Where Does AI Actually Belong in a Manufacturing Sales Process?

One of the easiest mistakes to make is starting with the technology. A company sees an impressive AI demonstration and starts asking: “How can we use this?” A better question is: “What does this actually change in our business?”

A demonstration is not a use case.

Look at what happens when a customer asks for a quote. Someone has to understand the requirements. Someone needs to identify the right product. Someone may need to configure it. Engineering may need to check the solution. Pricing needs to be calculated. Information may need to be collected from an ERP or other internal system. Then the quote needs to be prepared and sent.

AI may be useful in several of these steps. But it may not be the best solution for all of them.

Customer requirements often arrive in forms that aren't structured for your internal systems. An email might contain dimensions, load requirements, product preferences, special requirements and information about the application — all mixed together in natural language.

AI is particularly good at working with this kind of information. It can help turn an email, RFQ or conversation into structured information that the next step in the process can actually use.

The important distinction is that AI doesn't necessarily need to make the final decision. It can prepare the information so that the right person or system can.

Complex manufacturers often have extensive product ranges, and finding the right combination can require significant product knowledge. A traditional search assumes someone already knows the product number or specification they are looking for.

AI creates another possibility. Instead of asking “Which product number do you need?”, a customer or salesperson can describe what they are trying to achieve. The system then helps identify the relevant products, options and questions.

This is particularly interesting when combined with a product configurator. AI can help understand what the customer wants. The configurator can then determine what is actually possible. That distinction matters.

This is where AI and CPQ or product configuration become particularly interesting. A manufacturer may have hundreds or thousands of possible product combinations. Some are valid. Others aren't. Some options affect dimensions. Others affect load capacity, pricing, components or manufacturing requirements.

AI can help guide a user through this complexity. But we don't believe AI should simply be allowed to invent a configuration. Product rules need a reliable source of truth — and that is where traditional configuration logic, CPQ systems and structured product data remain extremely important.

AI can interpret the customer's needs and guide the interaction. The configuration engine determines whether the resulting product is actually valid. The configuration can then be translated into geometry and engineering data, allowing the appropriate engineering process to validate the result.

That translation — from configuration to geometry to validated data — is exactly the kind of engineering automation we build for manufacturers

AI doesn't have to replace product configuration. It can make product configuration easier to use.

Quoting is another area where AI can be useful. Not necessarily because AI should generate the final price — rather because there is often a lot of work surrounding the quote.

Automating these smaller tasks can make a significant difference to quote turnaround time. And in complex-product sales, speed matters.

A faster quotation isn't simply an administrative improvement. It can mean a salesperson responds to more opportunities without adding the same amount of manual work. It also means fewer quotes waiting in a queue — the pattern we wrote about in why your engineers shouldn't have to approve every quote.

An AI model is only as useful as the information it can reliably work with. Manufacturers often have enormous amounts of product knowledge already.

The opportunity isn't necessarily to replace all of this. It's to make it more accessible and useful. AI can become a new interface to information that already exists across the organization.

AI doesn't fix bad product data. It just makes the consequences faster.

But this is also why structured product data and connected systems matter. AI becomes much more powerful when it isn't operating in isolation.

This is perhaps the most important point. Not every manufacturing process needs AI.

Sometimes a simple automation is better. Sometimes a product configurator is better. Sometimes an ERP integration solves the problem. Sometimes the process itself needs to change before any technology is introduced.

And sometimes the best solution is a combination: AI plus product configuration plus automation plus integrations. The objective shouldn't be to put AI somewhere because AI is available. The objective should be to remove friction from the process. The cheapest automation is the process you no longer need.

And finally: is AI actually the right tool for this problem?

That last question is important. Because the goal isn't to build an AI-powered sales process. The goal is to build a better sales process. AI is one of the tools that can help get you there.

The real opportunity is not any one of these technologies. It's connecting them. That's where complex product sales can start to feel very different.

The technology is the easy part. Finding where it creates real value is the interesting part.