Want to make AI part of your team’s DNA? Start here.

Prompt and Prosper: that was the name of the day Sona Iliffe-Moon, Yahoo’s CCO, designated this past July for her 20+ Yahoo comms team to begin to build their AI strategy. “We made it very intentional, but low pressure,” says Iliffe-Moon.

The reason for the plan was simple: “When we had talked about deploying AI across the team,” she recalls, “we got feedback like, ‘I’m curious but I don’t know where to start,’ or ‘It feels like cheating.’ ‘This feels technical. I’m not technical,’ or ‘I don’t have the time,’ or ‘What if I do something wrong?’

“We wanted to get over that hump. Our hypothesis was that if they just played around with AI for a few hours, it would open a world of possibilities.”

Before the day began, every person on the team had to be prepared for three tasks:

1. They had to gain new knowledge, building comfort with various tools, exploring them, taking an online course, or listening to a podcast like “How I AI,” by Claire Vo.
2. The second was to automate a task that team members might perform regularly, that AI could make faster, easier or more efficient. “They didn’t need to overhaul entire workflows, just things that were maybe repetitive or time consuming. We encouraged the team to come up with ways that could add real value right now, not just in theory, and to help them think critically about where automation made sense. So even if it was saving five or 10 minutes, this exercise would help them uncover those opportunities,” says Iliffe-Moon.
3. The third was to create something, whether that was a prompt library, a resource, a template, or something that could be used across the team. “This was not about coding or building a full-scale product,” says Iliffe-Moon. “It was to shift the mindset from ‘How can I use AI?’ to ‘How can we use AI?’ for it to have team-wide impact.”

The results were immediate, according to Iliffe-Moon. “Everyone learned about different AI tools and nuances. I don’t think you can just squarely focus on one LLM, because each has its own personality, its own tone. Even testing out LLM updates was an important learning. If you love Claude for writing, you should try Gemini and ChatGPT too, to see what the different strengths are. That was one shared outcome.

“Several teammates created custom GPTs to help generate content and the voice of their specific executives for first drafts of emails, speeches or other internal or external materials. So now we have different bots for different people,” she says.

At the day’s end, the team had compiled a list of products to build; tracking the construction of those resources is a team KPI that’s covered in Yahoo’s standing Comms team meetings.

Post-meeting results

The Prompt and Prosper plan provided inspiration, says Iliffe-Moon. “One person on our team created AI-powered audience personas to help with message development and testing. If we’re rolling out an internal update, we can use themes from our employee engagement surveys and other internal insights, like tone preferences or topic sensitivities, to adjust to what different employee bases might be interested in. This helps ensure that we’re delivering the most relevant information to each of our audiences. The team has already started to use these products.”

Daily workflows and findings are informing next steps. “As communications professionals and leaders, we can envision what a future team looks like,” says Iliffe-Moon. “We will always be managing and guiding people, but we may have a suite of AI bots in specific roles, like a research assistant, a first draft writer, a sentiment monitor, the coverage assistant, with each having clear responsibilities. We have conversations in our one-on-ones asking: how did you use AI here? Or, could you have used AI here? Now it’s part of our team’s ongoing conversation.”

Robbie Caploe is director of strategic initiatives at Ragan Communications.