NTT Data: How AI at the Edge Can Unlock ROI for Industries

AI isn’t just about ChatGPT, Perplexity and any of the other popular large language models (LLMs) on the market. 

AI at the edge will enable exceptional productivity as AI capabilities are embedded in products we use every day. However, to build and manufacture these products, edge AI has a vital role to play as well.

Since ChatGPT came onto the scene in 2022, AI has drastically reshaped business operations as enterprises embrace its benefits, including greater efficiency and faster decision making. 

AI is a game changer in many ways. But in industries that are considered to be blue-collar, such as on the factory floor, in mining operations, airport baggage handling areas, healthcare settings and supply chains, businesses are still working to determine the best way to implement and use the technology in order to gain a competitive advantage. 

AI in blue collar industries

In these complex environments, AI is most useful when combined with edge computing solutions that gather and process data at the edge. 

This allows operational technology (OT) data to be combined with information technology (IT) data in a way that provides a robust return on investment through improved machinery and energy management. 

AI combined with edge computing also allows for the analysis of data in real time for actionable insights—and more quickly than in larger digital transformation projects.

For example, edge AI can assist in the predictive maintenance and management of parts and machinery by proactively alerting managers to potential issues, and recommending repairs before equipment breaks down. 

Additionally, these systems can be trained with AI-driven reasoning to monitor and notify if goods on a conveyor belt are being manufactured or assembled correctly.

How AI at the edge works

Manufacturers and other businesses will push over the coming year to unlock the value that comes with the latest IoT technologies, automating business processes to advance innovation, while also avoiding concerns such as latency and security. 

This is why analysts such as IDC estimate that global spending on edge computing will reach $378 billion in 2028. 

This also presents a perfect opportunity for edge AI services to correlate all the data produced by these IoT devices and integrate it with data from more traditional equipment, such as thermostats.

Doing this leads to better efficiency, productivity and faster decision-making.

Edge AI differentiates itself from GenAI by using smaller language models. These are specifically task-oriented, requiring fewer resources, and can operate independently from larger compute environments. 

An edge AI model may have only a few hundred parameters compared to LLMs with more than 50 billion parameters, which makes it easy to do real time processing at the device level. 

Security is also enhanced as the data is collected, processed and kept locally. The power of these smaller models brings edge AI to life, where it can have real impact with actionable data for quick response times. 

Another way to understand edge AI is to see it as ‘actionable AI’. It goes beyond just collecting and reporting data. Edge AI actively computes, learns and takes real-time action on data outcomes. 

Unlike traditional AI, which requires querying and processing on central servers (often introducing latency), edge AI operates locally. 

This enables the kinds of instant responses that are critical for applications like autonomous vehicles, industrial automation and real-time video analytics. This, in turn, saves resources, enhances the user experience and reduces operational costs. 

For instance, imagine in a 1M+ square feet factory floor setting, a company is struggling to manage energy consumption sustainably. As such, the environment must be a steady 72 degrees, leaving the operations manager overseeing over a hundred thermostats, each with its own app. 

Fluctuations are continually happening due to heat from machinery, worker movements, doors opening and closing and more, making temperature control a full-time job. 

With edge AI models, data from all these thermostats can be collected and managed automatically, while factoring in real-time changes. This frees the manager to focus on other tasks.

What are the future implications?

For businesses with ultra-fast, low-latency Private 5G networks already in place, adding edge AI will open the door to new applications and real-time data interactions. 

This arrangement facilitates the streaming of real-time data from OT devices, forming a data lake of operational information that factory floor workers can use to improve overall performance. 

Edge AI deployments – especially within industrial markets – will expand this year and beyond as businesses see the benefits in action. Rather than struggling with expensive, time-consuming Gen AI and digital transformation projects, an edge AI solution can be installed and operational within days, if not hours. 

The ability to process data in real-time on a factory floor is a game-changing technology breakthrough, and a significant competitive advantage. 

Edge AI removes the worries over security by keeping data on location where it is created and removes the latency concerns. CIOs and CTOs are just beginning to scratch the surface on the use cases of how to transform the real-time secure data to business benefits not yet exposed.

Explore the latest edition of AI Magazine and be part of the conversation at our global conference series, Tech & AI LIVE. 

Discover all our upcoming events and secure your tickets today.

AI Magazine is a BizClik brand