The One Week Generative AI Challenge

We challenge you to use generative AI in as many ways as possible for one week.

Midjourney prompt: /imagine an abstract way of illustrating a group of people trying something new that they think may be impossible for the first time

Last week, we wrote about how we asked our practice to list all the ways they were using generative AI tools like ChatGPT, Claude, and Copilot at home and work. The response was remarkable — we collected well over 100 distinctive use cases. But what struck us most was not the sheer number of applications but the affirmation that the best way to understand how these tools work, what they are good and not good at, and how they can make your work and life more effective and efficient, is to use them — extensively and purposefully.

Many folks who follow this stuff, including us, believe that spending around 10 hours hands-on with these tools is a tipping point where the aperture of possibilities really opens up, and you start to develop a keen sense for their strengths and limitations. With that in mind, we have a challenge for you. For the next week, try to use generative AI in as many ways as you can. But here’s the twist: the question isn’t what CAN it do — instead, try to use it in all the places you think it CAN’T do something. Push the boundaries and test the limits.

Use whatever model you prefer — Anthropic’s Claude, Google’s Gemini, OpenAI’s ChatGPT, or Microsoft’s Copilot. Or more than one. If you can, use the leading version of each (Opus for Claude, Advanced for Gemini, GPT Plus for ChatGPT, and the “Creative” mode in Copilot), but don’t let that stop you if the free versions are what you want to use. Ask them questions that you happen to ask yourself through the day. Give them things to read. Engage them in conversation. Upload images for analysis. Ask them to help you develop an application for your phone that you can use to track your travel schedule. Whatever. Stretch the capabilities as far as you can. If you’re wondering how to start, try asking it — for example, “Hey Claude … I have to plan seating for a meeting of 100 folks. How can you help me do this?” Or, “Hey GPT, I have garbanzo beans, leftover fried chicken, crusty bread, and celery. What can I make for lunch?” The goal is to try to use a generative AI tool to do or assist with every task you can think of, no matter how unlikely it may seem — even things you think might be impossible. If you’re inclined, share your experiences in the comments for this post or drop us an email at confluence@crainc.com. But the real purpose isn’t for us to collect use cases, it’s for you to gain hands-on experience and insight.

A word of caution: Generative AI models can and do make mistakes, so please don’t trust anything where there is risk to reputation or wellness without thoroughly verifying facts and information, and please follow your corporate policies if using it for work. But otherwise, for one week, immerse yourself in the world of generative AI. Use it as much as you possibly can, for as many things as you can think of, and see what revelations emerge. We’re eager to hear about your experiences on the other side of this one-week challenge.

---

AI Disclosure: We used generative AI in creating imagery for this post. We also used it selectively as a creator and summarizer of content and as an editor and proofreader.