The Fly on the Wall: AI, Instinct, and the Future of Internal Comms

The Fly on the Wall: AI, Instinct, and the Future of Internal Comms — Transcript

When the AI Gremlins Showed Up

VERNON: All right, cool, so we'll go ahead and start. Vernon Ross, your host, welcome to the Stories That Lead podcast. Every leader has a moment in leadership where something happened, something changed, and that's what we talk about here. We talk about real leaders and real frameworks in their real moments of revelation.

Today I have a very special guest. Rosemary Cassie is a senior communications leader where she works across executive communications, internal communications, and employer brand. Her career has centered on how organizations communicate change and build trust with employees in very complex business moments. She has great stories about how she navigates this in public, some really interesting takes on AI and how it relates to employees and engagement, and where the future of internal communications, and possibly some external communications, is going. Rosemary, welcome to the podcast.

ROSEMARY: Thank you, Vernon. I'm excited to be here.

VERNON: So I'll start with something interesting we talked about: how AI can not work the way you wanted it to, particularly when you're working with humans. I have a guest dossier process that builds an entire profile, takes my frameworks, and turns them into questions. Then I have a VA who validates everything: checking links, that kind of work. Well, she put a quote in there about something you supposedly said to your team. What did you think when you saw that?

ROSEMARY: First of all, I'm thinking: how would anybody know what I quoted to my team? Because it would be internal. Then I read more and I'm like, I don't think I ever said that. I think there's a few AI gremlins in your process.

VERNON: There were some gremlins. I've been talking about this a lot lately. When you're using AI in communications, particularly for the folks I serve in comms and learning and development, discernment is the superpower. Being able to know what you should publish, what you shouldn't publish, how to use it. The human in the loop decided to do the "too long, didn't read" version. She didn't read the research. The AI appeared so well informed, so confidently communicated, that she thought it was clean. She said, this looks great, and she just sent it. I didn't get a chance to look at it because I had near 100 percent confidence in her. She has never botched any guest research.

ROSEMARY: Well, it was funny, it wasn't just that question. Two of the questions had nothing to do with me. So when I first saw that first one, I was very discerning about the follow-on ones.

VERNON: Right. So I'm thinking, thanks, AI, you really helped me build trust with my guest.

ROSEMARY: I've been using AI enough to know that it hallucinates. It makes up stories based on lots of other factors. So I'm like, hey, Vernon, I think it's hallucinating.

VERNON: Yeah, it will make jumps. It will make big jumps.

From TPM to Communicator

VERNON: Okay, let's get into the interview. You've been doing executive communications for a long time, well before AI became a tool in internal comms. Before the tools changed, what did you already believe about what employees need to pick up from communications? And do you think that still holds with AI in the mix?

ROSEMARY: I want to step back even further, because this is an important part of how I use AI. Before I was in internal communications, executive communications, and a little marketing, I was a TPM, a technical program manager. I helped implement technology tools for solution architects of high-end server designs. So I've always been very comfortable with technology. I love to speak it. I love to work with engineers.

What I really enjoyed when I was doing that is, I could talk to an engineer for an hour about something and distill it down to a couple of sentences. That's when I found out, oh, this is what I want to do. So as I've moved through my communications and marketing roles, I've learned how to use technology to my benefit and to the benefit of the company. The more process and tools you use to take the friction out, the more time you spend on strategy and strategic efforts. That's where the impact is.

What AI has changed for me isn't really the quality of my writing, though I can do it much faster. It's giving me time back to think about strategy. Operations, especially operational work, takes time. And the reactiveness of any comms job usually throws your day off every day.

In terms of how AI is changing the role of internal comms, in any company I've been at, we're always moving fast. There are leadership changes, strategic pivots, technology pivots. I'm always trying to change the reactive nature of the function. Comms groups, because they're writers, creatives, often journalists, don't naturally think process first. But because my first ten years as a TPM, that's what was ingrained in me. So in senior roles, I used process to take out the friction, and then spent more time being strategic.

