How Everyday PR Pros are Using AI to Shave Hours Off Their Workflows

Doug Downs (00:05):

In the late seventies and early eighties, Bryant Park in Midtown Manhattan was a notorious hotspot for crime and drug use. Everyone avoided it and that only led to more neglect and deterioration. The city needed to transform Bryant Park into a safe, inviting and vibrant public space, but they'd already tried conventional urban planning. Now that wasn't going to do it. They needed something different. The City hired William H. White, an urbanist and sociologist to study the park and develop a solution. White's approach was rooted in cultural sensitivity and appreciation for human behavior. Instead of a top-down solution, he first spent time observing how people use the park. He saw how people interacted with the space, where they felt safe or not safe, and what activities they took part in. He engaged with residents, workers and visitors to gather insights and opinions on what they wanted from the park.

(01:00):

Then based on his observations and engagement, he proposed several design changes, including movable seating that enabled people to create their own social spaces, fostering a sense of ownership and comfort, visibility and access. He recommended opening up the park's, entrances and removing tall hedges that block sight lines that made it feel safer. Activities and amenities. He emphasized the importance of diverse activities and amenities to attract a wider range of people. Within 10 years, the park had become a vibrant, safe and popular destination for New Yorkers and tourists. This is an example of a very human solution. It's about observation, engagement, empathy, creativity and interpersonal skills to solve complex social issues. AI can't do that. Not yet. Anyway. Today on stories and strategies, there are things PR pros can do that AI can't like human relationships and crafting truly creative strategies, but there are also things AI can do that make our day a walk in the park. My name is Doug Downs. My guest this week is Michael Smart, joining today from Alpine, Utah. Hey Michael.

Michael Smart (02:23):

Hey, Doug. Thanks for having me.

Doug Downs (02:24):

Great to have you. And Alpine, how many miles south of Salt Lake City, just so I can sort of position

Michael Smart (02:30):

Myself? Yeah, it's about 30 miles south or as you might say, what would that be? A little bit 15 or a little bit less than 60 kilometers.

Doug Downs (02:42):

Am I getting Yeah, Canadian terms? Yes.

Michael Smart (02:44):

Yeah, yeah, and I was on top of 3,200 meter mountain yesterday. That's 11,000 feet, so that was a milestone for me. I hiked up there.

Doug Downs (02:55):

That sounds beautiful. And the real Canadian way to say 60 kilometers by the way is about 10 Tim Hortons and then every Tim

Michael Smart (03:03):

Hortons. I get that. I've seen enough Tim Hortons to get that reference.

Doug Downs (03:08):

Michael, you work with large and small PR teams to earn more media placements and level up their writing. You've trained teams from Salesforce, Zillow, ally, and the University of Toronto, ding, ding, little Canadian reference and others. So let's dig into it. What are some of the specific examples from brands and agencies where PR pros are getting more done or going home early thanks to the use of AI?

Michael Smart (03:36):

Yeah, Doug, this is still an emerging tech of course, so I'm not claiming that this is widespread, but since I've started training clients on this in January of 2023, some of the early adopters who've powered through that adoption curve where you're going to maybe fail as much as if not more than you succeed, some of those are doubling their output. I'll give you one example from a big Fortune 500 brand that if I named everyone would've heard of, I asked her to go back and look at her output, her productivity before January of this year. We're taping this in 2024 and after, which is when her application of the AI skills we've talked about really took off, and the way that she did her workflow is on the four months prior, she put out three press releases and two, what they call market reports. Think of it as like a detailed corporate blog post with a lot of data, and then in the same time period after adopting ai, she was surprised to tell me that she doubled that output. So six big press releases and four market reports. The Go Home early reference makes me think of an independent PR pro. So she runs her own agency, just her, and she wrote to me and said, yeah, I had this client request, wanted me to do a video script. It's not something I normally do, but I took output that I had created for that client before Add chat GPT, format it into a video script. It took three hours less than I thought it would take, so I left early and went sailing.

Doug Downs (05:22):

Oh, nice.

Michael Smart (05:24):

Yeah, she's in the San Francisco Bay area.

Doug Downs (05:28):

And when we say use ai, is chat GPT kind of it or are you hearing about other tools and other software folks are using?

Michael Smart (05:36):

Well, we're taping this in early August and the word on the street is that Anthropic's model, Claude 3.5 sonnet that they just released last month has slightly overtaken chat GPT four and four oh overtaken in terms of the type of work that most PR and marketing Pros would do, creating content and copy with text. The assumption is that open AI won't let that linger, that they'll come out with either chat GT five or some sort of improvement pretty soon because they don't like being number two. Oh,

Doug Downs (06:13):

No kidding. And just dropping a value nugget right there, right at the top of the episode. That's beautiful.

