Are We in an AI Revolution? Part One of Three

It seems that no matter where you look, AI is the subject that is being talked about most. At BA Insight, it’s a daily topic both internally, as we design new capabilities, and externally as our customers and prospects want to learn about how AI can benefit them. This is for good reason, as my belief is that AI is poised to usher in a period of dramatic change in the way software interacts with users and data.

I think we are at the precipice of what history will view as the “AI Revolution” that will be talked about in much the same way we talk about the Internet Revolution.  A time in which the advent of the internet drastically altered the way in which people communicate and share information.

Think a little about the story of Amazon. Their initial business surrounded purchasing books via the internet, rather than a brick and mortar retailer.  Their success proved that internet-based commerce was a viable business model.  They could have stayed as a web retailer, but instead they innovated.  They expanded to more sectors, selling everything from electronics to appliances.  They invented concepts like “Amazon Prime” where buyers paid a subscription cost for special benefits.  They practically invented “platform as a service” with the market leading AWS service.  The snowball of innovation that led from those initial risks taken on this new thing called the Internet paid back massively. I think this story is bound to be repeated on the AI front.  AI technology is in the same stage that Amazon was in when it proved that internet-based commerce was real.  We know AI is real. We are now entering the next phase where we will see rapid advancement and capability stem from it.  It will be a fun ride.

I’ve spoken to several industry influencers who are quick to talk hype and skepticism. This is good advice, because at the root, the advice is to be cautious.  You should understand what AI is, understand how AI capabilities can help, and learn how to separate hype from what is real. The purpose of this blog is to help with that. I will break down the major players in AI, the categories of capabilities, and provide some guidance that can be used to evaluate AI in terms of search applications.

AI Players

I see three categories of players in AI.

Category one is the behemoths.  Huge companies with deep pockets who can either invest resources to innovate or acquire promising technology that has already innovated.  These are the Microsofts, Googles, Amazons and IBMs of the world.  Even Facebook is making an entrance here.  I’ve recently posted a blog just about them, so it’s worth a quick read in this context.  IBM’s Watson was likely the first experience many people had with a real example of AI, as it went toe-to-toe, and beat, two of the best human players to ever play Jeopardy. Personal examples have popped up, allowing individuals to interact with AI daily.  From Amazon Alexa, to Apple Siri and Microsoft Cortana, these nameplates are pumping out AI tech at a furious pace.

Category two is the little guys. These are often venture-backed smaller companies who are innovating niche areas of AI.  They are creating problem focused-algorithms or chat bots focused on a specific scenario.  These are companies that many have not heard of, but they are playing an important role in the maturity of AI. These companies tend to either get acquired or fizzle. I wrote in a previous blog about the rapid pace of acquisition in the AI space, and I expect this to continue and increase in pace. This category is worth mentioning as they will drive two big areas of AI growth.  The first being the need for AI frameworks accessible to the everyday developer, and innovation around specific features instead of generalized capabilities.  Niche players solve niche needs.  Once they prove they can do that, then they will likely see their tech get absorbed into one of the behemoths, which makes their innovations that much more accessible.  I’ve got two good examples in this area.  For those in the marketing technology space, look at Pattern89.  They’re building AI capabilities that help companies discover the best content to use for paid social initiatives.  A perfect example of AI innovation focused on a niche market. My second example is Semantic Machines.  Semantic Machines focused on conversational AI interfaces aimed at enabling people to communicate naturally with computers.  And to further drive home my take on the “little guys”, Semantic Machines was just acquired by Microsoft.

Category three are your open source initiatives.  Some of these overlap with the behemoths, depending on the strategy of the behemoth.  Microsoft, Google, and even Facebook contribute to, or have open sourced, pieces of their AI frameworks and capabilities. Some of these came from higher education, with a great example of this being Caffe, which is a deep learning framework that found its start at UC Berkeley. The great thing about open source initiatives is that it lowers the barrier of entry for cost conscious implementation, ensuring that AI is not just for the rich.

My expectation is that a complete AI strategy may involve pieces from all three categories, depending on the needs and use cases.  Ensuring that you understand the options and characteristics of each type of player will help ensure that you choose the right player for your specific needs.

Check back soon for the second part of this blog, where I’ll break down AI capabilities into categories and get into some details on specific capabilities.  In part three I’ll bring it all back around to search.