Issue link: https://iconnect007.uberflip.com/i/1406534
66 SMT007 MAGAZINE I SEPTEMBER 2021 where we built AI products, we learned that if you want to have a product that's infused with AI and derives value from it, first you must dig- itize a process, own the interfaces to the right stakeholders, and start collecting meaningful data. at's exactly why our product here is a workflow improvement automation tool that collects the right pieces of information and stores them, so you can derive value from it in the future. For now, it's an RFQ tool for EMS with the most magic happening in the BOM importer and the automatic sourcing. As an EMS you can select BOMs of any file type and pull them into our tool without pre-formatting. Within seconds, LumiQuote extracts all relevant in- formation and creates a neatly structured over- view in a digital version. e person handling the BOM at the EMS can then check the iden- tified information and manually edit details if necessary—for example, if we didn't find the part number (MPN) from an unknown man- ufacturer. Moving to sourcing, certain scenar- ios can be set, like "I want to have a prototype with 10 pieces, I want it as fast as possible, and I want it from only my preferred suppliers," and our tool automatically gets the different prices from all the different suppliers through connected databases. Apart from that, we have the individual sup- pliers integrated for our customers and they get their specific custom prices directly in the tool. Potential issues with selected parts from certain suppliers, like long lead times or cheap- er prices for larger quantities than selected, are marked in the sourcing overview. is way, time from manual work is saved and the focus can be put on these "critical" parts. If we didn't find anything, it's marked with no sources. At this point of the process that I just explained, they could basically just go back to the dash- board, hit export, and get a quoted BOM. Nor- mally it takes them two weeks, but now they have all the prices and data within minutes. You can do your downstream process and, in the end, deliver a quotation to your customer. In the manufacturing feature, we allow us- ers to do a bottom-up calculation and to lat- er compare it with actuals aer production. With the blocks on manufacturing costs and material costs, we only need a flexible calcula- tion module for you to be able to do a full end- to-end digital quote. at's where it starts to get interesting, because then you can give the EMS company this tool and OEMs can do sales quotes on their platform. OEMs can start up- loading their BOMs, and we can start using these modules and bring them closer into the OEM workflow. But, for us, the process of de- veloping and prioritizing features was guided by how we can get to market as quickly as pos- sible, get customers, and get actual data. Holden: Sebastian, how does this handle the bare PC board? Schaal: at is something that you current- ly have to do with a manual quote, but we are co-developing with a German manufacturer to work with parametric quotes. We will have a Gerber extractor that delivers some param- eters like the minimal trace width, the num- ber of layers, etc. In addition, the user will have to enter extra information like the color of the PCB. If the PCB manufacturer has the neces- sary infrastructure, we can use this data to get you an instant quote or it will still be a manu- al quote, but the supplier will still get the ex- tracted parameters. If we can get an API con- nection with the supplier, this is our preferred choice. Matties: I imagine there are other soware so- lutions doing a similar task on the market? Schaal: ere are products out there. For us, the unique idea is the connection and transfer of knowledge into the OEM space later on. It's this larger vision that we have. In other prod- ucts, most of the stuff is built on legacy tech, so building a similar infrastructure like ours to turn data into insights is very hard for them.