In Stay Sharp Episode 152, OpenBOM’s New AI Agents Solving the Problems Everyone Ignores, co-hosts Juliann Grant and Jonathan Scott sit down again with PLM expert and friend of the podcast Oleg Shilovitsky. In a previous visit (Episode 118, Digital Product Series: OpenBOM), Oleg explained that OpenBOM was founded to improve the way data is organized and flows between organizations involved in a product’s lifecycle. This time, the trio discusses how OpenBOM is taking improved data flow a step further by harnessing product memory and using AI to solve “the boring problems.”
The Flywheel Supporting Your Work
In a recent podcast (Episode 149, Product Memory: The Context Layer Behind Every Product Decision), Oleg shared the idea of product memory, a context layer that collects and contains pieces of information from different systems and activities about the product. Product memory in part became important to OpenBOM because of what they call the Product Memory Flywheel, a circular, repeating process for connected product knowledge. The first segment on the wheel is capturing information—files, design, BOMs, emails, Excels, etc.—to make it available for every part of an organization.
Oleg Shilvitsky, PLM industry expert and CEO of OpenBOM
The second is review or validation of data, such as a BOM review. That’s about checking, but also organizing a team to do the review seamlessly and with minimal effort from a human organizer. Oleg says, “It creates a new type of environment where everyone can go for validations, make a comment, get a task, do this work. We make a structure for the process that today is unstructured.”
The third segment is flow, which is the distribution of information. For example, when the system checks a BOM, it then updates files and sends the updates back to the original design system, ERP, or anywhere else that needs the information. Oleg makes the point that while OpenBOM can do some authoring, mostly it’s working with other systems where the data is authored, like CAD, ERP, or PLM. “As we spin this flywheel, the information, the product knowledge graph and product memories, gets richer and richer,” he says. “It’s accumulating information, it’s accumulating actions, it’s accumulating comments, it’s accumulating decisions, and this is where it all comes together. It’s progressing over time. That’s why it’s a flywheel.”
Tools that Feed the Flywheel
These days, Oleg is thinking a lot about how to solve the more tedious aspects of the product lifecycle. He says, “The more boring the problem, the better it is because if we can solve a boring problem, it’s important.” The first problem they’ve tackled is managing files, which he acknowledges may be the most boring task on an engineer’s plate. After repeated requests from customers, OpenBOM developed their own view of PDM. Then they created the CAD File Agent, agentic behavior for PDM that essentially hides file management from workers, allowing them to do the work via a chat. He explains, “It’s an agent that can help you and hide all pain of PDM from you, and the experience is completely different.”
The second boring problem they’ve addressed is reviewing or checking your BOM. “This is the premise people talk about all the time,” Jonathan notes. “A mistake found early in design costs less to correct.” In response to a flood of customers wanting APIs in order to put their BOMs in front of an AI, OpenBOM created the BOM Review Agent that works with any LLM you want. Oleg says, “We can put the bill of materials, also electrical, also mechanical, also software, also files, everything, structured data. And we can start running checks on this.”
The BOM Review Agent comes with a variety of pre-defined checks—for empty properties, circular dependencies, and an overall BOM AI health check.
In addition, an organization can customize their own unique checks. “Someone whose name is Joe knows that they always need to go to this place and validate this thing, they can program it,” he explains. “Or they have a very specific design that they want to check in specific way that everything is completed.”
Oleg admits that for many organizations, these tasks are painful and even “hateful” functions, but he loves figuring out how to solve them. “The boring problems are becoming my favorite thing,” he says, “because boring problems and amazing technology together create a lot of excitement.”
Learn More About OpenBOM’s New Capabilities
The full podcast contains more discussion and details on a variety of topics, including how they’re helping give AI the necessary information to make decisions with, the connectors they’re creating to ease the tedium of sending data to ERP systems, the company’s plans for making the platform more intelligent and helpful, and more.
Be sure to check out Stay Sharp Episode 152: OpenBOM’s New AI Agents Solving the Problems Everyone Ignores, and join us each week for a new podcast.
Follow the Razorleaf Podcast, Stay Sharp in Digital Engineering on:
