Google’s first Chief Decision Scientist, Cassie Kozyrkov, wrote recently that “The biggest challenge of the generative AI age is leaders defining value for their organization.” Among leadership considerations, she says, is a mindset shift, one in which there are “endless right answers”. (“When I ask an AI assistant to generate an image for me, I get a fairly solid result. When I repeat the same prompt, I get a different perfectly adequate image. Both are right answers… but which one is right-er?”)

Kozyrkov’s overarching conclusion is that confirming the business value of your genAI decisions will keep you on track.

In this episode, Neville and Shel review Kozyrkov’s position, then look at several communication teams that have evolved their departmental use of AI based on the principles she promotes.

Links from this episode:

The next monthly, long-form episode of FIR will drop on Monday, May 26.

We host a Communicators Zoom Chat most Thursdays at 1 p.m. ET. To obtain the credentials needed to participate, contact Shel or Neville directly, request them in our Facebook group, or email fircomments@gmail.com.

Special thanks to Jay Moonah for the opening and closing music.

You can find the stories from which Shel’s FIR content is selected at Shel’s Link Blog. Shel has started a metaverse-focused Flipboard magazine. You can catch up with both co-hosts on Neville’s blog and Shel’s blog.

Disclaimer: The opinions expressed in this podcast are Shel’s and Neville’s and do not reflect the views of their employers and/or clients.

Raw Transcript

Hello everyone and welcome to four immediate release episode number 4 63. I’m Neville Hobson. And I’m Shell Holtz reports on how communication departments are moving from AI experiments to serious strategy driven deployment of Gen AI are proliferating. Although I’m still mostly hearing communicators talk about tactical uses of these tools.

The fact is you need to start with strategy or don’t start at all. That’s the conclusion of Cassie. Kako, Google’s former chief decision scientist who warns leaders that Gen AI only pays off when you define why you’re using it and how you’ll measure value. She calls Gen AI automation for problems that have endless right answers.

Now that. Warrants a little explanation. Traditional ai, she says, is for automating tasks where there’s one right answer using patterns and data. It’s gen AI that automates tasks where there are endless right [00:01:00] answers and each answer is right in its own way. This means old ROI, yardsticks won’t work.

Leaders have to craft new metrics that link every Gen AI project to. Not just a cool demo. This framing is useful because it separates flashy outputs from real, genuine impact. With that in mind, we’re gonna look at a few comms teams that are building gen AI programs around a clear, measurable strategy right after this.

Well, let’s start with Lockheed Martin’s Communications organizations, which set a top down mandate. Every team member is required to learn enough gen AI to be a strategic partner to the business. They hit a hundred percent training compliance early this year. They published an internal. AI Communications Playbook filled with do and don’t guidance Prompt templates, a shared prompt library, and monthly newsletters that surface new [00:02:00] wins.

There are a few reasons that this is a worthy case study. First, the team generated savings. You can count, for example, a recent video storyboard project ran 30% under budget and cut 180 staff hours. The team has fostered a culture of experimentation. Uh, there’s a monthly AI art contest that they. Host inviting communicators to practice prompting in a low risk environment, helping them learn prompt craft before they touch billable projects.

And the human in the loop discipline is built into the team’s processes. Gen AI delivers the first draft or first visual. Humans still own the final story. The takeaway, Lockheed shows that enterprise rollouts scale when you train first, codify governance. Next, then celebrate quick wins. Qualcomm corporate comms manager, Kristen Cochran Styles said Gen A is now in our DNA.

Qualcomm’s comms team is leaning on edge based gen AI, running models on phones, [00:03:00] PCs, and even smart glasses to lighten workflows while respecting privacy and energy constraints. Uh, they have a device centric narrative. They don’t just talk about on debate on. Its comms group uses the same edge pipeline that it promotes publicly.

They have faster iterations occurring in their processes, drafting reactive statements, tailoring, outreach to niche reporters and summarizing dense technical research all happen at the edge, shaving hours off typical cycles, and there’s alignment of their reputation because they’re eating their own dog food from their own silicon powered AI stack.

Qualcomm’s comms team reinforces the brand promise every time it ships content. Let’s. Take a look next at VCA, uh, chain of veterinary clinics. One of them was the one that I take my dog to. Joseph Campbell’s, a comms leader at VCA and he’s echoed the strategy first mantra. He noted that 75% of comms pros now use gen [00:04:00] ai, but more than half of their employers still lack firm policies.