AI isn't just speeding up the process work. It's giving us time back to think and spend energy on the questions. What's the right message for moments like this in the company story?

Owning Every Channel, On Purpose

ROSEMARY: Now that I have time and a team that's also embraced AI, instead of throwing out a one-and-done message, we have time to think it through and go more omnichannel. I have the benefit of many different channels, and we can look at each channel and find the right message for it.

We just had a huge company moment, and because I owned all the channels and had the time, my team organized the promotion of that milestone across every channel. It was noticed. It made a difference in the way employees perceived it, absorbed it, and celebrated it.

VERNON: With the time AI has freed up, what tools are you using? And I'm interested in how you're listening, because the important part of comms outside the communication is listening. You need feedback to know the communication is effective. What are you using for listening, and what are you using for production?

ROSEMARY: We have the luxury, we're an engineering-led company, and we're given the tools we need. If you can put a business case together, you get it. My team and I have the luxury that our engineers are brilliant. They're using many of the top AI platforms. Because we help employees understand which tools they can use, we get to use them as well, if not first.

As AI has grown in capabilities, we started out, yes, it helped draft things faster. One of the funniest things we use it for, and it drives a lot of impact, is on Slack messaging. AI is great for putting emojis into your messaging without having to hunt and peck.

VERNON: I didn't expect that. It is really good for that, particularly in Slack.

ROSEMARY: That alone is saving us time. But it also gives more depth and texture to our Slack messaging. When I told my tools and productivity team I was using AI for emojis, they looked at me like, what?

The Push-Back Rule

ROSEMARY: We use AI on both sides. Like any good comms group, we have sentiment that's both qualitative and quantitative. We get that through surveys, through listening in team meetings, things like that. There's no magic button. We use AI to distill survey results, listening results, things like that. It helps us spot patterns, which we then synthesize.

What was really startling to me, and I'm excited to say this: while AI gives us all that, it really takes instinct and experience to interpret it. I didn't realize how much instinct I had until AI handed me a pattern and I had to decide what to do with it. I'm always asking myself, is this real? Did it hallucinate? Is it taking content and adding its own sentiment, its own perspective?

It's really important that all communicators push back. You don't just take AI's first pass. You push back with, hey, from my experience. And it will frequently say, oh, you're right.

VERNON: Great catch.

ROSEMARY: So by taking the qualitative and quantitative data, applying my instinct and experience to interpret it, I'm always asking, what does this tell me about what we should keep doing, what we should do more of, and what we should stop? Stopping is the hardest part of that conversation. You've been doing it for a while, you've invested time. It's like a car you've owned. When do you keep fixing it, and when do you decide enough?

In addition to the AI assistance, I also have the luxury of sending my team across to all of our global offices to talk to employees, get the context, really understand by sitting in the room. That's where you get the deepness. Then AI helps us interpret it, and we apply our own instinct and understanding of the culture.

Each office has its own microculture. Each function has a culture underneath that. So we shift. We've not just shifted how we listen. We've also figured out how to work smarter and engage employees in new ways they resonate with. Because our engineers are smart, and they can sniff out an AI-driven message like that.

Employees Are Consumers Too

VERNON: It's interesting that AI has enabled you to focus more on the human connection with your employees, not just rely on sentiment from whatever tool you're using. You take the information from listening, put out the communication, and then physically visit, talk to people, see how they feel. That further validates whether the AI was off when it told you something about how people are feeling.

Communications is often responsible for chatbots or AI agents that get deployed. Are you doing any of that? Do you feel that should be leveraged in a communicator's role?

ROSEMARY: I don't think I quite understand the question. We're not responsible for the chatbots. We just take advantage of them being available to us.

VERNON: How do you leverage the chatbots that are used internally?

ROSEMARY: Like I said earlier, we distill the information down. We ask the chatbot, here's what we're hearing, here's what our survey results are saying, quantitatively and qualitatively. The bots help us summarize, and we apply instinct and experience on top.

