Julie Cliche-Dubois discusses the nuances related to collecting, processing and sharing factory data

Eric: Why don't we begin by having you introduce yourself briefly and explain what your role is within Cogiscan?

Julie: Sure. So, I've been with Cogiscan for two years in the Product Manager role. And now, funny story, it's my second tenure at Cogiscan. I was previously with them 20 years ago as a software developer. And then, I moved to a support role and left to find what I liked in life, which is product management. So, I came back two years ago. In my role, I take care of the development roadmap around everything regarding machine connectivity and factory analytics and our Factory Intelligence product.

Eric: And how large of a team is that?

Julie: The development team is about eight developers because we take care of two different products. So, that's a lot of work to achieve. And we have a QA Analyst to perform the tests.

Eric: Well, let's get into this because I think you can share some unique insights here. When I think of electronics manufacturing, I think about just, especially since digitalization became all the rage, just the massive amount of data that's being generated by our industry, right? And then, that comes in different formats, and you have different protocols between the machines, older machines, newer machines. And I think a lot of people tend to think, "Oh, this is easy to do, right?" But it's, in fact, a very complex environment. So, share your insights: tell me what challenges that you and your team have to deal with in this environment, and how do you help your customers achieve that connectivity and then to generate usable data from it?

Julie: That's a really good question, actually. Thanks. We're continuously focused on improving our connectivity products, specifically our Co-NECT platform. So, we're constantly enriching our database of adapters in order to collect the richest and the best available data. And as you mentioned, this can be quite challenging, as many of the machine vendors have their own way of storing and providing the data. And it's not always as straightforward tapping into their data source. Now, we're getting more and more standardized data with CFX, for example, but some of the equipment out there, vintage machines only can supply a text file of data which contains serious black holes regarding machine information. Sometimes we have limitations, we're not able to get machine status. For example, maintenance conditions. We don't know when the product enters the machine, when the product exits the machine, and don't even think about interlocking. There's no way to achieve that with these files.

So, we worked closely with equipment vendors to tap into what they have. And in many instances, work with them to develop a solution to better access the data our customers need. So, sometimes we're able to get more data from them, they're able to modify the way they provide data, and our customers can directly benefit from it. It's our commitment to get the richest data available to calculate factory KPIs and better eliminate where our customer needs to make process improvements so they can do proper gap analysis.

Eric: That makes sense. So, when you're saying that you're working with them to develop solutions to get better access to the data, you're talking about the machine vendors there, right?

Julie: Yeah, exactly.

Eric: Yeah, so you're working with them and then they're learning how to kind of tweak their programming's. Do they sometime change their format?

Julie: Yeah, yeah, sometimes the format can be changed, and it can provide more data because at the end of the day we and the customers all want the same thing. And whether you are a software vendor or a machine vendor, you want your customer to get the information they need to take the right decisions.

Eric: You know, I always think of it in terms of, you gather the data and it goes up a pyramid and ultimately there's this great business insight and you can improve your business, right? Nowadays, we need to think about what happens after you gather the data, right? … but the tricky part is not to gather all this data but making sense of that from a business perspective. So, kind of tell us a little bit about that process and how do you process all this data to convert it into something usable?

Julie: Yes, definitely. So, we've been collecting factory data for over 24 years now. It was initially driven by our need to drive Material Control solutions and later for Traceability with our track trace and control applications, and now we are focused on Advanced Analytics.

So, we are approaching Factory connectivity to drive more top-level production solutions. We're not only concerned by what the machine can provide us, now we're collecting data from the entire manufacturing ecosystem – machines, products, materials and process data. So, with our SMT-specific domain expertise, we are able to contextualize the data, and this is the key. We are contextualizing the data so it can be more usable and understandable. We take all the information from the factory and we correlate it together.

So, for example, we take the data from the placement machines, from the MES and from the scanning operation at the machines and we tie it all together. We don't leave that information trapped into silos. We are tying it to a specific job order for example. Then you're able to track down the data with your specific job order, and you are able to get the relevant KPIs that will help you to take the right decisions. At the end of the day, the customers want to maximize their overall efficiency, so they need to get a full picture, not on only related to a specific machine.

Eric: And for those KPIs – kind of the dashboards you’re providing. I'm assuming that Cogiscan has a library of developed KPIs, right? But then, can those also be customized, can this grow over time for a customer?

Julie: Yes, definitely. A customer can have access to our databases, so if they want to customize their own KPIs, they just need to get the access to the database, and from there, they can build their own dashboard, their own widgets, that will really talk to them and will help them solve their issues. So, they could even come and say: “Listen, Cogiscan, this is how we see this, but we'd like to tweak it this way. You can develop something?” This way we can develop something and then we can work with them to customize whatever it is that they're needing for what they're trying to accomplish. And the key to this is machine connectivity. Once you have the data from the machine, which is one of our biggest expertise, then you are able to build all the KPIs that you really need.

Eric: Right. It was interesting, you know when I was at APEX, I spoke to both Martin and Mitch about kind of the AI use cases and what they were doing… And you know AI is kind of the big buzzword these days. So. how does the change in advanced analytics and especially with the AI powered kind of solutions change how you gather data and then even how you share it?

Julie: Yeah, you're right. AI is like the big thing right now, and we're such in an exciting time as we continue to develop our AI use cases with the IIoT.Edge platform. So, the tricky part for my team is making sure that the vast amount of collected and shared data is both usable and understandable and not just a jungle of comma-separated values, you know.

So, for example, when we’re sharing inspection measurement data with the IIoT.Edge platform, not only we have to gather the appropriate data but also provided it in a specific format and structure. So that data needs to be generated in a way that the algorithm knows how to access it, read it, and interpret it. Essentially, it knows what to expect from it every time, and when we talk about measurement data, that's a LARGE amount of data. So, we really need to parse it in a way that can be understood. In fact, we've put in place a standard, and this standard is so clear that if you wanted to read it, you would be able to read it, and you would understand what's inside all this data… that's pretty amazing, and this is what we're using for our AI use cases as well.

Eric: I'll be talking with Martin about AI next week, but it just strikes me how everybody sees it almost like this panacea for all the problems that we're facing. However, it's really still about the humans who are programming it and working what’s behind it, and THEN it’s a great tool to work with. From what I hear from you, it sounds like AI kind of helps you make the complex more manageable, right? All of this data that the industry is producing, and it makes it somewhat more manageable for you. Is that it?

Julie: That's right. We have access to so much data. But then you have to do something with it, and I don't think that as humans we are able to process it all in a way that's usable, so we need some help from something smarter than us, exactly.

Eric: How long for, or is that data stored in perpetuity? I mean, is that forever then, or are there standards around how long that data must be maintained?

Julie: Well, it all depends on the different use cases. In our case, we will save it as long as it is needed. So, for AI for example, you need to keep a set of data for the algorithm to learn and learn and learn based on all the different scenarios. I would imagine that, in that case, it's worth it to keep it for a very long time because the algorithm will get smarter and smarter.

Eric: And I'm assuming then just as Cogiscan’s solution is growing in the industry that your work must be keeping you very busy.

Julie: Oh yeah. That’s for sure. I don’t really have time to feel bored, so that’s very exciting!\

Eric: That’s a good challenge right!

Julie: Exactly.

Eric: Especially in this current environment. Well Julie thank you. This has been excellent. I think you've helped me understand it a little better and that's, you know if nothing else, that's what I always appreciate from these interviews. But I think our audience will appreciate it as well and hopefully at some point in the future we can speak again and hear more about what you and your team are doing.

Julie: Amazing, let's do that again.