A featured contribution from Leadership Perspectives, a curated forum for enterprise technology leaders, nominated by our subscribers and vetted by the CIOApplications Editorial Board.

Weyerhaeuser

From Hype and Hallucinations to Headway: How Weyerhaeuser is Embracing Generative AI

Since the release of ChatGPT in November 2022, generative AI has dominated headlines, sparking both excitement and concern across industries. Whether generative AI is used as a personal assistant like Copilot or in a specialized manner for an agent, companies around the world are still figuring out their strategies for enabling and upskilling their employees to take advantage of this technology. Additionally, generative AI is not perfect. It can still hallucinate, get confused on complex topics and scare people with its potential long-term abilities.

At Weyerhaeuser, to counter these issues and concerns with generative AI, we’ve created a new team focused on enabling and upskilling our whole company to take advantage of the technology where it makes sense to use. We’ve conducted several

training courses—primarily focused on GitHub Copilot, M365 Copilot Chat and Copilot for M365 and have courses on agents in development. We begin our training by grounding our workforce in the historical context of AI, from Alan Turing’s foundational work in the 1930s to today’s transformer-based models. This context helps demystify the technology and we’ve found that our employees grow more comfortable and confident with generative AI once they know this history.

Treat It as a Tool, Not an Oracle

We emphasize that generative AI is not infallible. Users must “distrust and verify” its outputs, treating it as a powerful assistant—not a flawless oracle. We view generative AI as another tool for our workforce to utilize, like Robotic Process Automation, taking the robot out of the human. We also make it clear that generative AI can and will hallucinate. We cannot discount the importance of human intelligence in reviewing what is produced by AI to determine if its conclusions are valid.

  • Generative AI is not infallible. Users must ‘distrust and verify’ its outputs, treating it as a powerful assistant—not a flawless oracle

We’re not stopping with training and Copilot for M365, though. As mentioned earlier, we’re beginning our agentic journey as well. We feel there is significant value to be added by implementing agents for specific use cases across our company. We’ve formed a new team to support this effort and are currently working through multiple proofs of concept in different parts of our business. We plan on utilizing custom-built agents as well as pre-built Microsoft agents, such as the Analyst and Researcher agents. However, before we dive headfirst into agents, we’re spending time with our internal teams to make sure we have the correct governance strategy in place to effectively develop, deploy and monitor agents.

To summarize, generative AI is an exciting tool. The value it offers is significant, whether as an assistant or an agent. To unlock the full value of generative AI, it’s critical to set realistic expectations for all employees. Clear communication, thoughtful governance and continuous upskilling are key. I’m very proud of the team we’ve built to support this technology and very excited to see where the journey takes us over the coming months and years.

The articles from these contributors are based on their personal expertise and viewpoints, and do not necessarily reflect the opinions of their employers or affiliated organizations.