The Evolution of the Communications Engineer

Doug Downs (00:10):

The Red River runs south to north from the border of Minnesota and North Dakota up into Canada, eventually spilling into the Hudson Bay. When it gets north into Canada, into Winnipeg, Manitoba in particular, the topography changes and the potential for flooding becomes much higher. Back in the 1950s, there was a disastrous flood in Winnipeg, taking out four bridges and evacuating a hundred thousand people. One person was killed. Damages at the time were close to a billion dollars.

(00:46):

The leader of Manitoba Premier Duff Roblin championed a new capital project to build a floodway to carry water around the city of Winnipeg. But the cost 72 million. Now, people were divided on this. They felt the flood of 1950 was a freak occurrence once in a lifetime, maybe several lifetimes, waste of money. They called it Duff's Ditch. They said it was like building the pyramids of Egypt in terms of usefulness. The debate went on for years with Roblin, often feeling like he was alone in his argument, but he was convinced he was right. Finally, the project was approved and the floodway was built finishing in 1968 on time and under budget, and the Red River did flood again and again and again, but the floodway prevented severe damages each time. In 2023, Canadian studies suggest a floodway has prevented 40 billion in damages, and it's now designated a national historic site of Canada. People today still call it duff's ditch, not with derision, but with affection. Sometimes you need to have the conviction and the courage to be the only person in the room who's rights today on stories and strategies. We're now on the precipice of a new era in communications, one that will leverage artificial intelligence like never before, and those who made that shift years ago while the rest of us were sleepwalking are the ones in the front of the line now.

(02:51):

My name is Doug Downs. My guest this week is Aaron Kwittken joining today from New York. Hey Aaron.

Aaron Kwittken, Prophet (02:57):

Hey, Doug. Great to see you.

Doug Downs (02:59):

Well, and you're joining today from New York. So springtime in the big Apple, there's got to be a ton of things to do, number one of which is get the heck out of the city and go somewhere else, right?

Aaron Kwittken, Prophet (03:08):

Oh, go somewhere else. Or actually just take a lot of allergy meds. So I think we have another two weeks left for these trees to be dumping all this pollen.

Doug Downs (03:17):

Geez. Yeah, but at least the sunshine's around. Aaron, you're the founder and CEO of Prophet, the first generative and predictive AI SaaS or software as a service platform designed for the PR community. Prophet is the flagship of Stagwell Marketing Clouds comms tech unit where you're also the ceo. You've got 30 years in the industry, founded PR agency, K W T Global, where you're still chair of the board. You have your own podcast brand on purpose, and you are the immediate past president of the Public Relations Society of America, or P R S A New York chapter, which by the way is one of the most responsive chapters to our little podcast.

Aaron Kwittken, Prophet (03:59):

That's great to hear.

Doug Downs (03:59):

We love PRSA New York, they've been awesome. Aaron, when it comes to science and PR slash marketing, there is lots of science and up to date, most of it has been psychology and behavioral science side of things. That's not going away that that's definitely going to stay with generative ai, though bursting onto the scene, or as one client put it to me a week or so ago, that Chat G P T thing I've been hearing about, we're introducing newer science or sciences for using these technologies. Just at a high level, walk me through the world that we're kind of on the precipice of here.

Aaron Kwittken, Prophet (04:42):

I think you nailed it to date, not only is it kind of more behavioral science based, but we also, even though we don't always admit it, you know, and I have been using our gut and our instinct to provide direction around campaigns and advice based also on experience and expertise. And that's fine. I always found it quite frustrating in part because all of our clients, most of them think they're way more interesting than they are, and our marketing siblings have very specific data sets, audience segmentation to be able to prove out a concept for a campaign. Whereas we really didn't have much, maybe our charm, and in some cases, of course with me, my good luck. Just kidding. So I think that we've entered this era that I call a communications engineer, which is a mindset, not necessarily a skillset where we finally can actually use data to backstop our instincts and to also have, I think, deeper, more meaningful engagements on the client side, especially when it comes to crisis.

(05:47):

How often has it been where you're negotiating a statement or messaging with a general counselor, outside counsel from a client, and you say to them, is that PR advice or that comms, or is that legal? And he's like, well, it's like it's not legal. I'm like, okay, well leave it to me. I'm the expert. It's easier to have some data. So I really view what I think AI in particular is probably the most consequential technology development, not just for humanity, but for our industry. I view it in two sections, if you will. There's the predictive, using AI to predict trends, media interest surface, what types of influencers, podcasters, reporters might be interested in your pitch. And then there's the generative, which is the whole, like you said, the chat G P T thing. For the record, I think chat, G P T is amazing. I think it's a toy.