We have another luxury. We're a B2C company, and we have amazing teams that listen to our consumers. Our UX team, our user experience team. When I came in, I quickly realized: our employees are consumers too. With my marketing background, marketing is about creating demand for a product. From my point of view, communications is creating demand for our messaging. We want people to open the emails. We want them to react to our Slacks. We want them to attend our events.

When I first came in, I asked, how do we save time from all this process and tactical work and get more strategic? I tapped into this amazing team of UX designers, people with PhDs in this. I said, hey, will you apply those principles and help me run focus groups, so I understand what our employees want and how they want it?

We applied UX principles to our communication vehicles and channels. Our open rates, for example, went from the 50 percent range to 60 and 70 percent depending on the messaging. Because we met employees where they were. It was a combination of what the UX team taught us about focus groups, and us continuing that work.

Headlines, Blurbs, Long Form

VERNON: Most of the success you're describing comes from human touch, not just leveraging AI. I was at a conference yesterday. A power company described how their linemen are out in the field, climbing poles, often in rural areas without signal. They literally publish a magazine they give out at job sites and mail to homes. It's a weekly piece, four or five pages, very well done, with a longer digital version. That reminds me of what you're describing. Hands-on. We need to talk to people. We need to know what they think.

ROSEMARY: One of the things we learned in these focus groups, when we redesigned our weekly communication, is that some people want to know the entire story. Some want the headlines. Some want something in between. So we redesigned that weekly communication: headlines, a quick blurb (we are very specific, 28 words for the blurb), and then a link to something longer.

It's been fascinating seeing the data on which stories get the double-click. Sometimes by whom. And then if you walk through the halls, someone says, hey, that story. It's amazing that we've actually met our audiences where they are. It is not one size fits all.

VERNON: Who clicks on what is something many comms teams aren't measuring. They miss that feedback. Can you do that through your platform? Is it AI-enabled in some way?

ROSEMARY: Like any enterprise, we keep everything anonymous. So I don't know the person, but we can distill it up to a function type, or a region, or an office. One of the things we found early on was, even when we were getting good open rates, it was mostly in the U.S. The consumer team would say, great, we're getting good open rates. I'd say, you, we are getting good open rates. But we're missing this growth area outside the U.S. Data only tells you one thing. You have to interpret what you're missing or doing well.

Selling This Approach Up

VERNON: How do you sell all of this to leadership? You're doing more human-touch work than I expected to hear in an AI episode. Leadership is often gung-ho on, oh, we need to do this with AI, why are you doing this personal-touch stuff? How are you balancing that conversation?

ROSEMARY: Actually, it hasn't been an issue. They're seeing the impact. They're seeing employee engagement go up, sentiment go up, especially as we venture into this AI-driven world. They respect the value internal comms brings. So that has not been a fight.

VERNON: That was my next question. I often hear from communicators, we feel like we don't have a seat at the table. Has that come up for your team?

ROSEMARY: In any job I've had, at any company, there's an element of not being in the room. But the really good communicators I've worked with through the years, I always call it being a fly on the wall of a room you're not in. What we do as communicators, with or without AI, we've always had to absorb what's going on without being directly involved. We do that through our network, our collaboration, our stakeholder conversations. So nothing has changed with AI about that. Usually, ninety percent of the time, we're ready to react or plan, because we have a pulse on the functions, the leaders, the stakeholders.

VERNON: I've run into a few people from your company at different events. The way you all communicate, I can tell. It's succinct. It's not chaotic. There's a real synergy between how comms communicates and how leadership receives it.

ROSEMARY: And we're a small enough company. As a company gets larger, this gets much harder. We have the luxury of talking to leaders on a regular basis, in hallway conversations, in meetings. That is a luxury.

When It Goes Wrong, and What to Stop Doing

VERNON: What happens when it goes wrong? Have you had anyone produce content or send something out where AI was directly involved and it shouldn't have been?

