Leveraging AI & Blockchain in Manufacturing- Driving business value using Data

#AI #Blockchain #digitaltransformation #NassconCoE

Jaspreet Bindra

ProMFG delves deep into AI, blockchain and data science with Jaspreet Bindra, Thought Leader, Advisor, Author on Digital Transformation, AI, Blockchain, Future of Work , he has worked as Chief Digital Officer of Mahindra Group, Regional Director at Microsoft, COO of Baazee.com and an Officer of the Tata Administrative Services

Mr Jaspreet Bindra is the Special keynote at NASSCOM CoE-IoT & AI expert session to bring to the fore the ways in which AI & Blockchain can be used to empower and optimize the Manufacturing sector.

In conversation with Saikumar. S of Pro MFG he shares practical advice for adopting an AI strategy, implementing AI proof of concepts, and going beyond the hype on AI & Blockchain to achieve real business outcomes for the manufacturing sector.

How can data science and AI Help address manufacturing challenges?

This is going to be a two-part answer. First, we must look at how manufacturers can get AI ready and thereafter, we will look at how manufacturers must get AI ready by organising their data and second, we will look at how a proper AI set up, rooted in thorough data science can actually make a difference to organisations.

Data is exactly the place to start. Many organisations want to skip the data step and jump right into AI but this is counterproductive. Moreover, organisations already have the data necessary for AI to work wonders for their businesses. However, the problem with data is that it is disorganised and decentralised for most organisations. Additionally, it might not be in a format that can be utilised by for AI insights.

I would recommend using the CUPP model at the very highest level, and rolling it out across the manufacturing organisation, for the manufacturer to become AI-ready
● Collection of the right data internally and externally
● Unification of data, or centralising and tagging it appropriately is essential
● Processing of the data or building the right algorithms, building the right data models, training on the data which has been collected to derive the databr
● Presentation of data for the stakeholder in your mix that is going to make decisions based on the data - this stakeholder could be you, your customers or any other stakeholder, like your investors or your suppliers.

With all this in place, you are ready to make good use of the buffet of technologies that come under the AI umbrella. Machine learning, deep learning, robotics, computer vision, natural language understanding are only some of the technologies in the buffet of AI tools.

Here are some use-cases:

1. The most commonly and efficiently used one among this long list is machine learning or machine learning algorithms which get better and better and therefore keep delivering better insights. Machine learning is already being used for predictive maintenance. Machinery represents huge investment in manufacturing and preventive maintenance can help you pre-empt machinery breakdown and damage occurring from a lack of awareness that something needed fixing.

2. Another area is inventory management and buying inventory at the right time. Linked to this is supply chain management and effectiveness.

3. Computer vision is yet another use. For example: one of the largest steel manufacturers in India uses tons and tons of coal. Previously, this coal had to be physically weighed but today all one needs is a drone flying over the site. Computer vision helps to map and scan the coal to get a fairly accurate idea of the weight and volume of the material.

4. Robotics of course needs no introduction or explanation, and it is cheap right now. Robots are already utilized and can be further utilized for jobs like load lifting etc. Especially relevant at this point is “Co-Botics” or humans and robots working together which is the most effective way to go.

What is the role of blockchain in manufacturing today?

I'm going to be very upfront here: Most of the movement and potential - especially in the manufacturing sector - is in the areas of AI and IOT. The role of blockchain is still limited; still a hammer trying to find its nail. That said, lockdowns activated a series of supply chain problems that at least opened the doors for blockchain to gain some level of recognition and adoption.

For example, with all the supply chain issues that arose out of lockdowns, blockchain was able to give some manufacturers and suppliers a level of transparency and shareability in the whole process. We’re also witnessing some traction for the provenance and traceability that blockchain is able to usher into supply chain management.

Blockchain fosters decentralisation, authentication, and traceability and that is what makes it relevant amidst the lockdown-linked challenges. It enables trust by activating traceability. However, blockchain is still a very nascent concept and most of the movement is being seen at a POC stage. It is still to be rolled for high scale application.

