We Wrote This … Or Did We?
An Examination of AI and External Communications
Take a deeper dive into AI and the Internet of Behaviors at the virtual IABC World Conference 2021. In "Leading With Values — The Internet of Behaviors (IoB) and AI," trend report authors Joanne Henry and Donna Itzoe will examine scenarios in marketing, internal communications and PR to understand how ethics and opportunity need to be bedfellows. Don't miss this interactive workshop, taking place on 28 June at 4:40–5:10 p.m. PDT (12:40–1:10 a.m. BST, 29 June; 9:40–10:10 a.m. AEST, 29 June).
An Examination of AI and External Communications
Earlier this year, the IABC Trends Watch Task Force explored in a trend report how communicators can benefit from using artificial intelligence (AI) and the Internet of Behaviors (IoB). As we further discussed the risks, rewards and ways to use AI in communications, we realized a deeper dive was needed. There are many risks, particularly with respect to a company’s transparency and honesty, but many may not be as familiar with the rewards and ways to use AI to our advantage. We want to cover those here.
McKinsey predicts AI and its technologies are not going anywhere. “We estimate that the AI techniques we cite in this briefing together have the potential to create between $3.5 trillion and $5.8 trillion in value annually across nine business functions in 19 industries. This constitutes about 40% of the overall $9.5 trillion to $15.4 trillion annual impact that could potentially be enabled by all analytical techniques,” the organization said in a 2018 report.
If McKinsey and others are correct about the prevalence of AI, then communicators as leaders must be comfortable with these conversations and find ways to use the technology.
External Communications, Marketing and AI
For external communicators and marketers, AI means we can now get insights from massive sets of data in seconds. This requires working alongside data analysts and technology partners to embrace and use AI-powered technology.
As communicators, we need to be mindful of how messaging consumption is changing and how that affects our role and our capabilities. Net point: the toolbox of communication channels will need to change, which is not new for us.
We don’t have to look very far to understand how AI data analytics has helped communicators succeed in extremely challenging circumstances. Data has been one of the most important allies we’ve had during the pandemic. For instance, health systems learned where cases of COVID-19 were popping up and started communicating with local health officials to share news and education. Armed with data instead of misinformation, we could communicate safety precautions in our own companies, notify procurement teams to purchase protective personal equipment (PPE) and more.
According to McKinsey, “AI’s application to real world business problems extends across nearly every sector of the economy, but the biggest impact could arise in two particular business functions.” They estimate these areas to be marketing and sales (USD 2.6 trillion) and supply-chain management and manufacturing (USD 2.0 trillion).
Most notably in marketing, AI can help with pricing and promotion, customer service and acquisition, and lead generation. External communicators and marketing teams need to pay close attention to this trend to guide their organizations appropriately.
As consumers, the data being shared about our online sales and searches can be frustrating. The new light fixture you just viewed online pops up on every site you visit for the next several days. While this alone isn’t new, the algorithmic sophistication has leapfrogged from the early days of online marketing. Besides the light fixture, you might get other offerings for household goods; a refinance offer from a partner in mortgage lending, a credit card offer, or — if your light fixture was deemed too high for your ZIP code’s average income — credit counseling!
From a marketer’s perspective, the wealth of integrated data does paint an instant picture of a prospective customer, their price sensitivity, tastes, combined interests, what images make them happy, which channel they’re most likely to click-through to on an image for your products and where they’re most likely to respond to an image. We have the chance to communicate more personally and, thus, more effectively.
This is where our expertise helps. We know that meeting people where they are and talking about what they’re interested in is most effective. With AI, we can use complex data sets to prove our proposed communication tactics have validity. Post-launch, we can use the data to re-affirm those postulates.
The Need for a Human Pilot
In 2019, Forbes took on the topic of AI and the marketing function: “Empathy, compassion and storytelling are all attributes that machines can't emulate, at least not yet. At the end of the day, AI is not bound by human limitations.”
Gartner predicted that by 2018, 20% of all business content would be authored by machines. Before we let AI do all the work, we need to remember that our experiences and communications expertise are critical to driving its use.
For the marketing professional, there are clearly opportunities. How we use them is where we find our balance.
We can also heed the advice of journalist Christine Crandell: “Augmenting marketers’ skills, akin to being their wingman, AI will have more impact on the discipline than any martech product to date.” She adds, “Marketing will need to prepare for AI and carry forward the lessons learned from predictive analytics and today’s patchwork of martech solutions — automation can only offer recommendations that are as good as the underlying data sources. In the end, the human is still the pilot.”
The Use of AI in Public Relations
AI and the Internet of Behaviors (IoB) are PR gold if we’re ready to find it, source it responsibly and create ethical best practices to capture its value. As in the early days of the Gold Rush in the American West, you don’t always need a lot of money to try your hand at using some of these tools. Open source and free resources are available and growing in use. Sophisticated mining, though, will take more investment.