Michael Smart (06:20):

We aim to please, you don't bring Slouches onto your podcast, so I have a high

Doug Downs (06:23):

Part to live up to. Nope. You pass the vetting process, I must admit. So alright, we get more done. My experience, especially working corporate side is my reward for getting more done is

Michael Smart (06:35):

More work.

Doug Downs (06:36):

I get more work. Yeah. So how does that help?

Michael Smart (06:40):

Yeah, great point. So I mean, I could talk about this way longer than the time allotted and that is how PR pros in particular have. We've backed ourselves into a corner by assigning the value we delivered. We do this in our own head, correlating that value with the amount of time and effort it takes to create the value. So in the long term we're going to have to make a big shift to value-based billing from vendors. We can set that aside and come back to it. You asked about it in a corporate context. You got two options and I'll cite a study that Boston Consulting Group did in conjunction with some Ivy League schools like Harvard and also MIT. They had over 700 fancy white shoes consultants and they tracked them doing fancy consulting tasks. Half used chat, GBT half didn't. The chat GBT users had about a 37% productivity gain and produced higher quality work.

(07:44):

So when I talk to my clients, I talk about what are you going to do with your extra 37%? Are you going to set boundaries and maybe you work from home some of the week and you're going to keep delivering to use your exam, keep delivering what you're expected to meeting boss's expectations, earning your same salary, providing the same value, but this week maybe knocking off two or three hours to watch the Olympics or what most people say is they're going to deliver 137% and the key is there. You got to make sure you get rewarded in terms of compensation. The assumption being if you're providing 37% more value to your employer than you deserve to be compensated accordingly.

Doug Downs (08:28):

Let me ask a question in my real, really simple Dougie terms, if I before, was it November 20, 22, 20 22, right before that, if I had a client for whom I wrote a newsletter or blog article once a week and it took me two hours, did some research, did some writing fine tuned it two hours for a decent really good blog post, if now it's taking let's say 20 minutes because I'm ingrained into the company and I can inform chat GPT on how to write it, do I charge them for 20 minutes? Do I charge them for two hours or do I charge them this value-based hybrid and how do I come to the conclusion on what that is?

Michael Smart (09:14):

Yeah, I'll answer that kind of practically so your listeners can apply their mindset and then I'll answer it the way I recommend the best practice and where I believe that we're going to be all ending up, but not for some years. If you really look inside yourself and you were raised with a Victorian work ethic that you equal the amount of effort and sacrifice that you put into things, you're probably going to feel guilty charging the same rate if it takes you, sorry, I forgot the numbers you gave, so I'll just say it takes you 25% of the time. Whereas in reality, the way business works, if you work for a small enough business, you just need to charge what the value is and the value ought to be the same if it's the same quality to the end user. I'll use the real life example.

(10:11):

I am a former newspaper journalist. I am a terrible visual communicator, so I use a designer to help me with my slide decks. I have a desired outcome in my head when I send something to him. We've been working together for years and when he delivers me that desired outcome, I'm happy and I expect that I would pay him the same as I did last year. I don't know or care if he's using generative AI to do that in the same way that I've never cared that my accountant uses software to do my taxes. That's right. I don't know how to use that software and I wouldn't know what the right output, I wouldn't recognize the right output in the long term. I think we as PR pros need to change our mindset from a work ethic based one to we get paid because we know the difference between the good output and the bad output. That's right. Because your client, my client's going to be able to generate a newsletter article by themselves faster than it would take to even email us and ask us to do it. What they're going to need to pay us for is knowing the difference. So

Doug Downs (11:16):

With this 137% output that you referenced, we're getting more done because we're using ai, the adoption rate by public relations and marketing pros must be absolutely huge. We must all be piling into AI and just staying in there because it gets so much more done. Am I right with that?

Michael Smart (11:36):

You smelled my hesitation. I was about to raise my hand and say, Hey Doug, I'm being a cheerleader, but we got to also put on the brakes and talk about the real world. No, your question is apt because no, that's not happening. In fact, it is. The Michael smart of a year ago would be very surprised by that. But the Michael smart of now realizes human nature is we're not going to take the most people believe eight to 10 to 12 hours of hitting our head against the wall and failing that it takes to really learn what these tools can do and what they can't. It's not intuitive, and I'll give some examples. It's not intuitive. You think it's good at this one thing, you think it'd be good at this other and it's not. And so it can be very frustrating and I get that most PR pros who don't get paid to experiment with new technologies and train other people on them, which that's my job.