Avoid chasing blockchain (for implementation) at this time but do try to wrap your head around the concept. It is indeed a complicated concept. As a result, I would recommend that manufacturers adopt a gradual approach - the issues at hand are better supply chain management, cost reduction, faster decision making and increased safety, which can be achieved by AI in all its different forms.

Take the first step with AI and keep blockchain for a slightly later stage, but definitely do try to understand the concept. It is fairly complicated (to the point that I take 60-to-90-minute lectures explaining blockchain) and must be fully understood before it is adopted.

How can we cultivate better digital transformation mindsets?

This is not a race to adopt big, new, popular concepts like AI and Blockchain, but about achieving goals by using these as tools. They are not the end but the means to the end. Your business goals are the end towards which these means must be employed. Both customers/manufacturers and suppliers/tech brands tend to get carried away with putting AI and blockchain in place without defining goals.

But these tools are supposed to make organisations more agile. They also come with business models that allow for minimal investment as well as trial and error. I’m referring to the fact that cloud models allow us to adopt AI and blockchain services on rent as well as on a trial basis. As a result, you can experiment with a handful of different solutions and eventually roll out or scale up the one that works.

The main mindset change that needs to be effected, however, is linked to the need for perfection at the first attempt. Manufacturers have had generations of imbibing this mindset, “get it right the first time; there are no second chances”. However, with machine learning or AI in general, it’s very nature is such that you might fail in the first few attempts, because it gets better and better each time.

Start with your business problem and use AI to solve it. Perhaps you want to reduce costs or increase safety or improve efficiency or increase throughput. Zero in on measurable goals. Then you look for the AI tools that can bridge the gap. Most importantly, adopt an agile organisation and an experimental mindset. I have to be blunt and say that manufacturers need to accept that these solutions will not - in most cases - click into place the first time. However it won’t cost you much. Moreover, when it does work out, it will drastically improve your business.

How long does it take for AI projects to prove their worth?

It takes at least 12 months - after the data is ready - for the technology to prove its worth. And you need to be patient in the same way that you would with a brick-and-mortar project. For example, when you are building a manufacturing plant, you are extremely patient. It takes years, but the patience linked to such projects and all the delays involved, is ingrained in most manufacturers. However, when it comes to some new technology, there is a tendency towards impatience.

A senior auto executive from one of the biggest auto manufacturers, who I work with frequently gave me some interesting insights. He said that where a manufacturer might spend Rs 5000 to Rs 8000 crores on an auto plant for a new car model that may or may not work on the market, people are wary of even spending a fraction of that - Rs 100 crore - on new technology that can achieve the same level of benefits and profitability for the business.

How can we de-risk artificial intelligence implementation in manufacturing facilities be de-risked? Predictive maintenance is the lowest risk, highest (immediately provable) results area of AI that can be a gateway to AI manufacturers.

Another low-risk investment in AI could be in the form of forecasting. The type of forecasting might differ between companies but will certainly make a huge difference to all manner of manufacturers.

A third option is robotics - at whatever level the manufacturer can afford to implement.

Moreover, opt for these as a service on rent. Use free or even paid trials rather than putting in a large upfront cost. Managing the input cost is the best way to de-risk AI a in manufacturing and this is very easy because most AI is available on pay-as-you-go and rental models.

What is your favourite artificial intelligence project?

Well, a favourite could be the most successful from a business’ point of view or it could be something that you find incredibly exciting because of its novelty value and innovativeness.

As a businessperson, I am going to define favourite as the former. And my favourite from that point of view, is using data for forecasting and prediction. Hence I mentioned earlier, that forecasting is something which all types of manufacturers must adopt.

Did this article answer a ton of questions that you have had about AI and blockchain? Get further useful insights just like this and in-depth understanding of how these concepts work and how to implement them effectively while keeping costs at the minimum. Sign up for the NASSCOM CoE-IoT & AI expert session on 22nd Oct 2021.


Industry veterans and front runners of change will offer hands-on insights into the ways in which AI & Blockchain can be used to empower and optimize the manufacturing sector. You are sure to walk away with a thorough understanding of how to leverage AI and blockchain to make your manufacturing business more successful and profitable.