The two of us started Women Defining AI with a simple mission in mind: to take this moment of widespread change and artificial-intelligence adoption to flip the script and close the gender technology gap at work, helping women adopt AI in meaningful and impactful ways.
One thing that we’ve learned from a year of building this community is that most people really don’t care that an AI model went from 32,000 “token context windows” to 200,000 or 2 million even. What people really want to know is what does this mean I can use it for? What is something that I can do today with AI that is fundamentally different and powerful than what I was doing before? What are the really game-changing use cases?
Among the hundreds of women we’ve spoken with, the most common thing we heard was that they were using ChatGPT as a slightly more advanced form of Google search. And sometimes it was worse because it could make up facts, also known as “hallucinating.” To get people curious about experimenting with the power of AI, we had to find aha-moment use cases that make them think “I didn’t realize I could even do that.”
Here are some of our favorite aha-moment use cases that expand our idea of what’s possible with AI:
- AI as a friendly colleague: Simulate or practice difficult conversations with AI voice. We know how important it is to practice giving feedback before we dive into difficult conversations with a colleague. But it can be hard to find the time or the right person to practice with, especially when you already feel nervous about the situation. Use the AI chatbot function that lets you talk to it (in ChatGPT, access this from the microphone icon) to set up a practice session and get feedback on how you can improve.
Prompt: I want to practice giving feedback to a colleague of mine. [Here’s the context]. Set up a scenario for me and run through the simulation to practice a feedback conversation. At the end, please wrap up with any suggestions on areas I can improve on.
💡Expert tip: Practice any type of conversation scenario with this technique, like compensation negotiation, a company Q&A session, or an external industry panel.
- AI as a thinking partner or coach: Ask me questions one by one. Getting AI to ask you questions rather than the other way around is a simple but powerful paradigm flip to unblock yourself at work. Sometimes, it’s easier to answer questions one by one, as if you are being interviewed, instead of trying to write the perfect prompt or instructions.
Prompt: I'm facing this challenge of [describe challenge]. These are my goals [insert goals], and here are the people involved, [insert people]. Here are also the assumptions I'm making, [insert assumptions]. Please interview me by asking one question at a time to sort out this issue. After I answer each question, you can ask the next one, building on the information I provide. Always get my answer before moving on to the next question.
💡Expert tip: Pair this with AI voice features to make it really feel like an interview.
- AI as an editor: Take my presentation transcript and tighten up the script. How often have we spent time writing and rewriting our notes or script for a big presentation? Get a transcript of a previous time you made a similar presentation, from the video conferencing software such as Zoom that you used or an audio recording you made. Then ask AI to tighten up your remarks before you present them again. (You can also ask it to update or modify your presentation as well.)
Prompt: I recently conducted a presentation and have the transcript from Zoom/Meet/Teams. The transcript is quite verbose and includes some unnecessary details. Can you help me tighten up the script? I'm looking for a version that is concise, to the point, and retains all the key information and messages. Here is the transcript: [Insert the transcript here]. Please revise the transcript to make it more impactful and easy to present, ensuring it flows naturally and is engaging.
💡Expert tip: You can do this with nearly every communication you might want edits or feedback on—from an email or blog post to an onboarding setup guide or podcast script.
- AI as “perspective taker”: Provide feedback from multiple perspectives and help me find my blindspots. We sometimes struggle to look at situations from multiple perspectives. Even when we do think of various stakeholders, sometimes we forget about others. Use AI to create feedback from different people or functions to help identify blind spots or additional considerations for your projects.
Prompt: As we evaluate our current project of [insert project], let's identify any missing elements in our brainstorming and consider the situation from various stakeholder perspectives. Reflect on the needs, concerns, and expectations of [insert stakeholders]. Also, identify stakeholders we’re not thinking about. Think about the key assumptions, potential risks, and alternative approaches we might not have considered from each stakeholder perspective. Provide feedback on how we can improve our planning process and ensure alignment with our long-term vision and values, which are [insert vision and values].
💡Expert tip: This is especially helpful for managers when writing performance reviews. You can also provide a version of this prompt template to your managers, along with unconscious-bias considerations or perspectives that they can use to quickly and more systematically identify their own blindspots.
- AI as an expert data scientist: Analyze data and create visualizations for me. Most organizations have more data than any person can process themselves, such as data from customer surveys or organizational-culture feedback. That includes data that are easy to group and analyze (like numbers) or hard and qualitative (like open-ended survey responses.) An underused use case for AI is to upload these data sets and very quickly get charts and visuals without enlisting the help of a data scientist or analyst. (We recommend stripping the data of any sensitive or personally identifiable information first.) You can analyze data using the Data Analyst feature in the paid version of ChatGPT or in Microsoft Excel using Copilot (also paid.)
Prompt: [Attach or upload the CSV file.] Please analyze the data I'm about to upload. Summarize your findings. Then create some visuals and charts for me to illustrate these key insights.
💡 Expert tip: You can try to be more specific about the type of chart or visualization, such as a bar chart, scatterplot, heat map, or word cloud. Keep diving into the data and insights with follow-on questions.
With all the AI media attention, it’s easy to forget that we’re still so early in our maturity curve with generative AI and this is true for everyone. As leaders, we can help our teams get started discovering and making use of these and other aha use cases by:
- Making time to experiment with the prompts and use cases we talk about here. Build an intuition for how paradigm shifting this technology is by getting hands-on.
- Push your thinking beyond “What am I trying to replace or automate?” to “What can I do that I could never do before with this technology?” or “What can I do that will save me a lot of time?” Dream and experiment bigger and push your teams to do the same.
- Create a culture of continuous experimentation among your colleagues, sharing and remixing approaches to inspire new use cases and bring together teams and expertise in unexpected ways.
There are so many ways that AI can already help improve work today. The best way to continue to unlock and build more advanced AI use cases that truly supercharge our teams is not to wait for the perfect AI guidance, but to enable more hands-on experimentation and cross-organization sharing today.
Written by Helen Lee Kupp & Nichole Sterling, founders of Women Defining AI.
© 2024. This article is reprinted with permission by Charter Works Inc. All rights reserved.
Advertisement
An organization run by AI is not a futuristic concept. Such technology is already a part of many workplaces and will continue to shape the labor market and HR. Here's how employers and employees can successfully manage generative AI and other AI-powered systems.
Advertisement