Embracing AI: How will the hottest tech trend of 2023 impact Paperless Parts and our customers in 2024?

Since the launch of ChatGPT, artificial intelligence (AI) has been top of mind for everyone. Like with any groundbreaking technology, however, comes a lot of hype—and fear. I recently took on a new role at Paperless Parts as Chief Scientist, specifically chartered with leading the research and development initiatives that will bring AI further into Paperless Parts’ industry-leading quoting platform.

But what exactly does that mean?

Since our inception, Paperless Parts has leveraged cutting-edge technologies and approaches to streamline the quote-to-cash process. Our geometric interrogation engine, for example, is built on algorithms we developed to automatically analyze part models to identify key part attributes that drive costing decisions.

But 3D models are only part of the story when it comes to part specifications; most Request-for-Quotations (RFQs) come in via email and include 2D blueprints or drawings—it’s here that we see the next wave of transformative innovation at Paperless Parts.

Throughout the front office, skilled and valuable employees are bogged down by repetitive, manual tasks that are perfectly suited for AI advancements. Quoting a part today frequently requires lengthy email threads with the customer or vendors to clarify specifications and line up suppliers. Skilled estimators have to reverse engineer part designs from spotty technical files. Prints or drawings contain easy-to-miss fine print or specifications that dramatically change the cost of a part. An opportunity exists for AI to streamline these tasks in a way that surfaces critical information quickly and consistently so that the user can make the right decisions for the shop.

“AI makes suggestions, while humans make decisions.”

With Great Power Comes Great Responsibility

Whether it was Voltaire or Spider-Man, the quote rings true. AI is an especially powerful tool, but responsible use of AI is even more critical when servicing manufacturers that handle sensitive information, including Controlled Unclassified Information (CUI) that make up a significant majority of parts for the defense industrial base or a medical device company’s latest invention.

Shops frequently cite trust (“Will my prompts be used to help my competitors?”), security (“Will I be part of a data breach?”), and compliance (“Will I violate aerospace and defense regulations?”) as reasons to avoid third-party AI, and we want to address these concerns head-on.

First and foremost, it’s critical to understand where your data and our AI models live. As we announced earlier this year, Paperless Parts recently earned third-party attestation for FedRAMP Moderate Equivalency, which was the final step in preparing for the upcoming CMMC regulations. This means your sensitive data is stored and processed in our secure environment within Amazon GovCloud. We operate our own AI within this environment, so your sensitive data is not being sent to non-compliant third-party AI tools.

To further explain how we think about AI at Paperless Parts, we want to share a few key principles that are guiding our AI strategy to help put our customers’ minds at ease:

• Security & Compliance: Following the FedRAMP and NIST standards as a guide, we have implemented and maintain strict controls throughout our organization, computing infrastructure, and application. Your data resides in US-based, government-authorized data centers. These controls apply to the AI features in our application and the R&D processes we use to develop and train AI models. We do not send customer data to any 3rd party AI tools or systems.
• Transparency & Human Review: Our AI features assist users in quoting by suggesting values for certain fields. The application makes it clear when there’s a suggestion and makes it easy to review it. In short, AI makes suggestions, while humans make decisions.
• Testing & Continuous Improvement: We have a rigorous process for testing AI features. Before a feature is added to the application, we first develop benchmarks and measure how well it performs on these data sets. When results are strong, we test with users who have opted in to be beta testers. We collect feedback from these users, both by talking to them and by collecting data about how the features perform. From there, successful features are made more widely available.

Lastly, I want to assure the entire Paperless Parts community that we are using AI to make your lives easier and more efficient. We recognize that AI is an incredibly powerful tool for extracting and analyzing data and answering questions with clear objective answers.

What it’s not good at is the art of manufacturing. Each machinist thinks differently about the exact steps to make a part. Each shop has a unique set of equipment or staff with different experience levels. Each shop differentiates itself on different attributes—the ability to turn products around quickly, expertise working with specific materials, or the ability to handle particularly complicated parts, for example.

Our goal is not to automate quoting through AI; we think that is a fool’s errand that diminishes the level of expertise required to build a long-lasting, successful shop. Our goal is to use AI to surface objective information in a clean, intuitive way so that you can make informed decisions about how to cost, price, and make a part.

–––

Scott Sawyer is the Co-Founder & Chief Scientist at Paperless Parts. Scott began his career as a software and electronics engineer at Lockheed Martin. Later, he served as a researcher in MIT Lincoln Laboratory’s Big Data Group and as VP of Engineering at Ecovent before he co-founded Paperless Parts, where he led the initial development of the company’s product and core technology as CTO. As Chief Scientist, Scott currently leads research and development initiatives to bring AI into the Paperless Parts platform.