Artificial Intelligence (AI) is quickly playing a role in almost every business sector. There is a growing sense of urgency for organizations that have heard about the benefits of AI but have delayed taking action or don’t know where to start. In fact, history is repeating itself, as companies are facing a similar situation to those in the 1990s that lacked an internet strategy. Now is the time for organizations to formulate an AI strategy, and Intel is here to help.
We all know that technology is disruptive. 52% of companies in the Fortune 500 have disappeared since 2000, often due to technology disruption. For example:
- In 1888, George Eastman of Kodak invented the first camera that could be used by amateur photographers. In 1977, Kodak patented the first digital camera, which was productized in 2003. In 2012, Kodak filed for Chapter 11 bankruptcy.
- In 1998, Nokia became the #1 mobile phone brand, and developed a mobile phone with a touchscreen in 2000. In 2013, Nokia sold its mobile device business to Microsoft.
- Wal-Mart opened its first store in 1962 and reached $100B in sales by 1997. In 1995, Amazon was launched as a website that only sold books; by 2015 they surpassed Wal-Mart as the most valuable retailer in the United States.
Artificial Intelligence is not new – however, recent breakthroughs in deep learning that deliver accurate image and voice labeling are enabling a 4th Industrial Revolution – led by AI. Artificial intelligence sits at the top level for any data analytics capability. Analytics is a constantly evolving science that companies leverage for insight, innovation, and competitive advantage.
Analytics has changed over the years and continues to advance through five stages of increasing scale and maturity: descriptive, diagnostic, predictive, prescriptive, and cognitive. Leaders are moving up the data analytics capability curve to Artificial Intelligence and using data to sense, reason, act, and adapt to generate value.
When developing an AI strategy leaders have to develop the business case, manage the capability, and fund it. A stair-step approach works well by mapping the current assets and how they can be used. Start by mapping data sources, infrastructure, skills, and problems. The next step is to prioritize the problems AI can solve based on their business value. For funding and infrastructure, it is possible to use what you already have - most AI workloads are running on Intel® Xeon® processors today.
AI is a fast-growing workload in the data center, as more businesses adopt AI to deliver greater efficiencies. Most AI applications are for classic machine learning which lends itself well to traditional computing solutions. Deep learning is a growing workload, which lends itself to highly parallel computing solutions. Both are Intel strengths.
The transfer costs of adopting new technology and infrastructure can be very high at all levels, from the power supply in the data center to coding language for the developer, so starting with the current environment and building upon it organically avoids those transfer costs. This is great news for organizations because it means that they can run AI workloads on their existing data center architecture without acquiring and learning proprietary systems.
Artificial Intelligence combined with data analytics is key for organizations to meet goals faster, scale to meet business and customer needs, and free up resources to contribute to other value-added areas.
Intel is making large investments in this area but rather than talk about AI methods and models, products, scale and design capabilities, the most important parts of the conversation to me are:
- Finding Value in Data - helping organizations unlock new possibilities for their data - with our comprehensive stack of products designed for AI.
- Maximizing the Impact of AI – investing in applications to solve real-world problems and engaging with government, business, and society’s thought leaders to prepare and enable this transformation.
Artificial intelligence uses abound in Health, Manufacturing, Energy, Oil and Gas, Government and Academia. AI is being used outside fundamental research for precision medicine, drug discovery, oil and gas exploration, and operations, predictive maintenance, marketing, regulatory compliance, banking and trading and it is hard to find a place where this level of automation is not being done. The key message is change now, before you have to, and develop an AI strategy.
There are key ingredients needed to get started which include a clearly defined business case and usage and the data to train an artificial intelligence model. This first part of this process is the most time consuming and challenging.
Innovation science techniques are helpful in identifying AI usage models. Innovation as a discipline is not about creativity. It is about tools and methods, measurement and process management. It is possible to use the evolution of innovations to plan and predict the next product innovation.
One of the evolution trends is increasing automation. This follows a predictable pattern moving from a tool, to adding over the product life, a transmission, energy source, control and finally intelligence. Another predictable evolution is moving from a single or mono system to a bi-system, poly-system and then combined system.
|Sewing Needle||Treadle Sewing Machine||Electric Sewing Machine||Pre-programmed Stitch Menu||Robotic Industrial Sewing Machines combined with cutting and pressing.||From fashion drawing to automatic pattern with cutting and sewing codes|
Another powerful trend is the evolution from an immobile system, adding increasing flexibility to become elastic then moving to a liquid, gas, and finally a field.
|Single lens eyeglasses||Bi-focals||Tri-focals - “Progressive lenses”||Adjustable lenses with membrane and fluid||LED glasses with Bluetooth||Autonomous Navigation|
|Single Door||Bi-fold Door||Accordion Door||Roll up Door||Air Curtain||Infra-red door alarm or lock||Facial recognition Access|
|Immobile||Joints||Many Joints||Elastic||Liquid, Gas||Field||AI|
Consider that AI is at the top end of each of these evolutionary trends as an intelligent, combined, “field” based solution.