(06:37):

I don't think it's a tool. What I think the opportunity is for all of us is there are going to be tools like Prophet, for example, and there's so many others that are professional tools that are layered on top of the underlying technology behind Chat, G P T, which is open ai. And of course right now it's this arms race between Microsoft and Google whose large language model is going to be better. I think for communications professionals, especially the mid to junior ones, I think that they are going to be embracing more and more tools to be able to get things done faster and better. In my mind, it's performative, productive and predictive and all those mundane tasks, searching for the right reporter, doing research, putting together, be briefing books, bios, whatever, all that can be sped up. Things that used to take two hours can now take 20, 22 seconds and then a few minutes on top to come in over the top. Because we've always hired people over the years. We hope that they can get us 50, 60% of the way there. Now, I think these technologies can do that, and it's only getting better because the more people that participate in the Chat GPTs of the world, the more learnings and that's what that's happening is there's more training going on and it's actually for the benefit of everyone.

Doug Downs (07:52):

What's amazing. We did an episode with Stephen Waddington in the UK from the Chartered Institute of Public Relations, and it shared that there's already thousands of tools that fit within what I'm terming generative ai, but I really like what you just said about predictive analysis. Tell me how AI is going to help us on the predictive side of things, because I think that's really the number of it. And G P T being a toy, I love that terminology reminds me a bit of we have Bitcoin and then we have blockchain technology and people know Bitcoin, but they don't understand that blockchain is really the piece, but that's not what this episode's about. Yeah. Tell me more about predictive analytics and how AI is going to help us in the future with that.

Aaron Kwittken, Prophet (08:42):

It was the original premise behind Prophet four years ago, and many people still today probably thought that I was absolutely insane to leave my day job to move into AI and comms, and then I was talking about G P T two and nobody really cared. They all, they would roll their eyes at me and one of the frustrations that I had was people are just kind of downloading names of media from these databases that are largely analog and they're wrong. And then unlike you and I, we were trained to read and to research reporters and create relationships and what have you, people don't do that anymore. And people don't even speak to reporters anymore. PR people, they just email them. They span the shit out of 'em. So I was looking at different industries, particularly book publishing and film and also media, certain newsrooms like The Guardian, and I was noticing that especially in media entertainment, they were running scripts through machine learning and natural language processing and comparing those scripts against past books and movies to then determine whether or not these scripts have future commercial viability based on similar or adjacent genres.

(09:50):

I'm like, that's genius. So why can't we look at past media coverage to predict future media interests and sentiment? It's all publicly available. And as you know, and I think journalists would agree, most journalists have a very specific style and persona and way of writing. I know I do. You do. They don't really change that often. So it's a perfect case study for predicting media interest. And I think that, look, there are two things that we've been trying to solve for forever. How do I know who's going to be interested in my pitch, whether it's an influencer or podcaster or reporter, and how do I pitch more interesting? So how do I know who's going to be interested? That's the predictive stuff. How do I make it more interesting? That's the generative. And when you put those two things together, I think that it makes us way more performative as communicators way more performative.

Doug Downs (10:44):

You mentioned you started Prophet several years ago. It was probably four or five, maybe a little bit, bit longer than that.

Aaron Kwittken, Prophet (10:52):

Well, the idea was in my head probably for about 10 years,

Doug Downs (10:56):

Right?

Aaron Kwittken, Prophet (10:57):

Yeah. And the idea was really rooted first in crisis comms. Cause that's really, I don't have any superpowers, but that's what I grew up in is crisis and reputation management. And I always wanted a way to test the crisis statement before I ever had to use it, right? To determine would it accidentally accelerate coverage, would it deescalate? It would mitigate it. Oftentimes, most of them, they sit on a digital shelf somewhere, right? That's the hope because it doesn't see the light of day. And that was the original thought. And then when Mark Penn joined, then MDC Partners, the company that bought my agency back in 2010, he came from a private equity firm that basically that he built using some funding from Steve Bomber, about a quarter billion dollars from Steve Bomber, the former CEO of Microsoft. And that's where Mark worked as chief strategy officer.

