How Rep Cap uses generative AI to go fast, ensure compliance, and work like a Trekkie
a conversation with Chief Executive Officer Mary Ellen Slayter
AT A GLANCE
Days
Campaign delivery speed
50%
Reduction in production time
20%
Less client review time
LISTEN TO THE INTERVIEW
WATCH THE CLIPS
Rep Cap headquarters
Baton Rouge, LA
Industry
Professional services
Department
Marketing
Key features used
Snippets Ask Writer Style guide
Listen to the Rep Cap story or read the (edited) story below.
Writer is the full-stack generative AI platform for enterprises. We make it easy for organizations to adopt AI apps and workflows that deliver quantifiable ROI quickly.
Mary Ellen Slayter is the chief executive officer of boutique content marketing agency Rep Cap. She’s a soil scientist-turned-journalist-turned content thought leader who believes in battle-testing assumptions, and makes ethics a key pillar of her decision-making process. Read or listen to the story of how Mary Ellen and the Rep Cap team use Writer to transform work.
Tell us a little about you — your background, philosophy, and work style. You studied agronomy, math, and journalism?!
I’m Mary Ellen Slater and I’m the CEO and founder of Rep Cap. We’re a content marketing agency that I founded 11 years ago after a career in traditional journalism. I worked at SmartBrief, which is an email newsletter publisher out of DC, and before that, I worked for the Washington Post, where I was a financial editor and also wrote a career advice column.
My undergraduate degree is in agronomy, which is soil science. People often ask me what on earth that has to do with journalism or marketing. And the answer, it turns out, is everything. Because I learned to think in a really holistic way, where if you change one thing in a system, it affects other things. That is a philosophy that has guided the way that I think about content, the way that I think about my life, the way I think about how I run my company. We don’t make decisions and choices in isolation, ever. I always try to think about the broader impact of things.
The second thing that has driven my career is I took statistics. That was actually the reason I got a job as a journalist at the Washington Post at such an early age — because I passed the math test!
I’m very comfortable with talking about revenue, pipeline, and financial impact, and using those numbers and analytical skills to shape how I think about content.
Rep Cap is a boutique content marketing agency that specializes in HR, technology, insurance, and financial services. My goal is to help risk-averse brands sell to even more risk-averse buyers.
The CFO, the CHRO, the CTO — these are not people who are out there trying to take chances.
The Rep Cap homepage
What led you to pursue generative AI? What challenges were you trying to solve for your clients?
I’ve been super psyched about generative AI for a long time. One, I’m a sci-fi fan, so some of the conversations that we’re all having around here, they’re like, “Who would have thought this would have happened?,” and I’m like, “It’s an episode of Star Trek!”
I look at generative AI as a way to take the knowledge and the experience that I have, and that the team has, and do better work at scale with it.
My worry right now is there are a lot of people who think, “Aha, I can make hundreds of blog posts over the weekend in my basement.” But that is not going to work.
I’m good at coming up with ideas. I’m good at making connections and building relationships. I suffer from blank-page panic just as much as anybody else. I’m an editor more than a writer. If you give me stuff, I’m very good at shaping it into things. Generative AI gives me a way to take things that are generated by those other pieces that I mentioned, like I’ve got interviews with people, and if I can take that stuff and use AI to get me to turn that stuff into content that can be consumed faster, that’s what’s interesting to me.
If I put a lot of love and a lot of work into some big research report or some big asset, that lets me get more life out of that. It lets me extend the distribution to more channels and over a longer period of time and not be super time-consuming.
We are deep experts in what we do. I know that there is this idea that there are lots of writers out there and lots of people who make content. There are not a lot of writers out there that like to write about hedge funds or parametric insurance. So I actually have hiring challenges. I have to find people who are not only interested in the type of content that we make, but who know enough about it to meet our brand promise.
AI allows us to produce more content and serve more clients more deeply than we would be able to using the expertise we have in house. My content team jokingly calls Writer the fifth member of the team. It lets us scale our expertise.
“A human editor may or may not have had enough coffee that day. Writer has always had enough coffee.”
Mary Ellen Slayter Chief Executive Officer
What prompted you to select Writer? Did you have any particular requirements?
My buyer is the most risk-averse buyer on earth, and they sell to other very risk-averse people. The criteria were security and privacy. I work with enterprise financial services, HR, technology, and insurance companies. What that means is that I cannot be taking their content and feeding it somewhere in the hopes that it doesn’t show up in someone else’s messaging.
With Writer, I’m able to go to my clients and go, they are SOC 2 compliant. And they’re like, “Acceptable.”
This is going to sound a little esoteric. It was very important to me that the leadership of the company take ethics seriously. It is the old Spider Man quote: “With great power comes great responsibility.”
We’re putting very powerful tools in people’s hands, and they may not understand them. I absolutely evaluated the leadership teams for each of these products, and the two tools that I use are both led by women that are upholding the highest standards.
Let me put it this way – I know that if something goes wrong within Writer, the first person I’m going to hear about it from is May.
“My content team jokingly calls Writer the fifth member of the team, it lets us scale our expertise.”
The primary use case that Rep Cap uses Writer for our clients is in taking big pieces — big cornerstone assets as we call them — and getting more life out of them for campaigns over a longer period of time.
