Is “Enterprise AI” different than Business/Enterprise use of AI?

Spoiler Alert: Yes

August 6, 2018

By Mark Stadtmueller, VP, Product Strategy, lucd.ai

A florist using a van for flower deliveries is not a transportation company in the same way that an Airport Shuttle service is or a rental van company or a logistics company. The florist is using a van, but the Airport Shuttle, Van Rental, and Logistics company leverage vans as intrinsic parts of their business. They very much differentiate and compete with their Vans. Indeed, logistics companies can and do take over the delivery aspects of the floristry business because of their transportation capabilities.

With AI, there is a similar business relationship emerging. Amazon Echo contains AI. But, when a business uses an Amazon Echo, it does not make them an AI company. Businesses of all shapes and sizes are starting to use AI to help them do their business. However, in order to participate, grow, enter new markets, and gain competitive advantage in the new digital economy, AI usage is not enough. “Enterprise AI” is about building AI as a core competency integral to a business.  However, unlike vans that are core to specific industries, AI can and will touch every industry. Therefore, every industry will be rearranged with different AI enabled growth vectors. So, while every business will use AI, those businesses who adopt “Enterprise AI” will successfully navigate industry realignment and capitalize on those new growth opportunities.

AI definition

It is no great secret that AI is one of the most exciting new technologies of our time and its potential to transform business is significant. To define “Enterprise AI”, it is important to level set on AI.

AI is easiest to define in the context of the following terms: Artificial Intelligence; Machine Learning; Neural Networks; Deep Neural Networks; and Deep Learning. Artificial Intelligence is a marketing term that encompasses many things but in general it is a concept. The concept is getting a machine/computer/robot to simulate decision making and do what was previously perceived possible only via human interactions.  Machine Learning is a type of AI where algorithms are used to analyze data and detect, classify, segment, predict, and recommend things. The machine learning analysis involves looking for patterns within the data and then creating and refining a model/equation that best approximates the data pattern. With this model/equation, predictions can be made on new data that follows that data pattern. Y = mx + b, the equation of line, is a basic example of such a model where an output (Y) is proportional to an input value (x) multiplied by a weight/slope (m) and offset by a constant (b). Y is predicted to vary linearly with x. Therefore if you know x, you can make a prediction of what Y will be.

Neural Networks are a type of Machine Learning in which brain neuron behavior is approximated to model many input values to determine or predict an outcome. Layers of neurons are used to improve the predictive capability. When many layers of neurons are used, the type of Machine Learning is called a “Deep Neural Network”. Deep Neural Networks have been very successful in improving the accuracy of speech recognition, computer vision, natural language processing and other predictive capabilities. When using Deep Neural Networks, people refer to using “Deep Learning”.  Deep Learning is the act of using a Deep Neural Network to perform Machine Learning which is a type of AI. However, because of the active work and success of “Deep Learning” in many circles, people equate “Deep Learning” with AI.

The excitement and rightful attention in business around AI is specifically the potential that Deep Learning has to offer to transform business.

AI industry segments: Innovators, Space Race, Businesses

Often industries are segmented per market vertical or size of company or revenue or geographic coverage. But, AI is best segmented as follows:

AI Innovators: These are people and companies that are building a specific capability or product or service that is powered by AI (Deep Learning). They collect data, train and build a model on that data and then build an offering that includes and leverages that trained model. They then sell or provide that offering to their market. Amazon and Echo/Alexa is an example of an AI Innovator. They built a basic home personal assistant (Echo) that bundles in AI (Alexa speech recognition and NLP) and they sell it to consumers and businesses as a product. There are many AI innovators ranging from Geospatial Imagery Analysis to Facial Recognition to Picture identifiers and most AI Startups fall into this category.

Space Race AI: There is also an AI Space Race going on. So far, this Space Race is dominated by the likes of Google, Facebook, Microsoft, Amazon, Baidu and others. This Space Race isn’t so much about specific capabilities but about the realization that AI is becoming a dominant technology and leadership in this technology will likely determine technology dominance and survival in the next decades. Like TCP/IP or Search or Wireless in the past, AI leadership will determine the future technology leaders. So, they are pushing the envelope and needing ever more powerful capabilities like advanced GPU innovation from Nvidia. For these companies in the “AI Space Race”, scale, performance, new types of Deep Neural Networks, breakthroughs, and innovations dominate and are their AI focus. Those “Space Race” companies are certainly AI Innovators as well and offer AI products/services that are a result of their work.