(11:49):

And he formed this thing called Stagwell, which is a very technology, digital driven private equity holding company at the time that then bought agencies that were like-minded. He came in, they invested in mdc, he basically merged MDC and Stagwell, now it's Stagwell today. One of the first things he did was he said to all of the agency leaders, he said, come up with an idea that sit at an intersection of marketing and technology. We're going to do our own little shark tank competition and whoever wins 500 word essay and then eight minute pitch and whoever wins, I will award you up to a million dollars of operating budget to create the mvp, the minimal viable product. And this is back in May of 2019, almost to the date that we're speaking right now. Doug and I ended up winning, it was called Project Taylor, and for billions fans out there, it's named after a character in Billions who's this genius, clairvoyant, see around corners type person named Taylor who I'm like, wow, how do I hire a tailor for my agency and that Taylor does not exist in real life, so maybe I can try to build a tailor.

(12:55):

So it's called Project Tailor, it became Prophet, and then we actually used the name Taylor later on earlier this year to then name our feature, our generative AI feature. So we brought Taylor back, like kind of like Taylor, your Patch, it's like a double entendre if you will. So we built the M V P in 2020, soft launched end of 2020, and then really started selling the product, which is predictive at the time in 2021. And it was hard. It was hard up until maybe even I would say December, January of this year,

Doug Downs (13:29):

You must have faced criticism over it. You must have had, not only did you have to generate this idea on your own, there would've had voices of support, but there must have been banks not willing to lend you money for this. There must have been all kinds of criticism surrounding you on this.

Aaron Kwittken, Prophet (13:47):

So

Doug Downs (13:48):

Too many eggs in one basket, so to speak. Not that it wasn't a neat idea, but criticism that the human mind is not actually predictive, therefore we can't build software that predicts the human mind. That must have been the lines of the criticism. Well,

Aaron Kwittken, Prophet (14:03):

I'll tell you what the headwinds were at the time. The biggest headwind was a lot of the time the traditional media database companies have been forced, have forced our industry into a very complacent, very spray and prey mentality. Like I had said earlier, that was one kind of headwind. The second headwind was that we weren't there to replace a data media database company. We're there to make it smarter, to make that, to turn a media list into actual media targets. I don't believe in media databases, I believe in media targets. The other headwind wasn't money because we were funded internally. So that was great. It was being able to build the logic and find a data set that had enough integrity to be able to stand up to the results that we were showing on top of another headwind, which is this continuum of doubt.

(15:01):

And we're seeing, we saw that a little bit with ai. It is, you know, first start with doubt and then fear, and then kind of curiosity and then courage, and then adoption. That's that continuum of technology, especially in our industry. In our industry, we are skeptical people. We're trained to find fault just like the journalists that we pitch. So our first inclination is, oh, is de-risking. I am an optimist and I like to try to think about what's next. Years ago we thought about design thinking for communications, the difference between manifest needs and latent needs. Then we leaned into E S G and brand purpose, and now I think it's comms, tech, and PR. People are so good at telling stories for our clients, those digital and transformation stories. Yet we are like naval gazers. We don't innovate ourselves. Our greatest innovations are these processes or research.

(15:56):

That's it. There's no real tech behind it. And it's funny because we're in a very non-linear business. Inherently communications is non-linear, yet the tools that we use are very linear and not agile. So I'm just trying to create a different mindset. Again, this is communications as an near concept where we need an agility to learn how we can change our workflow and not in a way of automation, because we're a business of articulation, not automation, but in a way that makes us do things better, faster, and perf and better and faster and informed by data

Doug Downs (16:34):

More along the so-called customer journey and customer not being the right word here, but yeah. Right. So let's sew all this together. This concept of a communications engineer or a prompt engineer, someone who knows how to use the interface of a tool or specific tools like profit to the best of the tools capability to enhance their business goals. Let's imagine that we're trying to create, let's say, a news release to promote this episode. Just something tangible, walk me through figuratively how a prompt engineer would use Prophet, to set the tone, to find the right journalists, to create the message and actually create something tangible that does promote the episode the right way. How would that happen with Prophet?

Aaron Kwittken, Prophet (17:22):

So I'm glad you asked that question specifically because I've noticed since chat, GPT, all these, some agencies and some of our quasi competitors, because they're more me meaty database companies, they're like, oh, we built a press release generator. And I'm like, that is so insulting. Like press releases is like two, 3% of what we do. So we focused more on how do you take any piece of content and make it better, make it into a better pitch in seconds. In this situation though, and I actually just did this with my team because we're about to announce another partnership with a really cool company, that bat battles, that uses AI to battle missing disinformation. In this instance, you need human ingenuity as always and forever, right? AI's that going to replace that if the prompt, which is you, if you prompt the core, the key takeaways of the first, it's the whole first best only, what are the key points you want to make?