A gap he finds alarming. Campbell’s rule of thumb. AI can brainstorm and polish, but final messaging must. Obtain human creativity strategy and relationship building. VCAs approach involves sandboxing with teams practicing in non-public pilots before committing anything to external channels. Crafting guardrails is treated as urgent change management work, not paperwork.

So they’re developing their policies in a very deliberate way, and they have an ethics checklist. Outputs go through fact checking and hallucination screen steps just like any other high stakes content. Now these individual stories of teams employing gen gen AI strategically sit against an industry backdrop that’s moving fast with tripling of adoption.

Three out of four PR pros now use gen ai. That’s nearly three times the level from March of last year. Uh, and [00:05:00] efficiency gains are clear. 93% say AI speeds their work. 78% says it improves their quality, but speed. By itself isn’t value. Cassie Coser Cove’s Endless right Answers framework reminds us Comms leaders still have to specify which right answers matter to the business.

So let’s wrap this up with six quick takeaways for your team from these case studies. First, tie every Gen AI experiment to a business result. Whether it’s fast or first drafts, budget savings, or higher engagement, write the metric before you. Invest in universal literacy. Lockheed’s a hundred percent training.

Target created a shared language, a shared context, and without that, AI initiatives are gonna stall, codify, and update guardrails. VCAs governance, sprint shows policies can be an after, can’t be an afterthought. They’re the trust layer that lets teams scale gen AI responsibly. [00:06:00] Prototype publicly when it reinforces brand stories.

Qualcomm’s on device PR work doubles as product proof and keep humans critical in every example. Communicators use AI for liftoff, then rely on human judgment. For nuance, ethics and style communicators have next desktop publishing social. Gen AI is bigger than these. It won’t just make us faster. It will change how we define good work.

That’s why the strategic questions upfront, what does value look like and how will we prove it matter more than which model or plugin you pick. Good insights in all of that. Uh, shell, I guess the first thought in my mind, it makes me wonder how do those who argue against using AI and, uh, what, what’s prompted that thought as an article?

I was reading, uh, just this morning about, uh, an organization where the leadership don’t prohibit it. No one uses AI [00:07:00] on the belief that, uh, it doesn’t deliver value, and it minimizes the human excellence that they bring to their client’s work. I wonder what, uh, they would say to things like this, because there are examples everywhere you look and you’ve just recounted a load of the advantages of using artificial intelligence in business.

I was reading one of the other articles that you shared, which you didn’t talk about on the examples that Mons, uh, which is really quite interesting, itemizes, how they, how AI plays a large role in their marketing, uh, for instance, to create digital advertising content. Product display pages, uh, towards high level creative assets including social media content and video ads.

They talk about though the 40 ai augmented campaigns that they have implemented, which they say have led to measurable improvements in brand awareness, market share, and revenue. And that compliments all the examples you were saying. They also say, rather than replacing humans, AI assist the, in refining their ideas and generating content.

The key role of humans is to ensure brand distinctiveness and [00:08:00] originality. That simple. Those two simple phrases really resonated with me because AI assists the humans, and the key job of the humans is to ensure brand distinctiveness and originality. And that to me is, makes complete sense. So, uh, AI delivers significant value and they talk about the, uh, the metrics they have.

Uh, here’s a one, uh, they say when start delivering two.

And if you can do that 1% better, that adds up to significant volume gains and significant growth in terms of net revenue. Then, then it’s just the beginning and AI is delivering that according to, so these, these add to the, to the, uh, collection. Of, uh, what I call validation points for the benefits of using a particular tool, particularly when you focus on the human element in it.

So they’re all great examples. Uh, and I think you, you mentioned at the start that too much of the, uh, activity we hear about is focused on tactics, [00:09:00] and this is full of it. It links it all to strategic aspects. Uh, it’s not just the, uh, the improvement in this and the 250 trillion impressions, although that’s pretty extraordinary.

It seems to me these are real learning insights that you can get from all this kind of stuff. And, you know, I love reading all this stuff, so it’s good to see it. I have to say. I, you know, in communication we talk about strategic planning as a core competency in the profession and IABC conferences and in textbooks, the strategic planning process is outlined repeatedly.

I mean, there are, are are different models and different approaches, but it’s always based on what is it that you’re trying to accomplish. At the end of the day, you’re not trying to accomplish writing a good headline. Right. You’re trying to accomplish, uh, having somebody read the article because it had a good headline and walk away ready to buy your product or ready to vote for your candidate, or [00:10:00] whatever it it may be.

