I interviewed Jeegar Shah who discussed The Challenges of Industry 4.0.







You’ve been a Microprocessor design engineer, co-founder of a SaaS startup in the life sciences space and then a management consultant with McKinsey & Co - so what brings you to Falkonry?


Hi, Dustin, first up thank you for having me - excited to be doing this. Before I give you a long twisted answer - let me just cut to the chase and get to the 3 reasons as to “Why Falkonry?” - this is one key trait that you’ll see with ex-McKinsey consultants.


  1. Well, The obvious promise - There’s been a lot of buzz about Industry 4.0 and the $11T IoT opportunity with 50B connected devices - you’ve heard all the pundits so I don’t need to repeat myself, but the take away here is that Industrial IoT holds the biggest promise - about $4T. Yes, there is hype and uncertainty but I am convinced we are at the cusp of something big.
  2. The geek factor - I’ve been around geeks that bump into glass doors and I’ve been around buttoned up professionals that can sell a Lamborghini while biting into a doughnut. I am personally more comfortable around geeks - I’ve come around working with people who have bruised foreheads. And the geek factor at Falkonry hits the roof!! Yah not even the glass doors.
  3. The product - Machine learning/ pattern recognition may be buzz words but I was sold when I learnt about Falkonry. It all made sense. It was intuitive and I could connect the dots. Fortunately, my HW and SW background perfectly fit into the Falkonry product journey. Seemed like the best way to put my engineering, product and management consulting experience into practice.


Well, that’s 3 reasons but I did end up giving you a long twisted answer. Guess I was never cut out to be a management consultant after all and another reason why I am here.


In your opinion, what are some of the ways of capturing value from Industry 4.0 applications?


There really isn’t a text book answer for this and the answer will vary depending on the subject and context at hand. But if I were to generalize I would say:


Digital enablement:  In my opinion providing a digital gateway to manufacturing can accelerate existing lean processes, and help to build a data-driven mindset, thus laying the foundation for more advanced technologies that will in turn pave the way for efficiencies and optimizations needed for Industry 4.0 to succeed. Digital dashboards for example help support performance analysis to achieve OEE improvement by increasing engagement of frontline operators and management. People often don’t realize that performance data usually persists beyond the shop-floor whiteboard. It supports normalized calculations and reporting, allowing KPIs across previously siloed functions, plants, and business units to be shared and benchmarked for consistency and best-practice sharing


Automation: When we talk automation, we imagine robots taking over the floor and displacing humans. Yes, for sure we expect significant adoption of robots on factory floors. This trend is expected to further accelerate because of advances being made in sensing the data (and hence the buzz on sensors) and making sense of the data (which leads us to the discussion on Artificial Intelligence). So it’s not just the invasion of robots and automation on the factory floors. Automation can also apply to decision making, resource allocation, demand forecasting, order management, etc


Throughput optimization: Factories stay competitive by maximizing throughput and managing costs. So the digitized data that we spoke of before can be used for process control systems to optimize yield. Now, in manufacturing operations, a lot of operational intelligence characterized by time series data stays hidden. Identifying bottlenecks OR processes that are largely responsible for impacting quality OR levers that impact the yield are hard to decipher for a human. Throwing the right technologies to get a better real time measure of operational efficiency can lead to higher throughput and higher profits.


Predictive maintenance: Predictive maintenance typically is geared to see an increase in machine availability and a reduction in maintenance costs – based on the introduction of new predictive maintenance algorithms. Companies need three components to be successful in predictive maintenance: knowledge of the respective asset, strong advanced analytics know-how (such as machine learning), and the appropriate change management capabilities.


If you notice, all these capabilities are not orthogonal to each other. Digital enablement leads to Predictive maintenance which is turn leads to Throughput optimization. Automation is needed to ensure scalability of all these capabilities. But the take away here then is that a disciplined, data-driven and measured approach are the key enablers to Industry 4.0.               


You mentioned Industry 4.0 and the buzz. So where do you see some of the challenges?


Industry 4.0 has raised definitely high expectations, and not all have been met yet. And there definitely is the accompanying buzz of Operations optimization, predictive maintenance, inventory optimization, health and safety - you name it - not all of which has been measured in terms of expectations and measures of success. Everyone, wants a magic wand. But Industry 4.0 implementation is a multiyear process, and more applications will develop as technologies mature further. It is imperative that manufacturers across the world start now with a set of concrete applications.


There’s been a degree of disillusionment in the actual implementation and results so far are mixed. On the one hand, we still see high uncertainty among manufacturers regarding what implementing Industry 4.0 really requires of them – and many are still struggling to even get started. On the other hand, most technology suppliers have moved relatively fast in adjusting for Industry 4.0. Very few technology providers and manufacturers have an overall strategy in place and even fewer have assigned clear responsibilities and goals.


Industry 4.0 requires a plethora of technology requirements - both IT and OT. Whenever you have a confluence of varied technology stacks, you can easily find yourself in implementation mayhem where half-baked solutions that are often patchy and in hindsight the solution stack not scalable or portable. Executives are driving Industry 4.0 strategies without understanding and appreciating the nuances and limitations of machine learning, big data, OT/IT protocols, connectivity standards, etc. For eg. executives clamor for more data and throw resources at big data technologies. Big data is more often dumb data with little contextualization. The secret lies in marrying human experience with algorithmic power to identify patterns that are more conclusive, useful and actionable.



I guess you are heading somewhere with that - and so how do you think Falkonry fits the bill here?


Yah, sure! But before I get started with that - let’s dig a little deeper into the motivation. Peter Thiel in his book “From zero to one” - picks up on an interesting discussion. There’s this ominous outlook and concern of AI and machines taking over the world, humans being replaced from their jobs and Skynet beaming humans out of existence. This all calls for some interesting science fiction but he cites several examples where human skills are complementary to opportunities that can be enabled through computers and AI. Machines can toss gigabytes around but aren’t as great at making judgements. Case in point, you can train a model to recognize cats with 90% accuracy after scrumming through thousands of training images but that’s something a 4 year old can do easily.


Falkonry builds up from this premise. While Falkonry can use advanced signal processing and machine learning techniques to identify temporal patterns in time series data, the data gets enriched when a subject matter expert annotates his experience on top of those temporal patterns. With a more contextual perspective to these patterns, it suddenly makes much more sense. Thus, Falkonry is a tool for subject matter experts - a tool that empowers them to cut the cord of dependence on data scientists and programmers that work in their own bubble. It also helps minimize context that is lost in translation as humans talk to each other and instead provides a paintbrush for subject matter experts that understand the operational data best.


So to sum it up, the answer maybe lies in the fact that it would be misleading to look for an all encompassing solution that works as a magic wand - A wand that would minimize human interaction and propel us into an automated world of connected devices that make smart decisions. What’s really needed is a way to equip, enhance and complement human judgement with better data enabled decision markers - and that’s Falkonry to the Industry 4.0 challenge.



About Jeegar Shah






Jeegar Shah


Sr Director at Falkonry


LinkedIn Profile