Warehousing and Large Language Models (LLMs)

ChatGPT brought mainstream attention to the power of LLMs. It's now up to ISVs to leverage such technologies for their customers.


Less than a year since ChatGPT burst onto the scene, OpenAI’s captivating innovation is bringing mainstream attention to the power of large language models (LLMs) to software developers. And businesses across the board – including those in the warehousing industry – are looking at how these models can help them run leaner and smarter, leveraging huge troves of data as a competitive advantage.

What Is a Large Language Model?

In simple terms, a large language model is a system that uses deep learning to understand massive amounts of unstructured data and generate conversational responses. The algorithm that underpins the model is fed vast quantities of data. However, the answers it provides are only as good as the information it’s trained on. The whole large language model experience comes together in a question-and-answer format that many people seem to enjoy and trust. Users are finding that LLMs can capably handle even nuance queries beyond the entry-level basics.

How Warehouses Can Benefit from Large Language Models

Large language models, whether ChatGPT or other entrants in the large language model space, can be applied to different aspects of warehouse operations.

  • Knowledge management. Warehouses can use these powerful new models as central repositories for their organizations’ knowledge. Instead of losing years of on-the-job know-how every time a talented employee walks out the door, now these models can soak up everything about core procedures and processes for all staff to learn from.
  • Data analysis. Want to see what ChatGPT-style systems can do? Plug it into an ERP, WMS, and other key systems, and watch the magic begin. The better it knows the ins and outs of a unique business, the better it will respond with analysis that makes a difference. The platform’s built-in security ensures that a company’s data isn’t accessible outside the organization.
  • Demand planning. This is one of the most promising applications warehousing operators should look into. A large language model integrated with sales and operations (S&OP) functions, as well as marketing, finance, and other software, can unlock smart, well-informed demand plans.
  • Automate customer service. Take some of the pressure off customer service teams by lightening their load. Use tech like ChatGPT to field customer complaints, take orders and manage other tasks.

Avoid These Pitfalls with Large Language Models  

Although large language models have the potential to streamline a range of tasks, save time, and help build data-driven decision-making into day-to-day workflows, warehouse operations should approach their implementation with a full understanding of potential drawbacks, including:

  • Errors Don’t make the mistake of treating the answers from a large language model like the absolute truth. ChatGPT and similar solutions still need human intel following behind them to ensure everything’s in order.
  • The Temptation to Replace People – It might go without saying, but AI is a poor substitute for human interaction. Warehousing clients still value the face time that greases the wheels of every good business relationship. And when something goes off the rails, a customer will expect to reach a living, breathing person, not a bot.
  • Data Privacy Large language models can be used internally, or users may choose to use an open platform that anyone can use. Be aware that using an open large language model exposes the data users feed it when providing answers to others. Few companies will sacrifice their IP, which is why many opt for secure, internal use.

Answers In an Instant Will Become Competitive Differentiators

For the warehousing industry, the advent of ChatGPT and similar large language models is an exciting step toward enticing new efficiencies. Even though tech innovators are still working out the kinks keeping LLMs from their full potential, ISVs should start looking at how these models can help them increase automation and unlock value. It’s only a matter of time before the best warehousing players will see competitive benefits from these platforms. Explore how they can benefit your applications today.

Jay McCall

Jay McCall is an editor and journalist with 20 years of writing experience for B2B IT solution providers. Jay is a cofounder of Managed Services Journal and DevPro Journal.

Zebra MC9400
Jay McCall

Jay McCall is an editor and journalist with 20 years of writing experience for B2B IT solution providers. Jay is a cofounder of Managed Services Journal and DevPro Journal.