(12:36):

Most PR pros are just trying to get their job done. So most of them heard all this hype. You mentioned November, 2022 that spring, the next year the hype was this is amazing. And when we tried it out, it sounded like a 10th grader on amphetamines we're like, there's no way I could turn in this DRL at work. And so I get that. A lot of PR pros after that when they hear people me talking about, they're like, no, they're overselling it. It's not really that good. The reason I'm so bullish on it is it boils down to two principles. One, the technology has improved a lot CHATT three compared to CHATT four oh, night and day. By the way, the best models are now free to use, used to have to pay the 20 bucks to get the best model. You can use it free, just a throttle, limited number of prompts a day. And then the other reason why I understand where those pros are missing is they're missing just a little bit of some training and tutorial and knowhow and maybe we can get into that if you want.

Doug Downs (13:45):

Yeah, let's do that because is it a resistance to the initial learning? And I empathize with that. I'm 57 going to be 58 in a short period of time. There are lots of things where, you know what? I don't want to DIY this. I don't want to look on the app to figure out the myriad, the labyrinth of steps I need to get this done. I want to pick up the phone and talk to a human and understand how I'm supposed to negotiate my way through this and have them talk me through it. That's faster for me. I get that the company doesn't want to do it because it costs them money. So is that what we're seeing is this initial resistance to the up curve in learning on how to use AI properly?

Michael Smart (14:27):

I might've said resistance, and if I did, I shouldn't because it's more just inertia. I've, I'm good at my job. I've done it this way for 5, 10, 15, 20 years. I looked at this and it didn't help me do my job better. So why should I take those eight to 10 hours until I see dramatic differences? If it's okay, Doug, I'll give a couple examples where folks maybe could without any risk and with a very limited time investment, see some differences, some improvements, and then I'll tell you guys some examples where you could see some dramatic differences. So by far, the two most common use cases where I see PR pros applying generative AI effectively, the first one is headlines and subject lines, and this avoids two of the biggest risks that PR pros are worried about, one of which is hallucination, and the other is some sort of copyright violation.

(15:29):

I don't really get that second one, but it's out there. It's really common. And so if you paste in the content of your press release or of your draft email and ask for five or 10 proposed headlines or subject lines, it's going to draw from that content right there. It doesn't have to go out, it won't make anything up. It doesn't have to pull from anybody else's ip. The second most common use case is reformatting or repurposing. So one example, I had a PR pro that did a press release that advanced an event, it will happen on this day and you can come at this time. And then she wanted to repurpose that into a blog post about it after the fact. It's not a heavy lift. You or I could do that pretty quickly without a lot of brainpower, but the AI does it instantly and you just skim it once and it's done in less than three minutes.

(16:20):

Other repurposing is taking a press release or a blog post and asking for five versions of LinkedIn posts based on that. So those are all quick and easy. Where the advantage comes is when the PR pro, instead of just asking for a headline, it says, you could challenge it and say, give me five headlines that will challenge public relations professionals who subscribe to my newsletter. Give me five headlines that would reassure public. You see, the more precise and specific, the more variants you get, the more you can choose from. And then if we go down, if go up the learning curve all the way, I've got some PR pros who've built custom, which is a lot easier than it sounds. Yes, that's what we've done. Yeah. It's basically like a saved prompt that pulls from examples that you give it where they've given the custom GPT 1115 examples of them writing a case study for a certain client in a certain voice with some instructions that they developed over a period of an hour or two of trial and error, and then they just paste in the transcript with the client interview and then GPT writes a thousand word case study does everything else.

(17:40):

Yep. Yeah, those are rare. Those are the exceptions, but that's where we're moving and that's where we're going to get. And so that's why the questions you're asking about value-based billing and all that stuff are really important for us to start thinking about. Absolutely. Because happen faster than we think it will. Well, and

Doug Downs (17:55):

I will admit in the creation of this podcast, a podcast is perfect. You have a 20 minute conversation. I've got how many 500 word blogs if I want to write them depending, did we compartmentalize our discussion? I've got how many social posts I can create, all kinds of things. And I use generated AI kind of as a work buddy where it does the initial drafts for me, and I'll inform my work buddy. Like you said, something that challenges PR professionals, something that will surprise make the first sentence short and compelling, that kind of instruction. But I use it work, buddy. You have an amazing newsletter. Do you do something similar in the creation of your newsletter? Because it doesn't read that way. It reads like I'm reading Michael.

