Charles R. Goulding and Andressa Bonafe discuss how AI-driven sales tools are reshaping industrial manufacturing by reducing misconfigurations and accelerating deal flow.
As industrial manufacturing systems continue to grow in complexity, selling advanced equipment increasingly depends on the ability to translate deep technical expertise into clear, application-specific guidance. In areas such as additive manufacturing and industrial 3D printing — where systems are rarely sold as standalone machines — sales teams must navigate intricate configurations, materials, workflows, and use cases, often relying on knowledge buried in manuals, PDFs, or concentrated in the hands of a few specialists. We spoke with Le’ora Lichtenstein, CEO and Co-Founder of Corbel, an AI-driven platform built to help industrial equipment manufacturers organize and surface critical product and application knowledge across the sales process. In this interview, she explains how AI can support sales teams in additive manufacturing, reduce knowledge bottlenecks, and help customers reach confident purchasing decisions more efficiently.
In companies that sell highly complex industrial equipment, like industrial 3D printers, knowledge often becomes a bottleneck as organizations scale. From what you’ve observed, where do these knowledge gaps most commonly emerge?
We typically see two major fronts where bottlenecks show up: the sales side and the servicing/support side.
On the servicing side, people need deep expertise to diagnose what went wrong. There can be thousands of variables that influence the root cause, so troubleshooting requires highly specialized technical knowledge. That’s an area where we consistently hear pain: companies spend a lot of money getting experts up to speed, and turnover can be high because it’s a demanding job that involve constantly dealing with problems, unhappy customers, urgency, and time sensitivity.
On the sales side, industrial equipment is unique because it’s very application-driven. When someone is buying a 3D printer, for example, the first question is: can this machine do what I need it to do? That is the number one framework in which they decide whether to move ahead with the purchase or not. Everything else (price, budget, and so on) often comes second. The problem is that the answers are buried in long technical specs, user manuals, PDFs, and in the heads of veteran sales reps. It becomes a pyramid, where only a small number of people have the detailed information immediately accessible.
You mentioned that knowledge often concentrates in a small number of people, creating a kind of pyramid within both sales and service teams. Why does this type of expertise tend to remain informal rather than documented or systematized?
Industrial equipment, for better or worse, tends to be on the lagging side of technology adoption and modern go-to-market best practices. Tech companies often adopt new tools and processes first (things like sales playbooks and structured “market motions”). You don’t see that as much in industrial manufacturing.
Historically, this industry has run on trust, relationships, and handshakes (getting on calls, going out to dinner, building credibility over time). And that makes sense when someone is making a very large purchase. But because of that dynamic, many sales interactions aren’t recorded or documented. So compared to other industries, industrial manufacturing is still catching up to what is now considered best practice.
Corbel positions itself as an AI operating system for industrial equipment manufacturers. How can that approach help companies selling highly configurable systems, like 3D printers, where needs vary widely across applications?
A key point is that there’s often more documentation than people think. The issue is that it’s buried in long, dense PDFs and materials that buyers don’t know how to navigate. The real question is: how do you make that information easy to surface, digest, and understand? That’s where AI can be a great fit.
What we do is train AI agents on each manufacturer’s proprietary machine data. We capture information from PDFs, videos, and documented sales or servicing conversations when available. Then the agent can surface the relevant information in natural language, in a way that’s intuitive for customers. So when a buyer asks, “Can this 3D printer do X, Y, and Z?” the agent is trained on that manufacturer’s data and can respond quickly and clearly.
This helps bypass a common bottleneck: a customer asks a question, the salesperson doesn’t have the answer, and the technical expert is busy. If it takes hours or days to get a response, the lead may be lost. AI helps deliver answers when and how the customer wants them.
That makes sense for documented information, but not all expertise lives in manuals or PDFs. How do you handle experiential, informal knowledge that isn’t written down? Do you try to capture that as well?
Yes, that’s a key focus. We train the agents on whatever data exists, but we also recognize the gap between what’s documented and what should be documented. So we’re building feedback loops into the platform to capture more expert knowledge over time.
For example, if a sales rep sees an AI answer to a customer question, they can provide feedback on that specific response. That feedback becomes part of a positive learning cycle to improve the agent. We’re also releasing functionality that drafts responses for internal teams: when a highly detailed question comes in, the system can draft an answer, and the team can quickly tweak it before sending. Those adjustments then feed back into the model. The goal is to continuously capture tribal knowledge that historically lives in people’s heads rather than on paper.
As you start embedding this kind of expertise into AI systems, some people worry that AI will replace jobs. How do you think about AI in this context?
