This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Read our privacy policy

Manufacturing is the lifeline of any country.
03

Making manufacturing sexy again

Akhila Tadinada, Co-founder and CTO of Xemelgo

The right tech is giving manufacturers visibility into factory and warehouse operations. It could also help them attract the best and brightest employees.

GAVIN: What does Xemelgo do?

AKHILA: Our software provides real-time visibility to factory and warehouse operators and eliminates the need for manual data entry by employees.

Companies lack visibility throughout their entire operation. Our customers told us, “I don't know when things are going to get done, or if they got moved at the right time.” So we offer four applications: one for inventory tracking, one for work-in-progress tracking, one for tool tracking, and one for shipments.

It’s all pre-built; there is no IT infrastructure to set up. Once you plug in and start [RFID] tagging things, the data starts flowing to our platform. We start giving you real-time operational data right away. It’s pretty much up and running in less than half a day, which is like setting up your Alexa or Google Home, but in a factory or warehouse.

G: What does that do for the customer, and how is your offer distinctive within a very competitive market?

A: Sensor data effectively is dumb. What does it tell you? “I’m tag number 123 and I'm here.” Does that mean my inventory has gone up, or gone down? You have no idea.

So we give you business data. We tell you, “Hey, your blue paint can has come through the door, and you have enough blue paint. But you only have three more cans of red paint, so you’d better replenish that.”

Previously, how did somebody know they were low on something? They pretty much had a person sit down and count everything once a week, or once a month. It was time-consuming. But now we look across their end-to-end operation, and we can we can tell them if there are bottlenecks emerging on the floor, or if their shipment is delayed, or their finished goods haven't made it to the customer. This all happens in real time, without human intervention. 

G: How do you stop them being swamped with data? How do you sift through it for insights? 

A: To the end customer, it's a black box. Every second, the sensor wakes up and sends real-time data - there's millions of data points. We built an extensive back-end that de-noises and de-duplicates that data. So, when the customer logs into our UI, all they really see is how much inventory they have or what their bottlenecks are, or just, “Here's an alert.” They really don't see all the data being captured, because to them, it's meaningless. They just want to know if some something needs to be taken care of. 

G: How do you address questions about security and privacy? 

A: It’s central to every conversation that we have with customers – and we work with top Fortune 500 and Fortune 1,000 automotive, space, aerospace, and defense companies, so our customer base is the most paranoid when it comes to security. We are secure by design: everything from the connection from the reader to our cloud software is all encrypted in transit. Our data is also encrypted at rest. And we’re not transmitting any sort of PII, or Personally Identifiable Information.

G: What challenges do manufacturers face when implementing a smart manufacturing solution?

A: We've designed our software to be intuitive and simple, so as to not intimidate our customers. Most software is built for manufacturers, not with manufacturers. We work with them and say, “Tell us your problem, we will co-create a solution with you and then we'll take it to other manufacturers.” 

The customer really feels they’ve been listened to, and that’s the first barrier to cross. We also always ask our customers to take a crawl-walk-run approach. Don't try to boil the ocean. This is not ERP implementation where you have to change everything overnight. Start with a small area, set up sensors there, and pick the problem that's the hardest for you right now. Learn from that, then move to the next step.

We keep it super simple – not trying to complicate the software, really designing it for factory floor operators and truck drivers. Small details matter: we keep our buttons really large because it's easy for people to fat-finger stuff. And with an aging population, we need to make sure the font is visible to them. It’s about a very human-centered design. 

G: You've said that when it comes to ICT, manufacturing is just getting started. Where do you see the journey going next? 

A: With global competition, you can no longer just run businesses whichever way you want. You have to be extremely competitive when it comes to your efficiencies and at what rate you're able to build things. Here in the US, there's a renaissance going on in manufacturing, with a lot of on-shoring. But there’s also a labor shortage, and if I get skilled people, if I get millennials, they don't want to come in and be typing stuff or writing Excel files. So, offering access to technology and upward mobility is a huge hiring imperative in order to be able to hire the best and the brightest. For lack of better words, make manufacturing sexy again.

Secondly, everyone is gung-ho for AI right now. But AI and machine-learning is only as good as your data. It's garbage in, garbage out. We now provide machine learning models for customers to do real time inventory and production forecasting. We look at consumption and production trends to arm the supervisors on the floor with data on how they can best serve their customers. Think of an operator or a sales guy on a floor, trying to figure out which of these SKUs (Stock Keeping Units) they need to replenish. There's 40,000 SKUs sitting out there and all they can do is run a bunch of Excel reports and PivotCharts to figure out how they should be replenishing it. We now have a large language model, a ChatGPT UI, where you can just ask a question: “What SKUs require replenishing? What else do I need to do in order to serve this customer?” And it pretty much runs those queries for you in a big machine-learning model and then comes back with answers in human-readable form.

So, like NVIDIA founder Jensen Huang, I think AI is going to make software more human. It is actually going to reduce the barriers to entry. Before this, unless you were a data analyst, you couldn't really be an effective sales guy because you couldn't punch through all of this data. But now, if AI can do that for you, if you can call on all the data that you have, then it makes it a lot more human, because people who are not tech-savvy can still serve their customers the right way. 

G: The greater the tech gets, the more human the process potentially becomes? 

A: Yes, it really comes down to enabling operators to do their best job.  Manufacturing workers are some of the most dedicated and hardest-working professionals anywhere. But they are riddled by inefficiencies. And I'm not one of those who believes AI is going to take over our jobs – particularly physical jobs. Those are going to stay. There are operational experts who make spacecraft parts, and satellites, and to design these things – we will always need people at the center of this. But we need to ensure that they can do their best work, by giving them the best form of assistance. That’s what IoT, machine learning, and AI can do. If you subscribe to the idea that all of these jobs are going to be taken over by robots, then go visit a factory floor. To get these robots to work effectively requires a lot of very talented humans.

G: So, what holds businesses back from digitalization? What is your message to those who resist it?

A: A lot of the resistance is because it’s like crossing the chasm. Manufacturers want to see their partners and competitors having done this before they take the step They don't want to be the first ones putting their foot out there and then having a false start or a failed project. That requires leaders to really encourage innovation and risk-taking. Again, don't boil the ocean. Start small. Start with your hardest problem. If you empower the right people in the right places, the right results are going to happen. 

Contact us! transform@huawei.com