And it seems like. Even though we have embraced this as a profession in general, we have by and large forgotten it when it comes to Gen ai just because we get so excited by the immediately evident capabilities, the ability to gimme five headlines in different styles. So I can. Pick one or, or adapt one to, uh, to, to, to what I wanted to say, create this image.

I mean, there’s nothing wrong with that. These are all great uses of the tool, but ultimately we have to look at where it delivers value that aligns with the goals that we’re trying to achieve on behalf of the organization. And you talk about those organizations that say there is no value. I, I would suggest either they’re not looking, they have a, a bias against it at the leadership level.

Or they have people at lower levels who haven’t figured out how to demonstrate that value, and therefore leaders are convinced that there isn’t any. But if you look at the examples we’ve shared here today, it, [00:11:00] it’s clear that you can align what you’re doing with Gen ai. To your organization’s business goals and your strategic plan and your business plan and the like, there’s, there’s, there’s no question that you, you can, uh, the question is why aren’t more people doing it?

I completely agree with the decision scientists from Google’s belief that if you’re not being strategic about it, why are you doing it at all? Yeah. I mean, I think to me the, the key thing to keep remembering, and this could well be the kind of circling point you come around to, to repeat together again, as Mondelez says, while AI has been a game changer for them, it takes human ingenuity to get the most out of a technology that is available to everyone.

And that, uh, is a point you mentioned from one of the examples that you gave that, um, how AI. Augments as opposed to replace or instead of that people talk about. Sure. But this needs emphasizing, I think, in a much, much bigger way. So Mondelez says, uh, again, a real simple point, but it’s, it’s good to say it.

They [00:12:00] think AI is gonna help you do everything from creation of the brief all the way to actual actually trafficking the effort and putting it out into market. It’ll help you. So, um, that bears repeating, it’s not gonna do any of, all of that or any of that. It’s gonna help you do all of that. Hence, you know, AI augmenting intelligence.

And I saw another different use of that phrase the other day, which has escaped my memories. Obviously wasn’t very memorable, but it was another example of it’s the human, that’s the key thing. Uh, not the technology, the technology tool that enables these things. So people’s eyes roll my view, leadership.

No. And I think if leadership is going to pay attention to this in a way that is meaningful to the organization, there has to be an effort to bring managers into the loop to, so that managers can help their employees feel good about this. Understand, and we’ve talked about the role of the manager here before.

Yep. But this, this is a critical one, is the emotional [00:13:00] side of managing. When you have a team of people who are confused and distressed and, and maybe worried about their futures with ai to be able to assuage those concerns and pull people together into a team that works with these things so that they do deliver that value, that’s going to increase the value of that team and of those individuals.

So there’s a lot of work to be done here, and it’s heartening to see organizations like VCA and Qualcomm and Mondelez doing it. Well and doing it right and, and the more these case studies we can see, the easier it’s gonna be for other organizations to basically adapt those concepts. Yeah, I agree. And on the case of, on the part of Mondelez, the article was published in a publication called Knowledge at Wharton from, uh, the Wharton School University of Pennsylvania.

I was quite at the end of April. Uh, I was actually quite amused to see the final text at the end saying that this article was partially generated by AI and edited with additional writing by knowledge at Wharton [00:14:00] staff. Curious about what the additional writing is. Uh, but that there, I would argue that’s a simple but good example that’s fully disclosed of the role AI played in them.

Being able to tell that particular story. I don’t think that diminishes anything. If anything, it’s additional to it, hence the additional. Uh, in the, in the, I was gonna ask, did you, did you find the article less readable because it was partly written by ai? Well, now I know that. How could I tell? That’s the thing.

They disclosed it and, uh, it’s good for them. I don’t think they needed to do that. Again, it depends on how they felt. They don’t say what percentage of the additional was AI generated, but I would imagine, again, a good example. To me, it seems that you’ve got something that you wrote and you running it by your AI assistant to check for.

The flow tone, all those things you kind of do. With Grammarly a bit, I think at the very least, if you’re using Word, you can use the grammar checker and all those tools in there. Not very good. Nothing nearly as [00:15:00] good as an AI tool to do these things. So that’s already with us and has been for quite a while.

It’s getting better, but the human element is absolutely critical. So it would be interesting to know what that additional writing was said, but it’s a good example. It is. And that’ll be a 30 for this episode of four immediate release.

The post FIR #463: Delivering Value with Generative AI’s “Endless Right Answers” appeared first on FIR Podcast Network.

(Short Cuts) Reckoning With Reckoning