Metamorphic Transformation with Enterprisewide Artificial Intelligence

 

 


Metamorphic Transformation with Enterprisewide Artificial Intelligence

KYield Genesis

Most CEOs and boards today realize that in order to remain competitive they must apply artificial intelligence across their companies, but according to a recent McKinsey survey, only 21 percent report embedding A.I. into multiple business units or functions. Clearly, the piecemeal approach to A.I. is not serving the strategic or operational needs of business or government.[i]

Responses from the McKinsey’s survey highlight twelve barriers to A.I. adoption:

  1. Lack of clear A.I. strategy
  2. Lack of A.I. talent with appropriate skills
  3. Functional silos constrain end-to-end A.I. applications
  4. Lack of leader’s ownership & commitment
  5. Lack of infrastructure to support A.I.
  6. Lack of available data
  7. Uncertainty or low expectations for ROI
  8. Under resourcing for A.I. in line organization
  9. Limited usefulness of data
  10. Personal judgment overrides A.I.-based decision making
  11. Limited relevance of insights from A.I.
  12. Lack of changes to frontline processes after A.I.’s adoption

In addition to the barriers found by McKinsey, our direct engagement with hundreds of large organizations over the last decade found cultural resistance due to fear of displacement from A.I., which exacerbates all other barriers to A.I. adoption.[ii]  While fear of A.I. has been declining over the past few years, the combination of barriers to enterprise-wide A.I. systems represents existential risk to companies and nations. However, even the strongest leadership will fail attempting the physically impossible. Success in A.I. is largely dependent upon system design.

“Even the strongest leaders will fail attempting the physically impossible”

These barriers are actually interrelated systemic problems that require system solutions in order to achieve the bulk of A.I.-enhanced value. It is possible to overcome a highly complex and dynamic combination of physical, cultural and operational challenges simultaneously so that organizations can immediately benefit from A.I. value across the enterprise.

Pre-optimized architecture

When the A.I. revolution unexpectedly arrived with the 2012 ImageNet victory, no organization was pre-optimized for machine learning (ML) functionality, much less end-to-end enterprisewide A.I. systems. Exponential growth has since occurred in many areas. Deep learning (DL) has been applied to speech recognition among many other tasks, we’ve observed widespread use of deep neural networks (DNN), artificial neural networks (ANN), and reinforcement learning (RL) is evolving, all of which are reliant on data integrity and sophisticated optimization, hence the foundational A.I. systems challenge facing all organizations. In addition, commoditized infrastructure as a service available to anyone worldwide provides no competitive advantage.

Although most large enterprises have integrated a logical framework tailored to their needs, even the most robust enterprise architecture does not necessarily lead to a competitive enterprise. Indeed, less than 30 percent of traditional digital transformations succeed.[iii]  An entire new generation of purpose designed and developed systems are necessary to realize the benefits of A.I. across organizations, networks and broader economy.

“Companies must have end-to-end solutions to win in A.I.”

To no surprise, companies that have made the most progress in digital transformation have also led in A.I. adoption, which is the competitive driver for the logical conclusion: “companies must have end-to-end solutions to win in A.I.”.[iv]  This is identical to the epiphany I experienced when conceiving the KYield OS in our KS lab more than two decades ago.

With few exceptions, adoption leaders in A.I. systems are the small group of digital natives that were designed end-to-end from inception, which also represent the greatest risk as well as opportunity for incumbents. A recent Deloitte survey showed a remarkable change in perception of risk from digital disruption. In 2017, only 6% of executives considered digital disruption as a top external risk factor to their business, compared to 61% in the new survey consider it a top external risk factor.[v]  This leads to the logical question: Is fear of digital disruption preventing organizations from making wise decisions?

“Is fear of digital disruption preventing organizations from making wise decisions?”

A robust semantic enterprise is necessary but insufficient to enable an A.I.-driven enterprise.[vi]  Although a few companies invested billions of dollars developing their own custom A.I. systems, they started at a different place than most organizations and their needs are quite different. Large-scale custom efforts are also increasingly redundant. We found this grand challenge solvable, though it required many years of deep work.

