The ongoing Fourth Industrial Revolution is driven by big data that are aggregated from connections of all things (physical, virtual, biological) and processed by AI (artificial intelligence) algorithms on hardware and software. Hence, business logic and machine models are optimized and support the intelligent actions of humans or objects. Furthermore, Edge Computing enables data to be processed at the “edge” near/at where it happens to automate intelligent actions, like how an octopus‘ hands can think and react by themselves.
On the factory floors, various sensors collect data and robot arms execute functions accordingly. If AI algorithms predict failure might happen in a piece of equipment, a maintenance action or an alarm will be triggered. This is called predictive maintenance, or preventive maintenance, a part of the overall asset management. The factory has the self-awareness and decision-making abilities to orchestrate the process flow and utilize resources, and Digital Twins allow engineers to do experiments to find better plans in real-time.
Digital twins are digital replicates of the physical assets and systems, used for the entire design-execute-change-decommission lifecycle in real-time. Tesla has a digital twin for every car manufactured. Every day, thousands of miles of data from the cars, are fed into the simulation models back in the factory. Digital Twins have evolved beyond assets to include entire organizations covering people, processes and behaviors.
In the domains of consumer or corporate businesses, products or services continuously collect data from interactions with users and build up user pictures so that they can improve customer experience and recommend the next items customers will buy. The business models might transform from one-time product sales to service providers. Thus, revenue streams are improved. Digital supply chain management and participation from customers enable the production of personalized products at scale. Digital Twin models integrating manufacturing and logistics enable agile responses for faster delivery and reduced inventory.
Leveraging big data, insurance companies can automate the review process for insurance claims with better judgment. Banks can automatically detect fraudulent credit card transactions and money laundering, as well as decide loan credit lines or approve loans for borrowers. AI can identify diseases from medical images with better accuracy than humans can. Drones can monitor farms and treat problems with precision.
Robotics Process Automation (RPA) can track office tasks and automate repetitive processes across software applications. It significantly enhances productivity and decreases errors. Employees can spend more time working on tasks of higher value or may also be replaced. RPA can look over the shoulders of employees and identify opportunities for improvement. It can also handle unstructured data from diversified systems and automate data analysis tasks. According to Deloitte’s report, the average cost reduction of every process is 30%. RPA is just a small tool in the enterprise automation toolbox.
For most people, program trading in the financial domain isn’t new. Now, high-frequency trading players use FPGA AI chips to process market data and noises without human interventions in order to gain an advantage in faster response (within hundreds of nanoseconds) to market moves. And, Goldman Sachs deploys AI technology that can digest a huge amount of economic news and reports and summarize critical factors for stock prices. This can replace quite a few analysts.
IT security engineers can automate the monitoring of cybersecurity risks with machines. This not only improves productivity, but the continuous data stream can also enable dynamic reactions and constantly update the model for optimization.
The largest oil & gas enterprise Woodside Energy in Australia feeds its accumulated knowledge and information from the past 30 years (about 600,000 pages of documents) to an AI engine and builds a smart assistant to support employees with the right knowledge at the right time when needed. The assistant also helps new hires get up to speed faster.
The Google Brain team uses Reinforcement Learning to develop AI chips and layout place-and-route. The AI method can use trial-and-error to optimize the chip design and accomplish an even better chip performance within hours than that of a senior engineer who would need several weeks to finish.
The same applies to the Machine Learning process itself, Google CEO Sundar Pichai wrote that, “designing neural nets is extremely time-intensive, and requires an expertise that limits its use to a smaller community of scientists and engineers. That’s why we’ve created an approach called AutoML, showing that it’s possible for neural nets to design neural nets.” What he refers to is called “neural architecture search”, it allows us to discover architectures far more complicated than what humans may think to try, and these architectures can be optimized for particular goals. Now Automated Machine Learning products have built upon the initial foundation and added a lot more, they automate many stages of Machine Learning, from data preparation, feature engineering, model selection, training, hyperparameters optimization, model performance monitoring after deployment, and optimization. In short, AI can help build AI models and operations, that enable citizen data scientists or business analysts to use ML to solve their problems, and also save data scientists a lot of time.
The development process for new drugs relies heavily on trial-and-error. This is something Machine Learning is very good at. Many steps in the scientific process of labs can be done by machines. The standardized scientific method, including forming a hypothesis, can be automated. R&D management for innovations can use the help of AI as well.
Of course, AI can benefit many aspects of HR and talent development, including recruitment, assessment of capability and personality, competency modeling and learning path recommendation, matching employees and positions and projects, knowledge management and search, leadership training, collaboration analytics, predicting the risk of employee behavior compliance. Corporations can keep track of emerging industry trends from job market data. They can also use AI to help decide key performance indicators (KPI) or objectives and key results (OKR).
Talent strategy is crucial for supporting corporate strategy and competitiveness. Nowadays, CEO and HR managers need to re-examine their talent strategies and re-imagine how AI, IoT, Edge Computing, automation, and Augmented Reality will change their workforce. They need to examine their goals and metrics and rethink work models and processes to gain the most out of collaborations between humans and machines instead of just replacing humans with machines. Tesla changed its R&D methodology to crowdsource all drivers to teach autonomous cars how to drive. This kind of AI strategy is what decision-makers need to set up before anything else.
Not only will AI and automation reduce the cost of the workforce by replacing humans, but they can also create new jobs. It can help companies upgrade or create new business and operation models. But organizations need to handle change management. The overall ecosystem inside and outside an enterprise is complex. What strategy is needed for the transformation? Employees need to equip new skills to work with AI and robots in order to achieve this transformation. Education institutions can’t solve the issue of skill gaps, so big enterprises are reskilling their own employees at scale. How can they do so? What should be taught? And how can AI be utilized for efficient training?
Harvesting the power of digital economics doesn’t just aim to use fancy technology to enhance specific abilities for siloed divisions in a company. Organizations can integrate technology creatively to serve their mission. Many companies said AI is important to their competitiveness but don’t have a strategic data policy. Or, they collect as much data as possible but don’t know what to do with them. AI is like electricity, it’s a tool, deciding the goal is the initial step. First, decision-makers need to understand how AI works and its capabilities and limitations so they can rethink the whole operation and workforce strategy (AI + Humans) with an AI mindset.
This is a big topic, and the best practices will be evolving for years to come. To facilitate the exchange and collaboration between enterprises and innovative startups/scale-ups. We and our partner Qmarkets have set up a free open innovation platform to accept the submissions of challenges, innovations, ideas, and collaboration invitations.