Two, we produce that content way faster. In the past, you’d have some big white paper and you’d have to get a junior writer to go through and say, yeah, let me go and I’ll find some social clips and I’ll make all this stuff. I mean, that could take a week by the time you wrote it, got it edited, got it packaged up, and then sent it over. And I had to charge accordingly! It allows us to get things into market faster. If you think about that from a campaign standpoint, the difference between launching a campaign in a day or a week — that’s huge. That can make a big difference in very competitive markets.
I record every call that we’re on and produce a transcript. We use Otter, which we found to be reliable and consistent. As a human, certainly I can go through and read the transcripts and see if I can come up with some ideas. But what I now do is take the transcript and give it to Writer. And Writer tells me what’s in there, and I look at it and I go, “Hmm, that’s interesting. I didn’t even think about that part.” It’s not a replacement for my brain, but it allows me to start to pick up some ideas. I can also take that same content like a transcript and I can say, “Writer, write me a blog post based on this.” And I get that first draft and then I can start tearing it apart.
In my world, campaigns are maybe not as agile. We go through compliance, we produce things that have to get reviewed. I’m often writing about things that haven’t been publicly announced, for publicly-traded companies. We’re very careful. What I do not want to do is take that product copy or that deck from that product and feed that into somebody’s random LLM. Maybe it’s a small chance, but it is an unacceptable chance for my buyer. I think for other industries, it’s not such a big deal. My buyers? It’s a big deal.
One of my favorite things in the Writer platform is claims detection. People have been writing perfect B.S. for ages, and nobody had a claims detector. Like, “How do we know this? Can you prove this?” We were making claims long before AI was making claims and just letting them slide on through, so I like the claims detector also for human writing.
“My buyer is the most risk-averse buyer on Earth. I cannot be taking their content and feeding it into somewhere in the hopes that it doesn’t show up in someone else’s messaging.”
For our major clients, Writer instances are trained around their brand.
With the content we’re generating, the fewer humans that touch it, the better. So, I’m able to take the long form and create the short form triggers. We’ve got all kinds of things that we flag, like forward-looking statements. Even compliance goes faster because we caught all the forward-looking statements before it ever went to compliance. Writer catches that for us — not a human editor, who may or may not have had enough coffee that day. It’s really easy to overlook that stuff. So, Writer actually has fewer mistakes. Writer has always had enough coffee.
If you’re gonna use generative AI, it’s actually coming up with your pilot. The intentional project that we did was an SEO blog post, and we knew how much time it would take because we’re an agency. We tell writers we expect you to complete this task in roughly an hour and a half.
It wound up being more like 45 minutes. That’s a big deal. That’s like a whole person.
Writer writes clean copy. Yeah, it can be nonsense, but it’s clean. So the editor found that he could spend more time paying attention to the substance of the content. It went from being an average of 30 to 45 minutes per 1,500-word post to more like 15 to 20, and his type of editing substantially changed. There was more, “Why are we saying this? Is this the right source for this?”
The best proxy I have for measuring that is that when we send it to the client, how many revisions are required, and what is the extent? That also went down. Not as dramatically as that front-end stuff, but I’m okay with that, because that’s not even what my goal was.
For the majority of our clients, that number also went down on the pieces that we produced using AI declined by a fifth. It’s literally a whole person.”
“Even compliance goes faster because we catch all the forward- looking statements.”
There are two reasons I chose Writer: one was about security, privacy, and compliance. And the other one was about ethics and governance. And the ethics and governance part is very important to me. One of the things that I think gets lost in the conversation about AI and generative AI particularly is people are very concerned about tactical things about how to use the tools.
Tactically, these tools are really easy to use. I can teach somebody how to be a good prompt engineer in about 60 minutes. The hard part is not figuring out how to use it. It is figuring out when to use it and why to use it and when not to, and to put guardrails. We have a content manifesto at Rep Cap that I share with our clients. It’s important that, for any organization, for any marketing leader who’s going to use this in their work, I think you need your own manifesto.
You need your own guidelines for when you’ll use it and when you won’t. For example, one of my rules is, AI creates, but only humans publish. Under no circumstances will we automate a thing where AI generates content and it automatically publishes.
My other rule is to uncover your blind spots; don’t amplify them. Because AI, if you’re not really careful, will take the worst and the most basic part of whatever people think, and it will just reinforce that and sound very confident doing it.
“The difference between launching a campaign in a day or a week — that’s huge.”
Mary Ellen Slayter Chief Executive Officer
AI absolutely has a bias problem. It has a bias problem in HR when I think about using it for hiring. And that is going to show up in our content, too. Writer has built guardrails for this. So one, develop your own manifesto, like sort of your own guidelines.
Those things that exist were created by biased humans. If you sit down and you write a job description for an engineer, in any open model, the language is most likely going to assume that that person is male and is going to use a pattern of behaviors. It suffers from bias. Not just bias inside the organization, but societal bias that tells girls that they’re not engineers from the time they’re little.
Let’s not do that. So, if you don’t want to do dumb stuff faster, think about your audience that you’re writing for. When you write your prompts and you’re like, “Write a series of emails targeting heads of IT at tech companies,” it’s going to write language drawing on assumptions that are based in these societal biases.
As marketing leaders and as serious content people, we have to look at that stuff with a very close eye and say, did I just make something that amplified the biases that were already in the system or am I making something to counter it?
One thing I appreciate about Writer is you do have inclusivity flags set up, but sometimes those things can be subtle. And that goes back to why you don’t want to just hit publish.
