Generative AI tools are flooding the market and changing how we work at a pace that hasn’t been seen since the dawn of the internet. Learning and engaging with up-and-coming tools isn’t new in product design or in leading teams, but the speed of advances in generative AI make it hard to prioritize and find the right balance.
As a creative director, I find myself reflecting on how I personally stay up to date, get teams to try out new processes, and build space into the company to support that learning. We’ve evolved over the past year to become more efficient and rigorous in integrating these new tools with our work process. Here are a few ways we’ve gone about it.
Personal Level
I personally carve out regular time to engage with gen AI tools, both the ones I already know and new ones on the block. To keep up with industry trends I read newsletters and digests including Ben’s Bites, Evolving.ai, and thought leaders on LinkedIn. Every morning I read the digests and see if there’s a new tool worth digging into, and as I find interesting tips and tools, I share them on our company-wide generative AI slack channel to extend that knowledge and rise the tide across the whole team and company. My coworkers also take time to share what they’re finding interesting and leaning into to create interesting cross-pollination.
Through this exposure and dedicated time, a main mental model of how and why I use generative AI has emerged, mainly to refine strategic concepts and frameworks, generate more concise text, and create images. This current form of AI is best thought of as an unpaid intern or a collaborative brain buddy. I give it clear directions in the tasks I want help with, but don’t expect it to deliver the final version of that task. Taking the time to fact check any text-based outputs is important, as is giving feedback to refine either text or images, or even taking generated images into other tools to finalize how I’d like it to look. As a brain buddy, I use it to help me push on arguments and give me feedback on my work, but I still continue to assess any outputs with scrutiny because at the end of the day I’m responsible for the work and output of my team and this tool.
Specifically, when I work with ChatGPT, I start by giving it context into the meeting, presentation, or workshop that I’m creating.
- I give it information about the types of people who will attend or participate (e.g. the CPO of a fortune 50 retail company and their direct reports, who are each a VP of a different product vertical).
- I then share the intent and goals of the gathering (e.g. a workshop to inspire and help these participants better understand gen AI and how it will impact their industry and internal work processes).
- Finally, I frame a specific request for the AI to help me with (e.g. please create ten conversation starters for the lunch table that will help continue the learning and integration of the themes of the workshop).
- I continue the conversation based on the output of the AI, asking it for additional or different answers to refine the outputs until I get what I need and can take it from there (e.g. I like the first two questions, please make five more that reference the privacy, ethical, and data governance concerns for generative AI).
Team Level
Now that there’s a clear use case for generative AI and a mental model of how to work with it, I share it with the teams at work. They usually don’t need direct training on how to use individual tools, although we’ve done some of that too. What’s most helpful is the higher-level why we should use these tools, the process to engage with them, and the impact that they have on existing processes and outputs.
This regular learning time is good modeling to teams of designers and mentees — it shows that we need to have dedicated time for learning and continued education so that we aren’t left behind. Team culture is built from the ground up and the top down. Being towards the top allows me to demonstrate the growth mindset that I expect in design teams. We’re always looking for better ways to deliver high-quality work, so seeing a leader learning and experimenting emboldens designers to do the same.
I’ve been both impressed by and learned from other designers on our teams as to how they use gen AI to speed up their process.
For a really specific example, a designer shared how to use bracketing in MidJourney to process multiple prompts and generate lots of images at the same time, speeding up the image creation process and leading to my current flow of “scatter shot, then refine” for image generation.