ROSEMARY: No, we haven't. We have processes and gates in place across my team. I've got a great leader who works for me with a great team we've been building up. We have enough checks and balances that we haven't dealt with that. Has there been a message we could have done better? Of course, because we're humans. In the celebration of that milestone I mentioned, when we thanked the teams, we missed a couple of teams in the shoutout. That was because we didn't know. So those are the kinds of things we tend to miss. But we haven't had anything drastic, like what happened with your assistant.

VERNON: Right. Luckily, it was insulated. As we're wrapping up, when you're mentoring the folks underneath you, what framework do you use to help them understand when to use the tools, and when to go by their gut and what their training is telling them?

ROSEMARY: Initially, when the AI tools were coming online and usable inside our firewall, I have a lot of early-in-careers. I love mentoring them. They're energetic, smart, they grew up in the digital world.

What was interesting was, we first had to break through a barrier. We had a session, we gamified it a little bit. What was startling about my early-in-careers was their initial reaction when they started using the tools. It felt like cheating. Because they had always grown up in high school turning papers in through turnitin.com, which checks for plagiarism. So we had to reframe how these tools could work to their benefit.

I have a great team. We're very transparent. We talk about the issues, the topics that are coming across at the company level. We talk strategies. My director, who is very experienced, talks with me about what we see and hear. We explain the back story. My early-in-careers get judgment that way. Every week we talk through our top priorities. What's the story behind them? Why are they priorities? We have gating and process. As they venture out and write their communications, their feature stories, their profiles, they know to ask in the areas they're not comfortable with, because they either haven't heard the context or they haven't been exposed to it. So I've been fortunate that, regardless of years of experience, the team understands the process. They understand who to ask. They're curious themselves.

Depth, Breadth, and Lifelong Learning

VERNON: What's one piece of advice you would give someone early in their career who's looking to move into leadership, that you wish someone had told you?

ROSEMARY: There are so many things. But the one I'll start with: depth and breadth in communications is more important than promotions. To really advance in your career, you have to understand the function. You have to go deep. Understand how engineering works. Understand how finance works. Understand how marketing works. Then you have to go broad. You have to become a generalist across functions. You have to keep doing it. Not just once. Throughout your career. You have to grow breadth and depth and keep doing it, regardless of what level you're at.

You need to do it in your early career, your mid career. I'm still learning to go for breadth and depth. I'm going deep in AI, just because I'm curious. I love technology. How can I use AI to the benefit of the team? Then I'm also becoming a generalist, understanding how each function works as we go through AI. How do you communicate to the different functions on their benefits and use cases?

You can only do that if you understand how they work, how they operate. You have to build personas. The most classic way to think about it from a comms perspective is, what does this persona need to think, feel, and do? That's the best way to create the type of messaging, and the transformation, the change management, that brings people along on the journey.

VERNON: That is amazing advice to close on. Rosemary, thank you so much for coming on the podcast. I really appreciate it. Amazing advice. A different direction than I thought we were going to go, but I'm leaning into it.

When we first met on a panel, I had this impression: she's brilliant. I have to get her on a podcast at some point. And this advice could honestly apply to anyone, any career. Depth over distance is pretty much what you said. Lifelong learner. Something a lot of people fall out of when they get comfortable in their careers. They stop learning. And when you stop learning, particularly in this environment, you stop growing. You covered all of that. Depth, lifelong learner, constantly growing.

ROSEMARY: Well, thank you for having me. I'm very passionate about communications. I'm passionate about technology. AI is a game changer.

VERNON: It is. I love the approach you're using. You're leveraging it as a thinking partner. That's what I tell people. It's a thinking partner. That's all it is. But you're maintaining the human contact and validating what's happening with the actual people it's impacting. That's the message anyone using AI could take away from this. Talk to the people who are being impacted by the output and the content you're producing.

ROSEMARY: Yep. That's it. Exactly.

VERNON: Amazing. Rosemary, thank you again for being on the podcast. I'm looking forward to when this episode airs.

ROSEMARY: Okay. Thank you, Vernon.

*Stories That Lead | storiesthatlead.co*