While AI tools have been in use by the private sector and military since the 1950s, the power of AI is now more widely available to the masses, including PR professionals. Every time you enter a Google search, you’re using this free tool that does real-time monitoring of thousands of sites, an impossible task for any individual. A number of programs and apps will analyze text for sentiment analysis. A recent development is the use of audio analysis to monitor a customer’s tone of voice in a recorded customer service interaction. A headline from Tom Simonite in Wired cuts to the heart of it: “This Call may be monitored for Tone and Emotion.” One way PR professionals might use this tool is to analyze how empathetic their spokesperson is perceived — and whether that perception is reality.
Beyond assessment and analysis, AI combined with IoB can bridge the gap between communications and actual changes in behavior. In other words, we can verify an impact and know exactly what worked. If communications is both an art and a science, this technology allows the science to be much more meaningful and important. Whether you’re taking baby steps with AI, embracing it fully or waiting cautiously on the sidelines, it will be difficult to avoid the impact that it will have on the communications profession in the decade ahead. Imagine knowing within hours whether the campaign you ran has changed the behavior of your audience — it’s much different than polling them to see what they say they will do.
Analyzing Journalists’ Behaviors – at Scale
Knowing your audience, including stakeholders and journalists, is a core element of public and media relations. In an article from The Drum, author Kenneth Hein explains the potential of a new PR platform, PRophnet, to predict journalist’s future behavior. This tool samples more than 2 million articles from 12,000 plus reporters to yield a probability score for a specific reporter acting on a particular pitch. Imagine this: the next time the CEO insists you pitch a story you think has no merit, just run it through the algorithm.
Cision, Pitch Intelligence and MuckRack are other tools for this function. Following through on this process to demonstrate real PR outcomes and ROI, a next step would be to pitch specific stories to the reporters most likely to cover them and compare results to the data from pitches in the past without the use of automated tools.
The ideal scenario: free up staff to do what AI cannot — build personal relationships and scale their reach, without resorting to mass emails that include sending pitches to media channels and reporters with zero interest in them.
According to Hein, PRophet has software in the works to pitch podcasters, as well as the capability to test images. Other platforms already include image scanning. One question is how well the software will pick up potential threats, such as using an image that is culturally sensitive in one part of the country or among a specific audience group. Impact is one thing — sentiment is certainly another.
Other questions to ask potential PR technology providers include how well their database matches your own geographic, cultural or market niche. If your audience is primarily in Southeast Asia and the database of articles and journalists is weaker there, you won’t get the results you need.
Sentiment Analysis for Brand Marketing and Reputation
Natural language tools for sentiment analysis run from simple, built-in tools such as LinkedIn’s messaging tools, which scan how trustworthy messages seem, to platforms offering tools that gather sentiment changes on a granular level, pinpointing exactly when, where and who began a shift in sentiment.
For global communications, automated text translation is essential to productivity in a world that changes at gig speed. A variety of tools are springing up to provide sentiment analysis in a number of languages and countries, too. For example, Awario is a social media monitoring tool with a dashboard that categorizes data by country and importance. Talkwalker offers sentiment analysis across 187 languages to track stories in real time, including pulling data from television and radio.
Other examples of brand and reputation AI-based tools are Brandwatch, MonkeyLearn and the IBM Watson Cloud services that include the Watson Tone Analyzer. The Watson Tone Analyzer will examine tweets, reviews and customer conversations so chatbots can respond to a user’s tone.
We’ve seen the growing success of native advertising; now imagine how automated, tone-sensitive responses on all of your important social media sites can occur in real time, at scale. While all of these tools are designed exclusively for PR professionals, it’s not a leap to see how these advanced capabilities can be integrated into other apps or customized for PR use.
Read about more AI platform solutions for PR use via PR Daily here.
Crisis Communications Prevention and Recovery
AI sentiment analysis, and social media and conversation tracking can also be used to detect a crisis before it happens. This is what government agencies have been using all along, with increasing sophistication. However, there has been backlash, including profiling and protests over privacy and individual rights. Still, it’s difficult to deny the value to prevention if you have the capability to track sentiment suddenly moving against a brand, company or key individual at an organization. It could also be critical to stop misinformation quickly and identify the original source.
Crisis guru Professor W. Timothy Coombs and his co-author, Elina R. Tachkova, presented on the lack of research on the most effective responses for a preventable crisis at the International Public Relations Research Conference in 2019. They argued that a “preventable or human-led” crisis, requires a very different response than other types of critical events. While the authors noted there is plenty of research on how to respond to natural disasters and accidents, not much is available about preventable crises.
