Making GenAI Really Real
A stomach flu bug stopped me from presenting last week at ASUG Tech Connect. I was so pumped to give my talk "Making GenAI Real: Project Lessons Learned". I'm sorry, 40-minute-long slide deck - you may never see the light of day. However, I can salvage a few nuggets for you, dear reader! Here are three takeaways I've gleaned in the last year.
First, people don't know what generative AI can do. Your second order of business (after proving that it can work for your use) is show and tell. Educate people about what you're giving them. Show them examples. The team I work with has a standing order to create short explainer videos, and when the non-techies see them, they instantly cook up ideas of their own. When it comes to something totally new, people need to see to understand.
Second, re-think design possibilities. Generative AI is not just "you type text into a chat window, and you get back possibly interesting text." Agents can make intelligent things happen in the background without any user interaction! Documents can be created with pre-baked intelligent comments! Sending IMs back and forth with a GPT is just the beginning.
Finally, in the enterprise there can be speed shock. Now that certain classes of actions can happen just by prompting, the distance between idea and prototype is shrinking. Some teams - e.g. security, infrastructure, and support - may not be equipped to move with the speed that a cracked AI team can. Establish patterns to ensure that a fast-moving AI factory can pair with ready-to-go delivery mechanisms. PM
This story appeared in Issue 71 of The Boring Enterprise Nerdletter. Subscribe for more stories! https://boringenterprisenerds.substack.com
Senior Applications Engineer at OGE Energy Corp.
2 个月You can’t crossover, until you’ve performed UAT