Dan Simon
Dan Simon

News-writing software that can report current events in real time, with little or no human intervention, has obvious appeal to media companies and newsrooms, and it’s time for the public relations field to take seriously the eventual impact artificial intelligence systems will have on the practice of media relations.

This effect is most obvious in financial reporting, and as the founders of a fast-growing financial communications agency, my co-founders and I’ve studied these developments with interest ever since we opened Vested in 2015.

This situation is far more nuanced than worrying that the robots are coming for all of the PR jobs, or that artificial intelligence applications will eventually replace the need that clients have for consultants. These dire predictions are not grounded in reality.

O'Dwyer's Aug. '18 Financial PR/IR & Professional Svcs. PR MagazineThis article is featured in O'Dwyer's Aug. '18 Financial PR/IR & Professional Services PR Magazine

For practitioners, the real threat is failing to recognize these changes and evolve.

First, it’s important to establish what exactly is happening. In 2013, the Washington Post began experimenting with an artificial intelligence system that could generate explanatory articles around structured data. Up until that point, there were a handful of platforms that could generate bare-bones articles on one-dimensional, newsworthy matters that revolved around data, say, the box score of a baseball game, or whether a company hit its earnings target. The Post, which at the time had just been purchased by Amazon Founder Jeff Bezos, wanted to test the boundaries of this kind of automated reporting.

Instead of investing in deeply reported, long-lead journalism, which is expensive, Bezos saw an opening to grow the Post’s audience by using bots to quickly produce large quantities of stories that appealed to smaller groups or regions. Heliograf, the Post’s AI technology for reporting, debuted in 2016 by spitting out around 300 short stories and briefs on the Rio Olympics as a trial run of sorts.

Later that year, its developers found out how to apply a stronger editorial voice to other data-heavy news events, such as the 2016 elections and regional sporting events. That’s when things got interesting.

The 500 articles that Heliograf wrote about the 2016 elections produced 500,000 clicks, according to Digiday. “Not a ton in the scheme of things, but most of these were stories the Post wasn’t going to dedicate staff to anyway,” the publication wrote. “For the 2012 election, for example, the Post did just 15 percent of [the clicks] it generated in 2016.”

Across the media industry, the consensus opinion is that Heliograf was and is a resounding success and that using AI in this way helps newsrooms dedicate more time and energy to the kind of gumshoe, long-lead reporting that society depends on but which cannot be fully outsourced to machines.

This is good for society and good for the business models of media companies. AI-powered journalism is here to stay.

Now, in 2018, using bots to power financial reporting is no longer the exception; it’s the rule. There’s some element of automation and machine assistance in at least a third of the thousands of stories and headlines that Bloomberg News publishes every week, according to the company. Reuters recently made a big push into AI-powered journalism with Lynx Insight, a tool that can analyze large sets of data to suggest stories and write key parts of them for the newswire’s journalists.

Moving forward, “[n]ews is likely to get shorter, quicker, and more graphical,” wrote Bloomberg News Editor-in-Chief John Micklethwait.

This is what should strike fear in the hearts and minds of public relations firms that sell media relations as a core service. Journalists are simply more likely to value objective facts that AI-powered platforms can flag over subjective claims from publicists who spend most of their time trying to wedge clients into the news. Lynx also points Reuters’ journalists to the sources who were quoted in previous stories that relate to stories it suggests, further replacing the conduct that many media relations practitioners claim make them indispensable.

Professionals who practice in the financial services vertical are especially vulnerable to being made redundant by this kind of automation because finance is a sector that features a great deal of structured data. Markets gyrate, stock prices fluctuate and earnings come out during the same few weeks every quarter. These are all newsworthy events based on hard numbers that bots can make sense of in a few milliseconds.

Structured data is also disrupting the practice of media relations in other industry verticals. Sports features a similar statistical data structure as does finance, and the appetite for real-time reporting of sporting events is similar to the demand for real-time reporting of earnings announcements and other planned financial events. Politics has a similar dynamic around elections, although the process of structuring the data is different.

So, since financial stories are increasingly being reported by machines, and since financial journalists are turning away from media relations practitioners in favor of bots, what roles will continue to exist for traditional public relations service providers and in-house teams?

There isn’t a single answer, but rather, several possibilities.

The most obvious is that raw media relations will lessen in importance, and being integrated will become an imperative. Firms are already subordinating media relations for a different reason — the declining volume of reporters to pitch and the massive increase in public relations practitioners has created incredible competition for journalists’ attention — and the rise of AI-driven reporting will accelerate this.

Adjacent to this is the mandate for modern practitioners to use software applications to increase their own efficiency, or that of their firms or teams. Firms that use tiered billing rates, where administration and client management is billed at a lower rate when compared to strategic work, have already signaled an appetite to rely less on downmarket activity. The next step is automating it away.

The bottom line is that earned media will remain important, but as a smaller piece of a sound communications strategy, not the whole of it.

Another possibility is for PR teams to adopt the same data-first approach that newsrooms are. PR has never been truly about data, because firms are service providers, not technology developers. But PR has historically approached media relations with a reporter’s mindset, and for the practitioners who aim to speak the same language as journalists do, using modern technology in a similar way should help.

The biggest payoff for firms and practitioners, however, would come from the public relations function overseeing enterprise-wide data analytics, owning and structuring relevant data before it hits the bots that newsrooms use, or structuring private data in the same manner that public data is structured, and selectively pushing it out to journalists in a strategic way.

Much of AI-powered journalism relies on publicly available data; in the finance world, it’s generally data that’s disclosed because of a regulatory obligation, or market data that’s accessible to everyone. But even moderately large companies have a great deal of non-public data. Bots are helping journalists by probing public data sets for newsworthy relationships, themes and patterns; those same bots can do the same thing with private and proprietary sets of data.

This would empower the media relations practitioners of yesterday to act as the warden of arrangements that give newsrooms access to a company’s in-house data, ensuring it’s used in accordance with the company’s wishes and in ways that do not violate regulatory or legal obligations. If practitioners can get media stakeholders to rely on their data, they’ll have leverage over the process of reporting that PR has never truly had.

Information brokerage would become the prominent form of media relations under this model.

In other words, the future of media relations will begin to see public relations firms and teams extracting intellectual property from their clients and helping them build a repertoire of newsworthy material, rather than just pushing out what is handed to them.

All of this can either be an opportunity or a threat, but one thing is certain: It’s as exciting a time as ever to be a public relations practitioner.


Dan Simon is co-founder and CEO of Vested, an integrated communications firm focused on the financial sector.