(18:17):

I would put in a headline and a subhead. I would put in a sentence or two. In our platform, you just need a minimum of 30 words. I like putting in 50 or 60 words, and that is the fallacy behind a lot of this Chat GPT stuff. The truth is that the more specific you are and the more words you put in, the better the outcome because it still needs you and you still have to have those core ideas. I call them islands of safety with my clients. What are those three key messages? Never leave the island. What are those three key points you need to make sure you put them in there? And then I'll throw in quote from Doug, quote from Aaron, and then I would hit generate and I would pick a tone that I want it to be persuasive, exuberant, emotive.

(18:56):

That might be a little weird. Contrite, no, because this is not a crisis professional possibly. I've got all these different kind of tones that you can pick from and they'll see what happens. Currently we don't generate press releases, but we could we'll put a template in there for that, because I believe that the majority of what we're doing is we're trying to create pitches, not press releases, and you can easily turn a pitch into a press release. It's harder to turn a press release into a really, really good pitch. So oftentimes people will load in like a thousand, 1500 word press release, often very technical. We'll be able to truncate that to a 250 word pitch using a very specific tone, and then develop, which I think is the most, which is really cool, is we'll create four social posts very, very quickly, all like 22 seconds,

Doug Downs (19:42):

And find the journalists in a micro media world that are most likely to respond to this. In fact, you can research journalists within your tool and get an idea of what they've covered in the past, the tone and sentiment of the coverage, the words that they've used. All of that's possible using Prophet

Aaron Kwittken, Prophet (20:01):

All of it's possible. And I think what comms people need to ask of their software providers is are you just using bullion search keyword search or are you using semantic search? The future is in semantic search. The difference is semantic search is kind of like a spider, the core is the word. But then we we're able to determine is that word being used in the right industry? I'll give you just a quick example. For years I worked with a company called Stryker, S T R Y K E R, huge medical device company, great company. There also happens to be a Stryker assault vehicle on the market. Totally different business. Semantic search can tell the difference between the two. Bullion search, which is what most comms tech companies use, cannot tell the difference. And this actually is one of the biggest complaints for media monitoring services because you know have to, I mean, humans should over always check the work of another human and definitely a software program, but oftentimes these monitoring services will get it wrong. They can't tell the difference between Stryker Med Tech or Stryker assault vehicle.

Doug Downs (21:05):

Last question, and fairly quickly within a couple of minutes. Sure. The future for those getting into the industry, whether it's PR or marketing, which sadly I'd lump under a big umbrella. We've always focused on a lot of the psychology going through school, getting into the industry, trusting our cut, as you said off the top, would it be valuable to have some computer science under our belts going into this? Is that what we mean by what a prompt engineer is likely to evolve to be?

Aaron Kwittken, Prophet (21:34):

My, and I still might do this, one of my biggest regrets is I never learned how to code. I think that in the same way that you have core curriculums around English and math, when you go to university there, everybody should have at least a base coding class, just like you have to take econ or accounting. Coding has to be part of who we are. And what's really interesting in our business is that we're not computing numbers or computing words, but you have to compute numbers to get to those words. And I think that everybody needs to know that. They also just need to be very comfortable with terminology and also have more patience in on the strategy side of comms. Not just so quick to be transactional and pitch, pitch, pitch, but think very deeply, use more calculus if you will. Not real calculus but thought calculus, which is what I think that, I think, which is what most people will appreciate when they're going through the new process of joining, whether it's an agency or in-house marketing or PR team,

Doug Downs (22:39):

As it always has been. But thank you for saying that we always need to plan and have goals in mind before we just start stepping into communications. You're right. Yeah. Aaron, I really appreciate your time today. Thank you for this.

Aaron Kwittken, Prophet (22:52):

Well, thank you Doug. I really appreciate this opportunity and again, I believe this is probably the most consequential and maybe the most momentous development in our industry. And to quote someone who, I have no idea who said this, if you don't use ai, it's not going to take your job, but it'll help you keep your job, and I think we all need to embrace that.

Doug Downs (23:14):

Excellent. This is right up there with TV and heck, the worldwide web, which also changed the game for all of us. Sure. Aaron, thank you. Thanks,

Aaron Kwittken, Prophet (23:22):

Doug.

Doug Downs (23:22):

If you'd like to send a message to my guest, Aaron Kwittken, you can email him. The address is in the show notes and ask for a demo of Prophet. By the way, once you see it, in action, it's pretty mind boggling. I really appreciated the demo and a chance to see it in action. Stories and Strategies is a co-production of Jgr Communications and Stories and Strategies podcasts. If you this episode to do us a favor, leave a five star rating that signals to others that this is a podcast worth listening to. Thanks for listening.