Use cases in IBM study showing how to focus AI projects for maximum value

New report on business value creation opportunities using Artificial intelligence (AI) has a data point that speaks volumes about the long way ahead: by the end of 2022, one in four large companies, or 25%, are expected to have passed the pilot stage to get their work on AI operational. The remaining 75% are experimenting with or considering AI projects.

While that number is a modest 25%, it is significantly up from the 9% that occupied their AI business as of 2020, and just 5% in 2018.

Data contained in a report entitledHow to create business value using artificial intelligenceFrom the IBM Institute for Business Value (IBV), based on discussions with more than 35 organizations with AI applications—and includes dozens of case studies on the use of AI across many industries. Industries. The report aims to debunk common myths surrounding AI and, in doing so, create evidence for its effective use, particularly in business operations where it can have a tangible impact.

One of the report’s core recommendations is that “C-suite and other leaders disagree with some of the myths surrounding it, like ‘AI shortcuts don’t work’…instead, they need to make decisions based on the reality of AI.”

Build on a proven foundation.

The IBV report cites the benefits of leveraging ready-made and pre-trained basic models that can provide a quick and cost-effective starting point with AI projects. One important factor that plays into this recommendation: companies have struggled to capitalize on the work they have done previously data scientists To train data sets, each business problem was addressed with a new artificial intelligence model.

Now, shortcuts are emerging. These are pre-trained modules similar to off-the-shelf software that can be installed and used quickly. This approach is designed to help organizations speed up their work without having to create entirely new data sets for each application; Instead, they can make use of the knowledge gathered from solving a single problem to help solve related problems.

Examples of these pre-trained forms include Google Forms Bert and OpenAI GPT-3. These types of models have three main advantages: they improve the economics of AI by amortizing costs across multiple use cases, they improve results by offering greater accuracy from larger data sets, and they offer new capabilities to take on board.

Abbreviations have also gained traction in commercial products, such as RapidMiner (recently Obtained by Altair), which includes the latest version of the product”Fully automated AI“The ability to create models based on commercial experience alone, targeting non-programmer and non-data-driven scientists. RapidMiner described this as part of its goal to democratize AI.

To illustrate the benefits of pre-made models, Boston Scientific spent $50,000 while benefiting open source Artificial intelligence models to address their goal of automation Examining stents in medical products, IBM reports. Through this measure and others, Boston Scientific was able to generate direct savings of $5 million while achieving higher accuracy in inspections than it had previously.

Focus beyond cost savings.

Examples abound of companies looking to apply AI to automate or perform jobs that become prohibitively expensive when humans perform them frequently and there is little or no human added value. Through this lens, AI can be accurately viewed as having the potential to reduce costs and this is of course a good result.

But the report notes that cost cutting is not the place for AI applications. Indeed, leading organizations focus on growing their business and achieving competitive differentiation through the use of artificial intelligence. As the authors expressed, those companies at the forefront of AI focus on customer-centric and top-line growth.

For example, IFFCO-Tokio, an India-based joint insurance company, has deployed a form of artificial intelligence to assess pictures of damaged cars, classify models, damaged parts, and type of damage after accidents.

AI was able to determine if the parts could be repaired or the replacement required. Moreover, it provides a cost estimate while retaining the human evaluator. While settlement costs decreased by 40%, customer “acceptance” improved from 30% to 65%, resulting in increased customer satisfaction, retention and acquisition.

Know that one size does not fit all.

How to create business value It also includes something of a cautionary note, invoking the myth that AI is a one-size-fits-all proposition, or that AI can and should be considered in almost any application or use case to achieve business outcomes. Not so, say the report’s authors.

Before embarking on an AI initiative, the first order of business is to determine whether enabling AI can serve a larger strategic initiative or address a core business problem. In short, there has to be something that fits the purpose of the AI ​​initiative.

Designed uses of AI can solve distinct business problems – across geographies and industries. But the case studies presented in the report show that the right approach often becomes clearer after selecting the right data set to solve the problem. Laying the right foundation is critical to success – Boston Scientific and IFFCO-Tokio help illustrate this point clearly.

For more exclusive coverage of innovative cloud companies, check out Cloud Wars Horizon over here:

Leave a Comment