Michael Smart (18:40):

Great question. I think a good way to answer that will be if I share what I've sort of determined out of these AI and PR user personalities, would that be okay? Yeah. Think of these as kind of avatars and the use case that you're talking about is a relationship building, and I'm going to call it one-to-one. Even though my newsletter goes to 14,000 people, I write it to one person and I strongly advocate against using AI for any type of relationship building because there's some sort of, you said it doesn't read that way, right? You can sense it somehow. You can. And even if I could fine tune the AI to write just like me, I would feel like that was inauthentic in a relationship building mode. Same with pitching. So before I started training a lot on ai, I trained mostly on how to pitch journalists.

(19:41):

I still do that and I strongly discourage my clients from using AI to write one-to-one pitches because that's a relationship building approach. So if one of your listeners is, does 90% media pitching today and they're like, I tried ai, it doesn't work. I don't want to, I would say you are correct. Actually keep not doing that. But let's look at a different user personality. This is somebody whose strengths in PR are other than writing, they're good. They're good at creating story angles, managing people or clients or strategic thinking. That person can benefit from using AI as a writing partner, just going back and forth so they don't have to stress about grammar and syntax and they can let their real strengths flow.

Doug Downs (20:26):

Okay, last question. It's flipping this around to the audience or stakeholder or listener or viewer perspective. Back in the mid eighties, I think it was, I saw Tom Broka talking about cell phones and the rising growth of cell phones and how they had enhanced reachability. And you can get reached on your cell phone now anywhere, and he made this sly comment, the rest of us are looking for a cell phone where you can't reach us at any given time.

(20:53):

And it was because we were starting to be inundated with constant accessibility, and we all live in a world now where I don't know how many emails you've got in the morning when you wake up. For me, it's 40 or 50, and I'm, I hope I can delete all of these and that they're junk, but I can't. Right? And I wake up and I'm already smoked with my day. Now we're generating, just to repeat that statistic, you used 137% productivity. Well, if we're all doing that, we're pumping out more content, more content, more noise, more noise from the audience's perspective, is this good? And how does that change the way the audience receives that noise?

Michael Smart (21:37):

We give you a reassuring axiom and a world increasingly dominated by AI generated content, you stand out more and more simply by being human.

Doug Downs (21:54):

That's true.

Michael Smart (21:55):

So I'll break down what you just talked about, your inbox. I don't know about you, but I use Google Workspace, so Gmail, and it keeps offering to synthesize my inbound email for me, and pretty soon they're going to have it wired right, that it's going to be able to reply to most of those emails For me, I don't know that I'm going to use that, but that's what's going to happen. The technological problem you just cited of an overwhelming amount of inbound is going to be attempted to be solved with another technological layer. And what's going to happen is somehow we're going to figure out ways to establish and maintain relationships outside of that technology. And that's how we'll stand out. Don't get me wrong, I'm not saying that generative AI cannot be creative. That's a big misunderstanding. It doesn't pull from stuff that's already been created. I could go on and say, I would like you to give me three ideas for a novel about, or as a buddy sitcom between a Canadian podcaster and a Utah based want tobe, mountain climbing PR trainer, and it would come up with totally novel concepts, right? The question is, would any of those be good and would it attract our target audiences? I think that's where the human intellect and creativity is going to stand out. In a AI saturated world,

Doug Downs (23:23):

Real connections will become realer.

Michael Smart (23:27):

Yeah,

Doug Downs (23:27):

Exactly. To sound like Dr. Zeuss, I hope.

Michael Smart (23:29):

Yes.

Doug Downs (23:31):

Thanks for this, Michael. Really appreciate your time today. You're

Michael Smart (23:34):

So welcome. I'm bummed I didn't get to say process while I was on here, so there we go. I just said it. Well,

Doug Downs (23:39):

Now you've done it. We processed that. Okay. If you'd like to send a message to my guest, Michael Smart, we've got his contact information in the show notes, personal recommendation, go to his website, sign up for his newsletter. It's a winner. Stories and Strategies is a co-production of JGR Communications and Stories and Strategies podcast. If you like this episode, please leave a rating, possibly a review. That's human interaction. It's real, it's organic. It helps the podcast. Check out our YouTube channel. Segments from this episode are available in video. Thanks as always, to our producer, Emily Page. And lastly, do us a favor again to this human thing. Forward this episode to one friend. Thanks for listening.