Our core thesis is that AI is meant to superpower human beings. Technical teams at industrial equipment manufacturers are highly skilled experts. The goal is to free them up to do what they do best, and remove manual work that isn’t the best use of their time.
AI can also help when people aren’t available (nights, weekends, holidays, off-hours). But in industrial equipment, human beings won’t go away. Customers want someone they can call when they’re growing their business, when a machine breaks down, or when they’re deciding what to buy next. Technology should free humans up, not replace them.
Taking that human-centric approach into account, with complex equipment, misconfigurations and expectation gaps can still happen during the sales process. Can AI help reduce that risk?
Absolutely. There’s a natural link between sales and servicing: if the configuration is wrong upfront, problems appear downstream. If a customer buys a machine that can’t do what they need, or they push it beyond what it’s meant to do, the system may underperform, break down, or deteriorate faster.
That’s why configuration is so critical. AI can help guide decision-making by surfacing what the machine is capable of (and what a particular configuration is capable of) in a way that’s understandable to the customer. If you leave customers with a “here’s a 30-page PDF, read it and tell me what you need,” mistakes are almost inevitable. But if they can ask targeted questions (“I need to do X; is this configuration capable?”) AI can serve as a translation layer and highlight discrepancies early, before they become costly problems.
This seems especially relevant when systems become more complex. Additive manufacturing systems can be particularly challenging to configure. How is AI useful in a context like industrial 3D printing, compared to more standardized equipment?
One point I often share with our clients, especially those in additive manufacturing, is that they are on the cutting edge of manufacturing technology. Their equipment can do incredible things, and the sales process should reflect that level of sophistication. If you’re selling advanced machinery, your sales workflow shouldn’t be stuck in the 1980s. We want to provide tools that help sales teams sell in a way that matches the value their technology delivers.
In additive manufacturing specifically, we’re hearing about competitive pressure, including price dynamics from overseas manufacturers. In those conditions, you may not win on price. But the most important question customers ask is still: does this printer do what I need it to do? The faster you can get them to a confident “yes,” the faster you can move them through the decision journey – from “I need a printer” to “I need an SLS printer” to “this is the right printer for my business.” AI helps customers reach that confidence faster, which supports faster closes.
As AI becomes more embedded in these decision-making workflows, in industrial manufacturing it can be extremely powerful, but it also raises concerns around safety, reliability, and control. What kinds of guardrails do you put in place to ensure AI behaves responsibly, including when it comes to sensitive information and trade secrets?
Guardrails are a core part of how we think about building and deploying AI in industrial contexts. In manufacturing, hallucinations aren’t just annoying, they can be expensive and even dangerous if not properly controlled. That’s why we don’t use AI to run manufacturing operations themselves; our focus is on supporting the sales organization in a very guided way.
From a practical standpoint, we work closely with each manufacturer to understand which data should and should not be surfaced externally. We then build a proprietary RAG database for each customer that contains only the approved dataset the AI agent is trained on. The agent is not allowed to go outside of that dataset to generate answers. The moment you let an AI pull from the general web or broad external knowledge, you risk unreliable or inappropriate outputs.
If certain information, including trade secrets, isn’t included in the dataset, the AI simply won’t have access to it. Across the board, the key is leveraging AI in a highly constrained, intentional way, with clear protections in place to ensure it operates strictly within the boundaries it was designed for.
The Research & Development Tax Credit
The now permanent Research & Development Tax Credit (R&D) Tax Credit is available for companies developing new or improved products, processes and/or software.
3D printing can help boost a company’s R&D Tax Credits. Wages for technical employees creating, testing and revising 3D printed prototypes can be included as a percentage of eligible time spent for the R&D Tax Credit. Similarly, when used as a method of improving a process, time spent integrating 3D printing hardware and software counts as an eligible activity. Lastly, when used for modeling and preproduction, the costs of filaments consumed during the development process may also be recovered.
Whether it is used for creating and testing prototypes or for final production, 3D printing is a strong indicator that R&D-eligible activities are taking place. Companies implementing this technology at any point should consider taking advantage of R&D Tax Credits.
Conclusion
As this conversation highlights, selling industrial 3D printing systems is increasingly less about listing specifications and more about managing complexity, context, and confidence throughout the sales process. By using AI to surface the right technical knowledge at the right moment, manufacturers can reduce misconfigurations, shorten sales cycles, and scale expertise without over-relying on a small group of specialists. Importantly, efforts to develop and implement AI-driven sales tools — including software development, data structuring, system integration, and iterative improvement — may also qualify as eligible activities under U.S. federal and state R&D tax credit programs. For additive manufacturing companies investing in AI to modernize sales processes, these incentives can help offset development costs while supporting the broader goal of making advanced manufacturing technologies easier to sell, deploy, and scale.
