Slideshow Key steps to becoming an AI-enabled smart business

  • May 14 2018, 5:56am EDT
11 Images Total

What is a smart business?

A growing number of organizations are embracing artificial intelligence in the effort to improve efficiencies, reduce operational costs and increase profitability. Jack Norris, senior vice president of data and analytics for MapR Technologies, introduces the concept of an AI-enabled smart business and how organizations can tap into data, analytical and decision services to monitor the ‘pulse’ of business and succeed in those efforts.

BI alone isn’t enough

“Business intelligence is fairly mature, with BI tools accessing data warehouses and data marts in support of both strategic and tactical decision making across marketing, sales and finance,” Norris explains. “Yet, despite the maturity of BI in the enterprise, most insight produced today shows what has happened in the business over time. Most organizations need analytical systems to go deeper by preparing and analyzing any type of data, to deploy prescriptive analytics – and show what is happening in the business in real-time.”

Content Continues Below

Deeper Insight

"Not surprisingly, deeper insight requires new data, which means that new data sources of varying types are now in demand," Norris says. "This includes clickstream data, sensor data, text and semi-structured data like JSON and XML. All of this is being captured or planned with data volumes growing rapidly, and the rate at which data is being generated is also skyrocketing."

All Access Pass

"Many organizations have an equally pressing need for insights to be available on a more continuous basis and to everyone in the employee base (and beyond) working in everyday business processes to help them guide and optimize business operations," Norris says. "The vision today is therefore much more challenging. It is to become an AI-enabled smart business."

AI-enabled Smart Business

"An AI-enabled smart business can be defined as: 'Where data and analytics are used to guide people and applications so that they continuously know the best action to take and when to take it to dynamically (re-) optimize business operations, minimize risk, seize opportunity and improve customer engagement in order to maximize profitability,'" Norris explains. "Why do businesses need to become smart? The answer is simple – it is to survive, to drive new business opportunities much more dynamically, to continually engage customers, reduce risk, reduce costs and remain compliant."

Content Continues Below

Overcoming Challenges

"It should be possible to spot problems way ahead of time and weigh up the risks, costs, opportunities and actions needed to manage them," Norris says. "It is clear that to achieve this, point solutions are not enough. There needs to be a framework that can embrace existing investment while laying the foundation that manages data from the edge all the way to the data center. It should make it possible to encourage greater the investment in data and analytics to produce new and deep insight on a continuous basis so that a business becomes the more informed, pro-active, responsive and competitive when making decisions to improve its overall performance."

Building a Smart Business

"Building a smart business should be done incrementally to build a business strategy that aligns descriptive, predictive and prescriptive analytical and AI workflows using a common extensible analytical framework on top of an extensible common ‘data fabric,’" Norris explains. "The data fabric should stretch across streaming data in motion and data at rest. The former includes streaming data at the edge, on the cloud and on-premises. The latter includes data in cloud and on-premises SQL and NoSQL data stores."

Getting Started

"To facilitate organized incremental building of a smart business, the data fabric should support the concept of data and analytics projects and the association of these projects with specific business strategy objectives," Norris says. "It should be possible to define business objectives and then link one or more projects to a specific business objective."

Content Continues Below

Strategic Objectives

"An example of a strategic business objective would be 'To reduce fraud,'" Norris explains. "It should be possible to define this objective and then allow the creation of multiple fraud reducing analytical projects with each project focused on building an analytical pipeline to analyze specific data assets in this business context and attach each project to the objective. With this approach continuous AI and machine learning models can operate to make it possible for a CDO and other C-Level executives to see the complete set of data and analytical results that have been created to help achieve a common strategic objective aimed at reducing fraud."

Data Ingestion

"The intention is to build a framework that extensible," Norris says. "Also note that the data fabric should extend into data on-premises and in the cloud and can stretch all the way to the edge. Data, analytical and decision services are pipelines (workflows) that can be published as services in an information catalog and that can be linked together using a common workflow capability potentially based on the Kafka API. This would allow micro-services based pipelines to be created and open up the potentially capability to create a Pulse ‘state machine’ for the business that dynamically responds to events and continuously optimizes the business to achieve specific business goals."

Ultimate Goal

"Smart business is similar to a person’s health and fitness goals," Norris concludes. "The objective is to improve and strengthen while also responding to problems as and when they arise. The ultimate goal is to become not just a smart business but a self-learning, self-optimizing smart business and to learn how to continuously improve."