the optimal outcome, new solutions may be offered in response to the changing dynamics of the business environment. Whether that be behaviour of: the competition, the weather or your customers. A dynamic approach such as this massively improves the efficiencies of business processes.

 Artificial Intelligence

The definition of AI depends much on who you ask. It helps to distinguish AI from machine learning and define it as the application of algorithms and logic to scenarios where we do not have an existing solution. To your business, this means that solutions may be offered to as yet, unforeseen problems.

Consider the situations where the programmer is unable to provide algorithms suitable for the problem at hand. For example, Facebook‘s application of AI to create an effective facial-recognition algorithm set. People are excellent at identifying individuals, but there was no way for the developers to extract a human methodology and program it. Facebook had an enormous dataset, millions of faces photographed in myriad: lighting conditions, angles, and expressions – all conveniently tagged with the person’s identity. This data was provided, and the AI system began attempting to create predictive rules to identify a person. Development was an iterative process; with algorithms constructed, and their ability to provide the correct answer tested. Only the algorithms that created the best outcome survived to be modified, improved upon and applied.

Provided with the data, AI can provide accurate predictions of opportunities or risks. AI is the biggest fish in our data pond. With its ability to provide innovative solutions to problems based on data and the desired outcome, its potential is enormous.

 Consider the impact

Consider the complex vacation application and approval procedure that took one person a quarter of a year to administer, replaced with one algorithm-set. Or the essential service provider who required an overhaul of their route-planning strategy. They have seen a 20% cut in their daily mileage and all the associated costs.

Big data is exactly that when you are linked to the internet of things (IoT) and consuming the data streams they produce. Customers often need data analysed in such quantities that machine learning strategies are the only ones applicable to the task.

Next time you have a delivery arrive at your home, consider this. Did the person in the warehouse select the optimal path around the distribution center to pack your goods, or was it an algorithm? Algorithms are already improving your day to day experience and their impact will only grow.

Sources consulted

https://www.expertsystem.com/machine-learning-definition/

https://en.wikipedia.org/wiki/Artificial_intelligence

 https://www.techemergence.com/how-to-apply-machine-learning-to-business-problems/

https://www.cio.com/article/3276318/artificial-intelligence/what-can-machine-learning-do-for-your-business-right-now.html

https://ec.europa.eu/growth/tools-databases/dem/monitor/sites/default/files/DTM_Industrie%204.0.pdf

One thought on “Where is all the AI?

Leave a Reply

Your email address will not be published. Required fields are marked *

four + four =