There are countless buzzwords that make their way into the lexicon of an industry. Yet few have permeated as many as “artificial intelligence,” or just “AI.” And there are at least as many definitions of AI as there are applications.
So what is AI as it applies to communications professionals?
In broad terms, AI means machines performing tasks, oftentimes more quickly than a human could. The tasks range widely.
“From language translation, image identification and complex computational tasks to classification and automation, artificial intelligence offers tremendous promise to augment human capabilities and realize the full potential of data,” says Chad Latz, chief innovation officer at BCW.
Mark Stouse, chairman and CEO of Proof Analytics, breaks AI into three definitions: assisted intelligence, augmented intelligence and autonomous intelligence. While the AI most commonly referred to falls into the third category, Stouse argues that the other two are much more commonly used in business.
“Assisted intelligence enables communications professionals to perform many duties better and faster. It enables you to scale activities that are already being done,” he explains. “Augmented intelligence delivers insights, capabilities and abilities that did not previously exist or were not normally accessible, thus transforming the efficacy and efficiency of an organization and its operations.”
The former are those that typically deal with data collection and measurement, while the latter have automated analysis capabilities, enabling humans to make better informed, quicker decisions and measure the results of their efforts.
John Gillooly, SVP of data and analytics at Hill+Knowlton Strategies, uses the terms “segmentation” and “summarization” to describe his definitions of AI in comms.
“Segmentation is all about more robust or more thorough segmentation of data types, i.e.: coverage, themes of conversation and audiences,” he says. “Summarization means machines summarizing a body of text in a manner similar to a human.”
Another, newer application of AI in comms is message and content testing and synthetic creation of content. Compared to analytics, this use is still very much in its formative stages, notes Brian Buchwald, head of global intelligence at Weber Shandwick.
Scale and speed
Experts also point to two characteristics expected of AI: scale and speed. AI, at its core, can process, categorize, sift through and analyze massive amounts of information and in some cases even predict events or crises at a speed that’s not possible for a human being.
Yet while we are likely to think of AI as futuristic, it is increasingly a part of our everyday lives, note experts.
“Most people rarely think about the simple algorithms: the AI systems that run our lives and cities,” says David Sanchez, director and machine learning and AI strategist at APCO Worldwide. This includes everything from traffic lights to voice-enabled devices.
Yannis Kotziagkiaouridis, global analytics officer at Edelman, notes that these AI tools are “so integrated that we don’t really make the connection.”
Chris Bingham, chief technology officer at Brandwatch agrees. “AI has a way of making itself invisible once it becomes familiar,” he says. “Simply asking your phone for directions is an immense technological accomplishment that would have been science fiction not long ago, but it’s quickly become so commonplace that it no longer seems remarkable.”
AI. What is it good for?
There are at least as many ways to use AI in communications as there are ways to define it. At its core, it is a “capability extender” for PR, says Paul Quigley, cofounder and CEO of NewsWhip.
“It’s going to steadily improve and change the technologies that PR people already use, like media monitoring, social listening and media databases, making those technologies smarter, more personalized and more impactful,” he says.
Kotziagkiaouridis sums it up as follows: “AI can help us understand what people say, who they are and what is the next thing we need to tell them to have them take a specific action.”
More concretely, this means a number of things for PR pros. At Edelman, Kotziagkiaouridis notes that some of the more interesting applications of AI include using facial recognition tools to understand how someone receives a particular message, tracing how a mouse or finger navigates a page to understand biases or distinguish between stated and real preferences.
Perhaps most importantly, it could also mean enabling PR pros to start with granular data, creating a more bottom-up approach. AI helps identify an audience, specifically the individuals a campaign hopes to target.
“Then you apply algorithmic intelligence to figure out where those people consume media and which publications match that,” Kotziagkiaouridis says. Rather than choosing which publications or reporters to speak to based on national profile or subject area, AI can help a PR pro tell their client’s story in the place that the intended audience is most likely to see it.
“Decision making no longer has to start at 50,000 feet,” he explains.
At Weber, staffers use natural language processing and image recognition algorithms to quantify qualitative information, thereby increasing understanding of their stakeholder audiences.
“Our most typical models are proprietary sentiment or emotion models to comprehend how audiences feel and ‘action intent’ models that tell us what is prompting a candy purchase or a drug prescription or a travel booking,” Buchwald says.
In practice, Weber has used a host of AI tools for work with a pharmaceutical client.
“Through machine learning models generated from social media, forums and earned media, we were able to build discrete understandings of what drives favorable emotion for healthcare professionals, patients and caregivers,” Buchwald explains. “Simultaneously, we generated insights into why each stakeholder would take action. We then were able to synthesize these different elements into one cohesive messaging strategy for each of the stakeholders.”
Sanchez contends that clients still aren’t interested in the “really pioneering stuff like generative AI or advanced voice enable solutions.” Instead APCO applies AI to things like “performing large scale analysis of survey data, sensing if a news story is going viral 30 seconds after it is posted, automated audio description, helping clients build a reputation monitoring system and helping clients train chatbots to respond to COVID-19 customer service demands,” he says.
BCW applies AI capabilities to three overarching areas: AI for enhanced data analysis and insights; AI and data to create consumer experiences and drive engagement; and AI and content. There are a number of solutions that fit under each of these categories. Image recognition and data-driven approaches to influencer marketing fall under the first umbrella. Chatbots and virtual assistants and AI-powered reply-based social engagement are solutions that fit under the second, while predictive analytics and mass personalization are part of the third.
“The truth is machine learning and AI services power a lot of the systems, services and platforms that we use as digital, PR and creative professionals today,” Latz says.