KYield’s project Genesis: Rapid metamorphosis

By necessity, we began our voyage more than two decades ago focused on the needs of networked organizations and individuals, working backwards in overcoming barriers to those needs based on what was physically possible as components improved. After the basic architecture was in place for continuous yield management of knowledge, we were then free to focus on how to optimize the many processes involved in managing ‘an end-to-end’ system, which resulted in a significant body of proprietary know-how, trade secrets, and additional patents planned. Security would require a new type of encryption, scalability would require conversion from natural language to a more efficient language, improved compression would be necessary, and entropy would need to be mitigated. Once the processes were understood and minimal viability demonstrated, only then could we focus on bringing it all together to overcome the adoption problem.

Over the last decade we have been in regular discussions with large organizations about various options in organic adoption of the KYield OS. It became apparent over time that very few organizations, if any, would be able to make it to a fully operational A.I. system across the organization in a piecemeal organic process. Even if the engineering talent were sufficiently skilled to overcome the physical challenges, the economic, cultural and organizational challenges were simply too overwhelming for the super majority to overcome. Multiple senior executives in the world’s largest organizations admitted as much in our discussions.

Project Genesis was formed when it became apparent that customers would require a much more refined, nearly instantaneous approach for enterprise-wide adoption. The KYield OS pre-structures the modular network to optimize ML and other A.I. functions, such as governance, security, prevention and enhanced productivity, and then populates the data automatically across distributed organizations and networks.

Top ten goals of KYield Genesis:

  1. Pre-install, train and certify administrators on the CKO Engine to provide governance over the entire distributed system.
  2. Allow pre-approved individuals the ability to download their individual modules and immediately begin use with no training required.
  3. Use the learning functionality in the individual module to accelerate transformation with tailored training for each individual’s field and A.I. systems.
  4. Engineer so that the KYield OS begins to rapidly improve from day one of install.
  5. Scale rapidly across the organization so that benefits can be quickly realized and the continuous improvement loop can improve indefinitely.
  6. Provide customers control and ownership of data at all times.
  7. Provide the most effective security, prevention and productivity enhancement possible.
  8. Optimize the efficiencies of commoditized components while providing customers with a powerful competitive advantage automatically tailored to the needs of each entity.
  9. Make as transparent as possible to earn and deserve trust.
  10. Ensure that we over deliver on expectations, providing customers with the most superior entire experience available.

Organizations ready to lead with enterprisewide A.I. and enjoy a strong competitive advantage can be up and running with the KYield OS in about six months with the Genesis architecture. Once the leaders adopt and system is fully operational, the benefits can be extended to partners within a few days of approvals.

Mark Montgomery is founder & CEO of KYield, Inc. He developed the theorem ‘yield management of knowledge’ and was the inventor of the patented KYield OS.

[i] November 2018 survey, McKinsey & Company: A.I. adoption advances, but foundational barriers remain

https://www.mckinsey.com/featured-insights/artificial-intelligence/ai-adoption-advances-but-foundational-barriers-remain

[ii] Fear of Artificial Intelligence vs. the ethics and art of creative destruction (KYield)

https://www.wired.com/insights/2014/06/fear-artificial-intelligence-vs-ethics-art-creative-destruction/

[iii] October 2018 survey, McKinsey & Company: Unlocking success in digital transformations

https://www.mckinsey.com/business-functions/organization/our-insights/unlocking-success-in-digital-transformations

[iv] Artificial intelligence: The time to act is now

https://www.mckinsey.com/industries/advanced-electronics/our-insights/artificial-intelligence-the-time-to-act-is-now

[v] Businesses’ adoption of A.I. is expected to surge (Deloitte)

https://www.wsj.com/articles/businesses-adoption-of-ai-is-expected-to-surge-11556184602

[vi] Semantic Web and Semantic Technology Trends in 2019

https://www.dataversity.net/semantic-web-semantic-technology-trends-2019/#

Categories: Tags: , , ,

Leave a Reply