The good news: the gap is starting to close. Following Paine Publishing’s 2020 Summit on Reputation Management and Measurement, one of its sponsors, Fullintel, rose to the challenge. Katie Paine recently covered the work of Fullintel and Coombs in an article online here.
Working with Coombs, Fullintel gathered thousands of articles of traditional media coverage on three recent human-error crises: the Boeing 737-Max, the demise of WeWork, and an accidental toddler death. The research made use of different types of crisis response from Professor Coombs’ “Situational Crisis Communications Theory” (SCCT). Fullintel taught its machine learning technology how to identify a crisis, how to classify the crisis and how to identify different responses.
The results, Paine says, were surprising in just how close the machine-recommended responses were to the human responses. In tracking results by how quickly media coverage turned to neutral after a crisis (a sound measure of success), the AI results bore out what our instincts tell us — that the least effective response strategies for this type of crisis are denial, stonewalling, scapegoating and discrediting information sources.
Does this mean a reduced role for crisis expertise? Paine says no. But how much more successful will such experts be, armed with data on which strategies have worked and which have failed in human-error crises over a large span of time, and including hundreds of data points?
In crisis PR, when you have a (hot) seat at the table, always bring data. This doesn’t mean throwing out everything you know about empathy, trust and acting with transparency. It does mean that the AI tools to persuade people with facts are there, and they’re getting better every day.
Authenticity and Privacy
Authenticity and transparency are essential to building trust. Where do AI and behavior tracking fit into this environment at a time when trust is tenuous in so many situations?
Disclosure is usually the best policy when it comes to monitoring or using tools to post automated messages that sound like a real person. That policy isn’t always followed, although the proliferation and acceptance of chatbots may change that. Beyond the moral issue around disclosure, companies risk reputation and credibility if their AI responses to queries smacks of corporate speak or don’t match the actions of the organization.
Responsible communications departments and their organizations will establish their own thoughtful policies for how automated AI tools can be used in the absence of clear guidelines. For example, any information gathered should be disclosed in a specific way and at a certain point in the customer engagement.
Best practices, if not legal requirements, require the use of personal user data by permission. (A narrow example: “I accept all cookies.”) Regulations are adapting to ensure consumers know when and how their data is used.
In the absence of consistent law, we need to police ourselves and, perhaps, our industry. Few of us working in communications during 2018 can forget the image of Facebook CEO Mark Zuckerberg testifying on the social media giant’s use of data mining, its disclosure practices and the impact on consumer privacy. Today, the tools are there to track much more than our online behavior; we can monitor almost any act, and in real-time. As communicators, we need to weigh in on matters such as disclosure, informed consent, when we communicate that we’re watching/listening and more. We need to apply our ethical judgment and provide a perspective of how our company’s use of perspective will be perceived once it’s in the rear-view mirror.
For additional information and details, read “New Ethics Guide for Artificial Intelligence in PR” from the Institute for Public Relations.
The Risk of Perpetuating Exclusion Practices
One significant criticism of data mining to influence future policies is that AI tools can be used to perpetuate the status quo. Since many people of color, women and other marginalized groups are underrepresented in mainstream channels, scanning sites for “qualities of top surgeons” or “colleges attended by the most awarded architects” may produce results that encourage narrow selection processes and reduced opportunities.
The impact of AI tools on equity and inclusion efforts could be a full research paper on its own. The United Nations Educational Scientific and Cultural Organization (UNESCO) recently published a report that tracked specific uses of AI that posed an ethical dilemma. As a result, UNESCO is working on a legal, global document on the ethical implications of AI. Until then, organizations may want to consider expanding their ethical best practices to address both current and emerging technologies.
Conclusion
There is so much to consider for communications professionals when it comes to AI. For those hesitant or thinking you don’t have a voice in this topic, we can no longer say we don’t have a seat at the table. We are firmly planted there.
Last year in particular, our clients and companies looked to us, asking what to say and, in many cases, what do we do. Indeed, it’s an understatement to call 2020 a tumultuous year, with a worldwide pandemic, cultural clashes and protests, and political unrest that permeated the world. Many organizations looked to those in the communications industry to ask for help because we are purveyors of authenticity and transparency.
We must focus on how AI factors into our world and why we should care. We must use the positions in our organizations to take leadership in using AI. Unlike other technological changes and new products, AI may not be as obvious as embracing email.
This technology is embedded in our lives. We may not be its owners, but we should forge our way into conversations around it — the authenticity of our organizations depends on it. There are ethical boundaries to consider, true, and we must embrace those front and center. But let’s also look at the trends that help us be even more effective communicators, help us reach populations we haven’t been able to and give us data to understand what we may not know to accomplish all of this.