For Sarajit Jha, who is leading the digital transformation at Tata Steel aimed at delivering an EBITDA impact of Rs 2,500 crore in four years, the task was cut out from the time he got his job title: chief – digital value acceleration.
From deploying smart sensors to using an image recognition software for spotting defects in steel plates, every project must deliver value and contribute to the EBITDA goal.
In August this year, we got an inside view of technology at Tata Sons with Gopichand Katragadda, the group CTO. Last week, I sat down with Jha to hear and understand how Tata Steel, among India’s oldest companies of its scale and scope, is relying on digitisation and software to solve its problems. Edited excerpts:
The job title.
This role came about in May 2016. But we had started talking about digital from May 2015. Coming back to your core question – first we said that digital had to be about business, you can’t have a digital strategy unless you are a tech company but you should have a business strategy which is supported and enabled by digital.
Second has to be about value. Steel as an industry is going through difficult phase today, facing the challenges of margin compression, talent and so on. Earlier, we had 7-8 years to exploit something. Today, in 21 months or in 18 months we need to look at something new. Unless you have the speed to exploit something you are thinking of from idea to impact, you would not be able to create adequate value and digital by its very definition allows you to do that. That’s how it (the job title) came about.
Our MD (T V Narendran) has very clearly defined that the talent of the future will have three things – they will be curious because things around you are moving so fast that if you miss something it may come back to haunt you; it has to be about collaboration none of us are wiser than all of us by definition; and then the ability to curate because there is so much of information and there is an entire data exhaust pouring stuff at you. Until you know what to look at and what not to and also what to look at when and how, you would not do a good job… you would be swamped by the overload.
The 2022 goal.
We will go very far. I mean we have two stated ambitions. Our first ambition is to be a leader in manufacturing in India and digitisation by 2022 and be a leader in the mining and metal segment in digitalisation in 2022. For us, digitalisation has become the way of life but we continue to be a steel company or a material company.
Tata Steel should have productivity 30% more than it should have on people. 80% of my processes should have predictability which allows me to be a leader in predictive asset maintenance in terms of predicting customer churn, in terms of predicting NPS. I should be a gold standard in safety with zero error. I should have changed the career paths of 80% of the people in Tata steel, made their careers much more enriching. And, finally, I should have 70% NPS and 80% of all our journey should be mapped on our digital space. So, if we are able to achieve this, then I would feel very happy.
Broadly, what we are saying is that we want to take NPS to 70 plus to our customers; average is about 53. What we are saying is that 80% of all the experiences should be measurable and should be best in class for all our customers. You may not be able to hit 100%, you don’t want to hit 100%, if you are trying to give somebody the best experience then you can only give another person a standard experience unless you have infinite resources. So it has to be a Pareto [optimum]. We are saying we are looking at building platforms, we are building exciting platforms in the marketing side, and rather we are experimenting with it.
That is one bucket, the other one is the entire people bucket where we are spending a huge amount of money training people, we are creating an analytic centre of excellence, next will be IoT centre of excellence. We are also looking at creating an innovation hub out of Jamshedpur of some sort. We are working with five start-ups, we are looking at working with another 30 startups in areas where deep tech matters. Start-ups are the best design on the clock speed of technology. Finally, we are a steel company, we are a process company, we have a huge cost basis that can be optimised.
Early pilots, algorithms at Tata Steel.
I can talk about the early successes because these were all largely pain management programmes. For example, we have a tubes business, it is about a Rs 2,200 crore business, neither very large nor a very small but not minuscule either. But it has all the segments – manufacturing, distribution, it has B2B, it has B2C, it has people who process the materials for us.
It is a very complex supply chain and what we did there was this: The way (steel) tubes were counted before dispatch was that the guy used to climb on top of the truck and put straws in each tube and then they would remove the straws and then count it. That is how they used to do until a month back. We have created a machine learning (solution) which can take a snapshot and count the number of tubes. Enabled by digital, what happened is due to computation power, due to superior algorithms and machine learning, image processing has gone to a new level.
But that was the smallest part of the problem. The biggest problem was to convince the transporter and our guys that this process is better. The second part was to change the dispatch process, the third part was to try and tell the transporter, ‘Look your turnaround time has come down by 7,8,10, 15 hours; at least 7-8% you are saving, and half of that logically belongs to us.’ That is just one area.
Another example we have is we have about 200-250 in-plant vehicles in Tata steel – 90 of them are leased, 150 of them are our own; they have only 20-30% utilisation. So, what we have done is launched this programme called the Digiwheels. What we had done is we have “Uber-ised” all these vehicles. We now have 70% of utilisation levels, safety has increased, we have only trained drivers, we are tracking the drivers, we are tracking the movements. We have literally created our own Ola.
Deployment of sensors.
Currently, we have 4 lakh active sensors in various shapes and use. It has taken 110 years to get to that 4 lakh. Back of the envelope calculations show we are using 20% of them and we are capturing and saving 20% of the data. If we have to meet our digitisation goals, we need to go to 48 lakh sensors on a conservative estimate by 2020. You can well imagine the challenges of connectivity, bandwidth, changing what is in between the ears of all of us, creating the business cases for all of it, having the right partners to make it all happen, looking at newer protocols…
AI in steel making.
AI is something that no one really understands. It is more hyped than reality. Are algorithms getting more sophisticated? Yes. But, because they are getting more sophisticated doesn’t mean that they are right. So what has to be done is that there is repetitive high-velocity work those algorithms will do much better than human beings. Wherever intuition, innovation, emotional connect are involved, it will be humans.
Our view of AI from absorption and from a management perception is fairly simple. We look at AI in three levels – we look at AI as what we call ‘Smart Assist’, we look at which is very simple work you are doing today and the ways mobile helps you to do better. Then, there is augmented AI where we see things that you couldn’t do earlier. Like today you can predict the time taken to move from X to Y based on all the crowdsourced data that the Google maps has.
And, then, there is the third level of AI which is autonomous AI. When people refer to AI, they are looking at autonomous AI; what we are saying is that we don’t have a vision of autonomous AI. Let us get people comfortable with Smart Assist, let us spur them on to be thinking about augmented areas, and if there is a technological breakthrough then we have eyes and ears.
The dark side of digital.
Well, at a humanity level there are plenty of challenges. If people are not trained what would they do? If they are not, there will be three things. There will be social unrest, there will be inter and intra country conflicts, and finally, there will be a large drop in aggregate demand because if I am not getting paid and I am not employed, I cannot buy anything. That is at the global level and I worry about this. My day job is to digitise the world and my night job is to worry about what happens if I actually succeed.
The second part is at the enterprise level, cyber security level… even cyber security is a massive emerging field. There are countries and there are private players whose ambitions and the intent is highly questionable and they target the weak, the young… because the young are not so aware and the old because they are not just able to defend themselves. And, thirdly, those companies who are not able to digitize and make it a way of working they will see significant non-achievement of their ambitions.