Former PR agency leader is a building software product focused on improving PR pitches with two different types of artificial intelligence
I’ve started to use ChatGPT to rewrite and suggest alternate headlines. Recently, I completed an article for a client and used the generative artificial intelligence tool to suggest alternate sub-headlines in the piece.
In the review process, my client was presented with two such headlines: one that I wrote and one that ChatGPT re-wrote. I thought my headline was better, but my client chose to go with the headline generated by the AI.
While I joke my writer’s heart broke in that moment, the reality is, ChatGPT took something I created for a client – and made it better – at least from their perspective. As such, it’s a good example of how PR and marketing professionals should be thinking about AI: it’s a tool that augments your abilities and enhances your work.
I kept thinking about this example as Aaron Kwittken demoed the software his company, PRophet, is developing. That seems to be the driving philosophy behind this new startup in the PR software market – and the backdrop for this new PR Tech Briefing.
1. Background on PRophet
Kwittken is a former PR agency leader. In 2010, he sold his agency to an advertising holdings company for an estimated $10-$15 million. Both companies have morphed a bit since but suffice to today the holdings company is known as Stagwell (NASDAQ: STGW) and the agency is KWT Global.
Stagwell and Kwittken are investors in PRophet – and KWT Global is providing PR support. This past week, Stagwell announced it’s building out a “comms tech business unit.” The unit will be led by Kwittken and PRophet will be the unit’s flagship product.
That context serves two purposes in this briefing: First, Kwittken is a bona fide public relations person that’s grown an agency, sold it and then stuck around to see things through. He understands the business of PR and the pain that current tech tools do not solve. By contrast, many existing vendors in the space don’t even have PR people on staff.
Second, PRophet has the backing of a publicly traded company that specializes in the field. Stagwell has experience supporting other similar software companies including the influencer marketing platform Koalifyed. In other words, the evidence suggests they know what they are doing.
The company began developing the product in 2020 and went to market in 2021 with its predictive AI. In March of 2023, it added generative AI to its software. PRophet employs about 10 people at this time according to the company’s LinkedIn page and website.
2. PRophet product overview – two different AIs at work
PRophet has two distinct AIs at work in its platform.
One AI is a predictive AI, which attempts to match the content of your release or pitch – text – with reporters likely to be interested based on their work. I’ll dig into that more in a moment. The underlying technology for this predictive AI is based on a combination PRophet’s own code and that of partners.
The second AI is a generative AI called “Taylor.” Taylor is based on a combination of technology from a company called Open AI, which made ChatGPT, and PRophet’s code.
So how does it work? Well, the product workflow is straightforward and follows a three-step process:
Step A: generative AI (Taylor)
You can upload a release and ask the AI to re-write it – or you can provide the AI with a few lines of descriptive text, and it will generate a press release for you.
In the case of the former, a rewrite just might give you a different look at your copy. It’s another iteration that takes just a few seconds to produce. You can also have it re-write it several times over if you wish.
This is an important point because you literally have the AI rewrite a press release a dozen times over in just a few minutes. That gives you options to choose or edit – and apply your own human judgment to the copy.
In the case of the latter, it will provide you with a draft release. Kwittken is quick to point out that the draft probably won’t be perfect, but it’s a head start that can save time. The better your inputs are into the AI – the better your outputs will be too.
The same process can be used to create blog posts, articles, social media updates, or whatever copy you might need. The software offers you a choice in tone too. For example, you can choose a persuasive tone if you wish – or a professional, luxury, friendly, bold, or adventurous tone.
Here’s a screenshot of what that looks like using a fictional press release from NASA.
In addition to writing your release, Taylor will also write social media updates based on the text and tone you’ve selected. Here again, you can choose to use these or edit them to your liking before they go out.
Step B: predictive AI creates a list of reporters that might be interested
In this stage, you pick the geography you want to reach and then the predictive AI analyzes the text of the final draft of your press release. Next, it compares your text to the articles reporters have written and published in the public domain.
In comparing the two, it compiles a list of reporters you can consider reaching out to with your announcement. It provides an estimate of how likely that reporter is to be interested in the topic on a scale of 0% to 100%.
