Google’s Unfair Advantage — Why Your Company Should Focus on Artificial Intelligence

Why did Google spend over $500 million recently on two Deep Learning companies?

Why have all the other major tech companies followed suit snatching up as many artificial intelligence researchers as possible?

To stay competitive.

At the rate that technology is advancing, it’s simply riskier not to pursue AI.

Large companies wouldn’t be spending so heavily to monopolize talent unless they believed it will give them a powerful edge and unlock potential returns.

The real question is, if your business isn’t in the Silicon Valley bubble, is AI still important? Absolutely.

Predicting human behavior is such an important aspect in corporate decision-making that it dictates the future of new products and growth. AI provides companies a clearer and deeper understanding of the needs and wants of their consumers. This allows companies to continuously provide value and retain or grow their customer base instead of stagnating.

Technology Threatens the Dominance of All Great Companies

Big companies can’t ever out-innovate the market. To stay big, companies need to be willing to exit old businesses and enter new ones — and do it quite boldly. — Richard N. Foster, from MIT Technology Review

The average length of a company on the S&P 500 is shrinking fast, mostly due to the inability of larger companies to innovate faster and better than smaller companies.

Experts state that three out of four giants on the S&P 500 today — giants like Apple, Corning, Ford, Intel, and P&G — won’t be on there in the 5 years with the current turnover rate.

That’s shocking.

Big Data has been touted as the technology of choice for corporations to innovate and stay competitive. Many of which have been banging on the doors of the Cloudera’s of the world to find out how to store and make sense of the copious amounts of data being generated at an exponential scale.

It might be hard to believe, but not every company or industry has Big Data, even though people still think they do. Unless your company is generating data at the rate of Google, Facebook and the NSA, you’ll need more advanced technologies to glean useful insights out of your data set.

That’s where AI comes in.

Why Every Company Will Have an Internal AI Group

Techniques such as deep learning could help companies make smarter inferences about their customers. They will be able to identify preferences and make predictions, such as when customers are most likely to want to be contacted or which ones are most at risk of not renewing a contract. — Matt McIlwain, Partner at Madrona, a venture capital firm in Seattle

Not long ago, companies didn’t have IT Groups. Now they are commonplace. The increased productivity has been staggering to the point where it’s hard to imagine day-to-day life without a solid IT infrastructure.

The same will hold true for AI Groups.

But just like any good home, it’s important to build a strong foundation before putting the house on top. That’s why we’re seeing Google and others acquiring AI talent and AI technology. They know that bringing in the talent will help them grow their technology with the company and adapt to ever-changing needs.

Having an Internal AI group allows a company to fully comprehend the data they generate. This will be the main reason we won’t see external AI companies working on complex corporate-wide problems outside of logistics and efficiency analytics.

Most companies are sitting on a wide variety of data silos across their organization. Investing in an internal AI Group allows the necessary time to deeply understand each silo before attempting to glean deeper meaning and correlations from the data.

Specialized Deep Learning — The AI Technique That Bridges Data Silos

Data science allows you to see the context between the silos

It’s not how much data you have, but the quality that truly matters.

Your company most likely has hard-to-reproduce data that has tremendous value, but isn’t at the scale where Big Data techniques can be used. To make this data usable, you’ll need to rely on data scientists with domain expertise in the field in which the data resides to apply the appropriate Specialized Deep Learning to extract the real value.

Firstly, Deep Learning is a technique within AI that enables computers to piece together the context of data without a data scientist doing all the grunt work by hand. This makes it very powerful for analyzing datasets.

In our recent open letter to Yann LeCun, the Director of AI Research at Facebook, we stated that there are two types of Deep Learning, Generalized Deep Learning and Specialized Deep Learning. If your company has actual Big Data, Generalized Deep Learning works fantastic as the techniques applied are more data centric and supervised in nature.

For smaller datasets, such as those in the wearable or health spheres, Specialized Deep Learning is necessary. Specialized Deep Learning leverages the viewpoint of domain experts who know what to specifically look for to train the data for features that are not obvious.

It’s dangerous to think that a data scientist can throw deep learning algorithms at a problem without lots of data and expect it to work. No two problems are the same and therefore there is no one size all approach to AI.

Here’s an example of why Specialized Deep Learning is necessary: if you put a biologist, chemist, mathematician, psychologist, and computer programmer in a room and asked them to determine a way to use data to improve a woman’s health, you’re setting the team up for failure. Although each individual has domain expertise, the nuances of the small — but meaningful — insights that go into the creation of Deep Learning algorithms will be lost unless one person has been in the trenches of each discipline enough to put all the puzzle pieces together.

Ultimately, the Holy Grail is for Deep Learning to be coupled with Big Data. With the rate of data expansion, this will be realized in the near future. Therefore, it’s important to create an infrastructure that utilizes AI on the gathered data to then guide the decisions of the company on what necessary data is needed to gather to generate the most benefit.

The natural convergence of AI and big data is a crucial emerging technology space. Increasingly, big businesses will need the AI technology to overcome the challenges or handle the speed with which information is changing in the current business environment. — Dr. Jim Hendler, director of the Rensselaer Institute for Data Exploration and Application (IDEA)

Benefits Take Time — Better to Start Now

The greatest benefit from AI will come within large corporations with the level of resources needed to analyze the data and with the drive of shareholders to push the envelope of innovation and return.

Know this: AI is not a magic pill. A company cannot just throw AI or a few people at a problem and expect unrealistic returns to happen overnight.

AI takes time. Engineers need to lay the foundation within your company to utilize their own domain knowledge and created technology and integrate it with the company’s current data silos and direction they hope to go. Building the support needed to make technologies like Deep Learning usable to benefit the corporation is the real art and is the main reason internal AI Groups are necessary.

From there these engineers can create meaning from current technology and build the infrasture for the future data sources needed to take the company to the next level and keep them competitive.

Disruptive technologies take time, but once they are implemented and worked on, massive growth and insights can be realized.

Is your company ready for AI?

This article was originally posted on LinkedIn.