PR firms start using AI for mundane tasks

For instance, Shift Communications has a marketing technology team of 10, including six data analysts. The team runs advanced analytics, machine learning, paid search, social media and more for clients.

Christopher Penn, vp of marketing technology for Shift Communications, said that natural language processing involves determining whether people perceive a brand positively or negatively, and identifying the most important terms and common themes that emerge from speeches, social posts and articles. Meanwhile, his team uses speech recognition tools to turn voice content into words, and performs predictive analytics to project media trends based on historical data.

But, before you get any ideas, robots aren’t yet coming to write press releases and call up reporters.

“The use of AI is not meant to replace PR professionals’ jobs — instead, it allows us to do our work at scale,” said Penn. “So much media is created every day, and we humans can’t read it all, but the machine can. For instance, I was analyzing content for a client’s trade show this morning, and there were around 2,000 pieces of content around the show. I couldn’t read every article on my own, but I ran natural language processing on those articles in 15 minutes.”

Penn explained that in addition to natural language processing, tools for speech recognition like IBM Watson Speech to Text and Google Cloud Speech API help his team convert speeches into text, and then his team can perform natural language processing on that text-based content.

“The best services cost pennies or less per recorded minute, and so much value is locked up in non-searchable audio,” said Penn. “Think of everything we listen to in marketing: Conference calls, speeches and presentations, et cetera, and how much of that knowledge is locked away from search. We use speech recognition to turn client calls into transcripts, speeches into blog posts and so much more.”

AI can also help PR professionals plan their media outreach calendar in advance. For instance, Penn’s team could analyze the past three years of a company blog’s traffic to predict what the site’s traffic would look like over the next year. Let’s say the brand’s blog is predicted to see traffic spikes next May and next November, and its readers like the subject “precision marketing.” Penn’s team would relay this insight to the PR team at Shift Communications, which would then look for speaking engagements and contributed article opportunities around precision marketing for the client in May and November of next year, according to Penn.

Elsewhere, Crenshaw Communications uses AI to track where a client gets mentioned on social media and monitor how people perceive the brand, according to Chris Harihar, director of the agency. “You increasingly need AI for both the tracking and analysis pieces, especially for larger clients with high coverage volume,” said Harihar. “Sentiment analysis is the holy grail, but it is still in progress in terms of accuracy — at least from the tools that I’ve seen.”

One area where PR professionals haven’t applied AI, though, is writing news releases. Penn thinks that while AI can easily produce data-specific and highly formulaic financial news releases, it will take at least five years for the technology to write general news releases that require personalization.

Harihar agreed that news releases involve too many customized elements — including overall company messaging and point-of-contact writing preferences — for any software to manage.

“If an agency is claiming to have created some sort of a press release stack that automates release writing, I would call it bullshit,” he said.