Skip to content

Category: Purple

Purple in Practice: From AI Skeptic to Everyday User

From an interview with: David Bergsman | Assistant Professor of Chemical Engineering, University of Washington

David Bergsman approached generative AI with some skepticism, concerned about both the quality of AI-generated content and what happened to information entered into commercial AI tools.

As a result, he largely limited his use of tools such as ChatGPT and Gemini to basic editing tasks, until he tried Purple.

“I came in somewhat skeptical. And I will say I’ve been extremely impressed.”

An AI Assistant for Everyday Work

Bergsman uses Purple as a resource while developing ideas, solving problems, and refining his work. He has used it to:

  • Brainstorm proposal ideas
  • Summarize papers
  • Develop outlines
  • Refine sentences
  • Work through technical questions
  • Support coding and projects involving SharePoint and Power Apps

“I primarily use it as a copilot and find it very helpful in that capacity.”

It is a significant change from his earlier use of generative AI, which was largely limited to occasional editing assistance.

A Time Saver and Conversation Partner

Bergsman often turns to Purple for a quick question or a “reality check,” allowing him to keep working without stopping to search for an answer. At other times, a conversation with Purple helps him explore a research topic or solution and uncover ideas he may not have considered.

Purple responds to Bergsman’s ideas, serves as a sounding board, and provides perspectives that he can then evaluate using his own expertise. For him, that distinction is central to using AI effectively: it should support human thinking rather than replace it.

A More Secure Place to Work with AI

Bergsman’s biggest concern with commercial AI tools was what happened to the data he entered. That made him hesitant to use AI for research, proposals, or other substantive work. Without being able to share enough context, he couldn’t get meaningful results.

This is one of the areas where Purple shines. Data entered into Purple stays within UW’s Azure environment and is not used to train public AI models. Purple is also designed to meet UW’s security and compliance needs, giving users a more controlled environment for university work. It provides access to multiple AI models in one place, allowing users to choose the right tool for each task.

Bergsman quickly saw the difference. Being able to provide enough context and select the appropriate model made Purple far more useful than the free tools he had used previously.

“My skepticism went away very quickly. I was like, wow, this is actually pretty helpful.”

Today, when Bergsman needs an AI tool, he says he uses Purple almost exclusively.

Learning When AI Helps and When It Doesn’t

Bergsman has used trial and error to find the right role for Purple in his work. In one experiment, he asked Purple to significantly shorten an existing proposal. The initial result looked good, but getting it to the right level of detail required more work than revising the original. In this case, relying too heavily on AI made the task harder.

For important intellectual work, such as proposals and papers, Bergsman prefers a conversational approach. He develops the ideas himself, uses Purple to test or expand them, and requests rough drafts or alternatives. He then decides what belongs in the final product. For lower-stakes tasks, such as summarizing feedback, he takes a more direct approach: he provides clear guidelines, reviews the output, and makes any necessary edits outside of Purple, usually without much back-and-forth.

His experience has led to a simple principle: use AI to help you think, not to avoid thinking.

Looking Ahead

Bergsman sees opportunities to bring Purple into more of his daily workflows, including connecting it with other tools. For example, he could use meeting transcripts to create summaries, identify accomplishments, and capture action items. But the biggest change has already happened. Generative AI has gone from something he approached cautiously to a tool he uses regularly alongside his work.

His experience reflects a practical role for AI at UW. It doesn’t replace expertise, judgment, or creativity. Instead, it helps people explore ideas, get started faster, and work more efficiently.

Learn More and Get Started