Many PR software tools today prompt PR to select reporters based on broad generalizations of their coverage (beats) like “technology” and then blast out pitches and releases. This is a terrible way to do it and creates PR spam. By contrast, the better PR people will search for related news stories and reach out to reporters with relevant coverage.
In a sense, PRophet is mirroring the behavior of better PR pros. Instead of looking for reporters, in Kwittken’s words, the tool is looking at content and surfacing reporters based on their work.
In the screenshot nearby you can see the AI has an estimated 87% chance of interest from the first reporter on the list. Moving to the next column, you can see a space for feedback too. So if the AI suggests a reporter you know isn’t a fit, you can mark it as such and the AI will learn from you for the next time.
Step C: generative AI (Taylor) writes a personalized pitch
In the third and final stage, Taylor writes a personalized and pithy pitch about your press release for the reporters you selected one at a time. This time too, you can choose to accept the pitch, edit what Taylor has provided, or ask it to re-write the pitch again.
Kwittken was quick to point out this is a starting point – that PR pros can and should be engaged and supervising the AI suggestions. Take a close look at the screenshot nearby to see the personalized pitch.
It’s also important to note what PRophet is not doing as a software tool. The company is decidedly not creating lists of reporters and hosting them like most tools do. In place, they offer export tools for PR people to manage their own lists.
Second, the company is not a CRM and they are not tracking interactions. Kwittken was vocally suspicious of tools that do this – that these companies could be monitoring and data mining sensitive communication between PR people and reporters.
That’s not an issue I’ve thought of before, and I don’t think I agree with it. You shouldn’t be pitching reporters – or doing external comms – about stuff that’s a trade secret. Tracking conversations is important to keep the same team from following up with the same reporter and reviewing the historical interaction. However, it’s an interesting point that’s worthy of conversation.
3. PRophet pricing and customers
The company offers three pricing tiers:
- Free but is limited to one user and has strict caps on predictive searches for reporters and text generation;
- Pro which is $299 per month (~$3,500 annually) with seats for five and more generous caps on predictive searches for reporters and text generation;
- Enterprise with offers unlimited seats and usage which PRophet says can range from $20k-$100k annually.
The free version has a waitlist and will be available “starting on March 31, 2023” according to a press release. You can sign up for the waitlist and see the pricing details here.
Kwittken says PRophet has 500 users on the platform today. Those users are a mix of mid-sized agencies and large brands, including a global CPG brand, an account that was landed through relationships Stagwell has built.
4. Future plans for PRophet and AI in PR
Kwittken made it clear that PRophet does not intend to be all things to all people. In classic business software terms, PRophet is taking a “best of breed” approach.
The company is considering a partnership with a monitoring solution – though he declined to name the company at this time. He believes there may be a market need for tools that monitor the web for misinformation and disinformation.
He also has ideas around voice commands or perhaps the ability to run audio through PRophet’s AI. Any specific new features remain to be seen, although Kwittken mentioned their product roadmap as a long list of aspirations.
5. PR tech assessment
Anyone that knows me, knows I approached these briefings with a good deal of healthy skepticism. Having worked in B2B technology for so long has hardened me to the rainbows and unicorns vendors are prone to promise.
Even so, I walked away from this briefing genuinely impressed. The company seems to be very intentional about what it’s doing. It’s put thought into the workflow and matches how PR people work. It’s a very simple design and the demo itself didn’t take but 10 or 15 minutes, which demonstrates the streamlined functionality.
I welcome the fact startups like PRophet are challenging the incumbents. In my assessment, too many of them have become obsessed with dealmaking and M&A. They have been spending all their time integrating businesses – and not innovating. And then here’s this little startup that’s barely three years old that’s beating incumbents to market with AI.
* * *
AI in PR software is a bet
An assessment I read from a brilliant thinker in another knowledge industry, that isn’t PR, called AI a bet.
If you think AI is going to change the world for the better – that’s a bet. If you think AI is a disaster that will ruin jobs – that is a bet. Even “abstention is a bet.”
Whatever you might think about Kwittken, PRophet and AI in PR – he’s put his own money on the line – and is giving the PR community a risk-free way to give AI a try.
If you enjoyed this post, you might also like:
A Fairly Comprehensive List of PR and Comms Technology Vendors for 2023
Image credit: DALL-E “A robot dressed like a reporter typing on an old fashioned typewriter in the style of Van Gogh”
