AI and the Intelligent Cloud

By Michael Meyer, CIO and CSO, MRS BPO LLC

Founded in 1991, MRS BPO, LLC is a provider of Account Receivable and Back Office Solutions that enhance client’s return on investment and net back via compliant collections. Headquartered in New Jersey, USA, it also specializes in offering BPO services like 1st and 3rd party collections.

In the last 10 years, we have seen a number of new Artificial Intelligences (AIs) that have been born, spawned, or created. None of these AIs is yet--anywhere near the powerfully scary omnipresence of the Machine or Samaritan that was portrayed in the recent hit series “Person of Interest”. However, you know instinctually these AIs are growing more powerful and sophisticated with each new version that comes out.

To many of us, these new intelligences already live in the cloud and are able to touch our lives nearly everyday. At a high level, there are two types of AI that we can interact with, the first kind is consumer focused, friendly, free, and is always available to us. Well known examples of these intelligences are: Alexa from Amazon, Cortana from Microsoft, Goggle Assistant from Google, Siri from Apple. All of these are already at our beckon call. Viv from the creators of Siri (recently purchased by Samsung) will be next year’s free and cool new being to talk to.

The second type of intelligences are very powerful general systems that are used for business purposes: these are Amazon Machine Learning, Azure from Microsoft, Einstein from Salesforce, Adaptive Intelligent from Oracle, Watson from IBM and WolframAlpha (provides some Siri answers). There are many more AIs out there powering specialized services in different industries such as stock trading(e.g. Holly from Trade-Ideas).

These packaged and productized general intelligences are getting easier and cheaper to use. Amazon’s doesn't have a free test account and its pricing model prices their services by compute hour ($0.42) and the number of predictions that the software makes using a batch $0.10 per 1000 predictions and for real-time predictions they are $.00001 each. Microsoft provides a free account for their Azure platform that you can use for testing. The paid account provides the ability to scale up your machine’s resources in several ways such as the number of cores and memory available to run your calculations or programs with. They offer pay as you go, as well as monthly offers as well. Microsoft provides a calculator to help you determine your actual cost based on their options. IBM provides a free account to start testing their cognitive APIs that provide access to Language, Vision, Speech and Data tools. IBM recently cut their price to use their service by about 70% per query from $0.0089 to $0.0025. WolframAlpha provides their basic question and answer engine at no cost and their Pro offering has a fixed per monthly price of $5.49. Since its computational model is a question and answer engine, you are able to ask it about information that exists and for which an answer is available or can be calculated from one of the included formulas. WolframAlpha hasan amazing set of pre-created information, formulas and calculations that you can query easily.

The companies behind these AI systems and platforms are opening some of them up, in order to make them more useful using REST, JSONand other types of APIs. As more and more people start to interact with these cloud based intelligences, they will continue to get even more accessible, easy to use and more powerful. As computing power gets increasingly cheaper, these price of these cloud AI services will continue to decline and become more useful to a wider audience of business users.

Quote: “As more and more people start to interact with these cloud based intelligences, they will continue to get even more accessible, easy to use and more powerful.”