Davenport 1

pdf

School

New York University *

*We aren’t endorsed by this school

Course

MISC

Subject

Information Systems

Date

Oct 30, 2023

Type

pdf

Pages

Uploaded by PrivateKoalaMaster761

Artificial Intelligence for the Real World This article talks about the moon shot project at the beginning whose objective is to diagnose and recommend a line of treatment for certain forms of cancer using IBM’s Watson cognitive system. But in 2017 it was put on hold since the cost price increased to $62 million. At the same time, the IT group of the cancer center used cognitive technology by providing recommendation to patients' family and addressing IT staff problems. The result of this small project is much more promising, despite of poor performance of the moon-shot project MD Anderson is still committed to using cognitive technology. Another survey was conducted among 250 executives who are familiar with the cognitive technology concept. Out of 250, 3 quarters believe that the company will be transformed within 3 if they use cognitive technologies. However, in another study of 152 projects, many companies believe that highly ambitious cognitive projects are less likely to succeed as compared to simple ones which provide better results. 3 types of AI can support business requirements- Process Automation : Out of the 152 projects, common projects were automation of tasks using RPA- Robotics process automation. Bots are deployed into the system and perform tasks like data transfer, replacing lost credit or ATM cards and updating the system records, reconciling failures to charge for services across billing systems, and reading documents to provide information using natural language processing. NASA launched 4 RPA pilots, four projects worked well. In the HR application, 86% of transactions were completed without human intervention and are being rolled out across the organization. Now NASA is implementing more RPA bots with higher intelligence levels. It’s a common assumption that with the implementation of RPA people will lose their jobs but 71 RPA projects were reviewed and only a few projects led to reductions in headcount. Cognitive Insight : Another type of project is using algorithms to detect patterns in vast volumes of data and interpret their meaning. For example, studying purchasing choices of customers, identifying credit frauds, analyzing warranty data to identify safety or quality problems in automobiles, automate personalized targeting of digital ads. Cognitive insights provided by machine learning are usually much more data-intensive and detailed. Machine learning can classify data that is likely to be associated with some similarity. GE is using this technology to integrate supplier data and saved $80 million in its first year by eliminating redundancies and negotiating contracts that were previously managed at the business unit level. Also, Deloitte’s audit practice is using this concept to extract contract terms which helps to audit a lot of documents without human auditors. Cognitive insight applications are used to improve job performance only machines can do so it’s not a threat to human jobs. Cognitive Engagement : Projects that engage customers and employees fall under this category. This includes intelligent agents that offer 24/7 customer service, internal sites for answering employee questions on various firm services, and health treatment recommendation systems. Vanguard is piloting a virtual agent that helps customers to get answers to frequently asked questions. SEBank in Sweden and medical tech giant in the US are using the lifelike intelligent agent to serve the internal IT help desk. But these bots are not completely reliable, Facebook messenger chatbots are not so efficient and couldn’t answer 70% of requests. Thus cognitive engagement is not a threat to human jobs instead it will reduce the workload and help the companies to transfer a few people to more challenging roles by offloading this work. Based on research, there’s a four -step framework for integrating AI technologies that can help companies achieve their goals-

1. Understanding the technologies - Companies should perform a thorough analysis of the benefits and limitations of technologies. RPA is transparent in the process but it's not capable of improving whereas deep learning is capable of improving but th ere’s no transparency in modeling. Several organizations wasted time and money while pursuing the wrong technology. Companies should plan to leverage the capabilities of data scientists who have the statistical skills to use these technologies appropriately. 2. Creating a portfolio of the project : For launching an AI program evaluate needs and capabilities and develop portfolios or projects. Companies should conduct an assessment in 3 areas- a. Identifying opportunities: Select business areas which could benefit most from cognitive applications. In some cases, the lack of cognitive insights is caused by a bottleneck of information flow; knowledge exists, but it is not optimally distributed in the organization. Another reason is that it is difficult to scale the challenges, knowledge exists but the process of using it takes too long or is expensive to scale. In the pharma industry, Pfizer is using IBM’s Watson to scale these problems. With the help of Watson, researchers are finding hidden patterns that will help speed up the identification of new drugs. Also, It helps companies collect and analyze more data. Cisco Systems and IBM are using this to create tens of thousands of “propensity models” to determine customer buying patterns. b. Selecting the technology: To consider the AI tools that are being used are truly up to the task. Chatbots might not be that effective because it is complex to understand human problems but RPA can help to streamline the simple process but it might slow down a complex production system. D eep learning’s visual recognition is unable to make sense of a complex visual field. In the future, cognitive technologies will transform how business is performed but it's wiser to take incremental steps with currently available technology. 3. Launching Pilots : since we are not sure of AI capabilities currently companies should create pilot projects for cognitive applications before rolling them out in the entire organization. It is a good idea to create a cognitive center of excellence, this approach will help them to build technical skills and add capabilities. Pfizer has more than 60 projects across the company running in pilot mode. Business Process redesign- with the implementation of cognitive technology there is a need to redesign the workflows. These redesigning principles help to understand the customer or end-user needs. 4. Scaling Up : Cognitive technologies support individual tasks rather than the entire process, so scale-up requires integration with existing systems and processes. While scaling up companies should first think about whether integration is even feasible. Certain change management challenges may come up while scaling up. Firms should also focus on improving productivity. AI investment should on the elimination of the outsourcing part. Cognitive technologies act as a catalyst for making data-intensive technologies succeed which includes IoT, mobile, and multichannel consumer technologies. The biggest fear about cognitive technologies is people will be laid-off. Some jobs will be taken over by smart machines but certain tasks only humans can do. These systems perform tasks, not entire jobs. Most managers are committed to an augmentation strategy -integrating human and machine work in one of the surveys a very handful of executives indicated that they considered reducing headcount as a primary benefit of AI.

Your preview ends here

Eager to read complete document? Join bartleby learn and gain access to the full version