Derick Jose, Cofounder @Flutura Data Sciences
Lokesh Pyik, Chief of Business-connected industry @Bosch
Satish Cheeti, Cofounder @Cyient Insights
Rishi Bhatnagar @Aeris
Derick: Electromechanical work is like once you set it, it just works. Analytics needs iterations til it gets better. Customer was not happy first and then he got delighted with what AI did.
Lokesh: Business person shall validate or guide the data scientist.
Rishi: Technology can show you which solar panel has gone bad out of tens of thousands pkaced in hectares of a solar farm.
Lokesh: It takes a lot many personal connects to let the traditional industrialist see the analytics magic and trust it. Embracing the change is a challenge.
Rishi: Got invited to a cement industry. I wondered what I could say. A cement company in Australia started to sell cement in the form of an yarn. If you soak it water, it solidifies. This is disruptive.
Satish: There is a BPO entity that does mapping work. 2500 people worked on it. AI has potential to replace all of them in one go. Leadership is critical. A medical devices company asked us to predict an event of their machines. Then the sponsor stakeholder started to ask when would this start making money. CEO of an AC manufacturer company told us he would no longer sell the AC but will charge the customer as a SaaS. Thanks to IOT. Manufacturers are behaving like services company!
Derick: We are great at modeling the data and not dollars! Important to choose which problem to pick. We went to a tough southern Houston customer. We told him we have a data lake. He kept on saying “so what” and we responded then again cane “so what” and this went on many times. Finally we could say “you don’t have to goto rig and hence reduce the cost of trips”. He yielded.
You need these three to make analytics succeed in IoT of a specific industry.
Lokesh: We have Internal Startup Ventures at Bosch. Develop, Nurture and Align is our mantra. Once I told our internal team “Outside, young and energetic Startups are willing to do agile development and you better give aggressive timeliness as internal team else I outsource the work to them.” They didn’t like it.
We came up with an eye care product that converts retina scan as digital content and sends it to an expert who the assistance of analytics can say whether the person is prone to diabetes or not.
We have also started one more initiative that monitors pollution in a city. They have been doing pollution tracking in the heart of the city using a huge 10X20 room with sensors. We bought a small in-built sensor from Intel that got erected on a pole and it did the magic.
Derick: Lot of mechanical engineers got laid off in oil industry in Texas. One of the experiments we tried was to let engineer do the math instead of turning a data scientist into a serious engineer. The experiment helped. Houston is big on transforming laid off people into industrial data scientists.
Rishi: I was at AT&T in 2005. We wee supposed to test the first iPhone. I asked my team, how I dial the number! That device closed 49 industries today because of that touch device. There is going to be an impact on the skill set need to be acquired.
There will be 4.5 million more program managers are needed to manage 4 billion devices. They need 15 million developers to support. Each device is unique and.it csn be wifi, Laura zigby and so on. Re-skillkng is the key.
All of the following four are needed to succeed.