Business or Enterprise AI: These are companies that are looking to leverage AI that is based on data unique to their specific business and modeled to support their specific goals around new products and services, new customer interactions or new ways of doing business. This is different than a business that buys an AI capability from an AI innovator. Business or Enterprise AI (note: will use the term Enterprise AI for simplicity going forward) is about a core AI competency and ability to produce and innovate based on AI. It is a business that invests to grow business specific AI capabilities.

Characteristics of a business adopting Enterprise AI.

At lucd.ai, our experience working to turn data into successful enterprise AI outcomes shows that businesses with the following characteristics have the most success:

Data Handling is a core competency: As stated here, data is the fuel for AI and AI needs lots of that fuel. In order to differentiate itself, a business needs to handle data that is central to their specific business. That data will be both internal containing specific internal information that can be used for new growth opportunities as well as external public data that is relevant to their business and then must be fused with internal data. The data handling needs to be comprehensive handling volume, velocity, veracity, variety. The data handling needs to be flexible and open-ended supporting both current initiatives as well as ready to support new, not yet planned initiatives.

Data knowledge, access control, and attention are included in mission statements. Businesses adopting AI will have strong core competency in Data Handling as mentioned above. But, end to end data management needs to be governed by executive level understanding of how data is used, data compliance ensured, data security achieved, and proper roll-based access controls are in place. Making sure that AI is trained only on proper data that delivers positive outcomes for a business’s customers becomes capability that executives know they have and thoroughly understand.

The focus of Enterprise AI is different than Business Intelligence. While BI is a mature area for understanding, monitoring, and improving performance of business, Enterprise AI is different in focus that Business Intel. Enterprise AI is about new capabilities and new markets. Enterprise AI is a key pillar to digital transformation that results in new or better products and services, new or better customer interactions, and new or better ways of doing business.

Enterprise AI affords a business to easily incorporate third party innovation. The pace and scale of AI innovation cannot be overstated. Innovation in building advanced and better performing Deep Neural Networks is rapid and widespread. Modern AI was born in the era of open standards, collaboration across entities and innovation outside traditional corporate boundaries. A business adopting Enterprise AI needs to avail itself to all the innovations taking place. It is too risky to limit AI adoption to only internal or proprietary models or current vendor eco-systems. A business embarking on Enterprise AI needs to be able to “bring all models”. The ability to understand how to “bring their data” and “bring all models” as well as understanding how and when transfer learning can take place is critical for Enterprise AI.

Optimization and Efficient Scaling is critical for Enterprise AI. By its nature, Enterprise AI needs to afford a business the opportunities to capitalize on AI and therefore support many and varied potential innovations. Therefore, the ability to automatically tune deep neural networks without constant manual efforts as well as the ability to efficiently scale and optimize resources becomes critical. Enterprise AI is not a one off linear pipeline initiative like that of an AI innovator.  Therefore the ability to dynamically scale and automatically optimize models and resources is paramount.

Agile is critical to Enterprise AI. No business goes from zero to Enterprise AI immediately. Agile adoption and iteration is required where small sets of data, AI models, and accuracies are developed and deployed in an offering. That process is rapidly iterated and expanded rapidly by using the same common Enterprise AI platform and methodologies. Also, responsiveness is achieved within the common platform and processes. Agile capabilities are needed to empower a business to succeed leveraging Enterprise AI.

The future belongs to businesses that adopt Enterprise AI.

It is now somewhat of an old adage that a retailer that puts up a website did not become an internet company. There are core changes that needed to take place to become an internet company. This is now the same with AI. A business that uses an AI product does not become an AI company. Only through leveraging “Enterprise AI” does that company avail itself to the new market opportunities that AI offers. So, yes, there is a difference between “Enterprise AI” and a Business/Enterprise that uses AI. And that difference will very much define the future of those two types of companies.