The Road to Productivity Revolution: An AI-Enabled Journey

Artificial Intelligence is one of the hottest topics of last year, with revolutions such as ChatGPT taking over the headlines. From something as simple as autocorrect or autocomplete to complex applications like autonomous vehicles, AI has seen tremendous growth and widespread application over the last few years.

As per the 2023 Gartner Emerging Technologies and Trends Impact Radar, technologies have been divided into 4 categories - Critical enablers, Productivity Revolution, Smart World, and Transparency & Privacy.

Productivity is one of the major concerns across industries today. Taking the example of our firm, our strategic integration of both manufacturing facilities is followed by the need for a productivity revolution, to ensure that we utilize our production capacity efficiently.

Considering the importance of productivity in our daily life, we shall explore two AI trends that have the capability to revolutionize productivity - Edge AI and Generative AI.

Edge AI

Edge AI an application of AI that aims to reduce restrictions due to speed and enables devices to react quickly without the need to access cloud.  To understand edge AI better, let us first take a look at edge computing.

Cloud computing is a widely employed technology, which delivers services over the internet. In contrast, edge computing systems operate on local devices such as IoT or a dedicated edge computing server and are not connected to a cloud. 

So, why are we reducing our dependence on cloud? We have always been on the quest for faster, more accurate results. With edge computing, we are not affected by parameters such as latency and bandwidth, or even downtime, all of which are critical when we depend on cloud.

Edge AI is a combination of Edge Computing and Artificial Intelligence, where AI algorithms are run on edge computing devices. Operations occur locally and data can be sent to the cloud once internet connection is established.

Now, one might think, in a time when internet is so widely available, what is the relevance of Edge AI?

There are several reasons why we prefer Edge AI, and it is not all about internet availability. For example, computing on the edge ensures that only relevant, processed data is sent to cloud. Since volume of data is reduced, the time isreduced considerably.

Edge AI is also useful in areas with frequent network disruptions, and situations that require quick decision-making. An example is that of self-driving cars that need to detect obstacles on their path and make the right decisions. Here, we cannot afford limitations like bandwidth, latency, and disruptions.

 A study was conducted to analyse the performance and scalability of edge computing, using it to read a QR code and send data to the edge server, which took 16.22 ms. The same process would take 80 ms if the data was sent to the cloud, giving us a clear idea on how fast edge operations are.

Now that we have an idea of Edge AI, let’s see how this can be implemented in manufacturing operations:

Applications of Edge AI:

1. Quality Control

The cost of poor quality can impact profitability significantly. But to ensure the best quality, there is huge requirement for qualified personnel and regular inspections. Even so, we, as humans are prone to errors. Recently, several manufacturing firms are shifting towards computer vision. Computer vision is an application of edge AI, where we train the technology using images of defective and defect-free products. Based on this training, computer vision can inspect every sample available within a restricted manufacturing space and identify defects quickly and accurately.

One of the benefits of this application is the ready availability of data. All inspection data is automatically uploaded to system and can be easily accessed. Further, this data can be used for assessing defect patterns and statistics for better process control.

2. Predictive Maintenance

Preventive maintenance is used to avoid equipment failure and downtime. AI controllers available in the market utilize adaptive intelligence to observe and analyse patterns in machine operations, recognize variations, and use edge devices for predicting failures and prolonging equipment life. It allows manufacturers to avoid unplanned downtime by using edge AI to accurately assess material condition.

3. Factory Optimization

Moving towards a lean and agile environment, factory layout optimization is of utmost importance. Through Edge AI, manufacturers can identify the various movements that occur in the factory and understand how well their space is being utilized. The information is processed using edge AI and made available to supervisors, who can analyse to identify ideal movement of materials and people across shopfloor.

4. Safety

Safety is the most important part of our lives, be it personal or professional. Safety at the workplace is essential to both employees and employers. Sensors and video cameras that use AI-enabled video analytics can identify unsafe conditions and unsafe acts, and raise alerts. The low latency allows the technology to take quick, remedial actions in such conditions, such as raising an alert on the situation. The data can also be easily recorded as there is no need for manual entry.

Generative AI

Moving on, let us now explore one of the hottest technologies of last year, Generative AI.

It is a technology that allows us to generate a wide variety of output by analysing and learning patterns from existing data. Output can be images, videos, text and even music. Needless to say, GenAI is going to drastically change the content creation space, and we shall see how.

One of the most popular foundation models that applies GenAI is ChatGPT. There is no introduction required for this text-based AI tech that has taken over the internet ever since its launch in November 2022. Within 5 days of its launch, ChatGPT had more than 1 million users, which went up to 100 million users in just 2 months of its launch.

So, what makes ChatGPT so different from regular chatbots or search engines? ChatGPT stands out with its versatility. While it can take a conversational tone, it can also frame emails, business pitches, complex essays, and write computer programs. They’re different from search engines, which can give us information from their huge database but cannot generate new/unique content.

Now, Generative AI has not only influenced the written content, but it has also seen massive developments in the realm of visual content.

DALL-E is a model developed by OpenAI that can be used to create images and variations of existing paintings through text input. A similar technology is generative design, which uses AI algorithms to evaluate specific input parameters such as weight, material, design, and fabrication method to generate multiple permutations of designs, which can be then be analysed and selected.

Using generative design does not have to be restricted to just iterations of a design with same input. The prototypes generated can be fixed with sensors to get real-time data on their performance. This data can be fed back into the system and the software redesigns with this newly added input to optimize the design. A recent example is the collaboration between GM and Autodesk to produce a seat bracket that is 20% stronger and 40% lighter than the existing component.

Now, let us see how generative AI can contribute to our journey of productivity revolution.

Applications of Generative AI:

1. Customer Service

Customer service is one of the most important areas for any firm, and service for prospective customers is equally important. The most common example of AI used for customer service is chatbots. Most websites use chatbots to answer customer queries, however, they are limited to a small set of questions. To enhance these areas in our business, we can train AI GPTs on datasets specific to our business.

Customers can have conversations with our customized chatbots that have been trained on a repository of knowledge on our product groups. Over a course of time, we can add features that can support order enquiries and help give quotations to customers.

2. Chatbots for Internal Knowledge Management Systems

Chatbots using AI can also support employees as an internal database of information. They can be trained with past order records, customer and supplier details, and information on different specifications. Employees can then use the chatbot to -

• Clear their queries on products and specifications.
• Easily get information on specific clauses or verify whether project specifications are in line with standards.
• Get consolidated and relevant information on past orders, based on parameters specified by the enquirer.

With the ability of AI to learn and adapt, there is a huge scope for growth and application in the future. We can use them for framing emails, marketing copies, and even graphic designing for social media posts.

“Investing in Tomorrow’s Technology Today is More Critical than Ever.” - Bill Gates

As we move forward, we need to invest in technology that is going to make our operations smoother and more efficient. With widespread availability of AI that presents both ease of use, access, and potential to improve processes, we can shift our focus to other areas of improvement and simplify workload. Together, we need to drive the productivity revolution and set an example to the world.

Author Name: Sreelakshmy Sivadas | sreelakshmy.s@lntvalves.com