You'll be aware of the recent explosion of interest in AI, and most likely have seen it in action too, in the form of smart assistants on phones, desktops (think Cortana) and elsewhere. Pretty soon, the ability of machines to achieve a greater understanding of the real world in all its messiness will become commonplace, in all aspects of both our personal and professional lives.
The question here, however, is how to make use of it in your business.
At the heart of the ability of algorithms to create sense out of the world as we humans know it is big data. AI requires large volumes of information to analyse, and has largely garnered its recent traction through the availability of low-cost processing and mass storage.
Where does the magic happen?
The kinds of AI - such as Cortana - that we encounter frequently as individuals are using back-end cloud services and data to process queries and deliver answers. Other applications that are becoming more realistic and so prevalent include facial recognition, self-driving cars, and spoken language processing.
Business applications include chatbots to provide service and support to customers, and the prediction of demand for products based on a range of criteria, such as the weather, time of day and date, along with other, mainly business-specific criteria. Automation of manual processes is also high on the list, especially but not exclusively tasks traditionally seen as routine work: journalism and healthcare are just two areas where AI has made headway; for good or ill, be aware that this blog was created entirely by a human.
But these are still early days; there is a long way to go, and challenges abound.
AI applications require levels of processing power and mass storage that only the latest technologies can deliver. While the cost of CPU and storage have fallen, the kinds of hardware and software that AI demands remain specialised tools.
And the volumes of data that AI needs to be successful are vast: fortunately, IDC's estimate of the volume of data that will exist globally by 2020 - 44 zettabytes - is not the responsibility of any single organisation, but a significant proportion of it is likely to be needed for analysis by AI purposes. IDC predicts that worldwide revenues for suppliers of big data and business analytics will grow from $130.1 billion in 2016 to over $203 billion in 2020.
AI and the cloud
It's clear from analysis of AI's hardware and software requirements that cloud economics will play a key role in the development and deployment of AI-based applications in the enterprise. Right now, PaaS services exist to help in this process. Given that we are in the early stages of the technology, what is required is ability experiment with data and algorithms, allowing you to manipulate data and train a model using machine learning algorithms.
You need then to score the model, evaluate the results, and generate the final values. Once the process has reached a satisfactory conclusion, it should be possible to deploy the model as a web service for the business to use and evaluate. Of course, it's not as simple as that. It will take a number of iterations of these somewhat over-simplified steps, and probably the training of a number of models as part of the search for an optimal solution.
All that said, however, when expecting professional results from a technology as complex and specialised as AI, it makes business sense to call on the services of professionals, especially - as happens with cloud business model - they are able to amortise the costs of infrastructure over a range of customers.
You could even argue that AI is the cloud's next killer application.
 Worldwide Semiannual Big Data and Analytics Spending Guide. https://www.idc.com/getdoc.jsp?containerId=prUS41826116