27th Apr, 2020

Mindful Messaging: How Persado is Adapting its Language Architecture to a Post-COVID-19 World

Never before has digital content and brand messaging been under such an intense microscope – truly. 

I can’t think of another time when promotional emails and branded social media posts were the target of New York Times articles, viral tweets and widely shared memes with such frequency and an overall agreement that at a time like this – and at the risk of making a self-serving observation here (see our tagline) – words matter.

What’s different about this crisis

The world has seen its share of local crises but the current pandemic represents the first time, maybe ever in recent history, where the social context has been rewritten so rapidly and dramatically that it overshadows everything else and resets the priorities of our collective consciousness. And yet, life needs to somehow continue. Not just after the crisis but during the crisis. 

The stakes are high for all business leaders

What we’re experiencing is by all accounts and in many ways unprecedented. One of those ways, the one that we monitor more closely at Persado, is the way companies speak.

For business leaders responsible for communicating to anybody in this context – employees, suppliers and customers alike – whatever your brand says at this time is naturally being perceived in the context of what’s happening in the world. Even by not making a comment, you’re making a statement. How do you strike a balance between finding a way to continue operating and not appearing to be “off color,” at best or, to quote this Slate piece, “weird” and “random,” at worst? For Persado, an AI company where we “mine” hundreds of different messages on a daily basis for all sorts of channels and communication types, the first question that we needed to answer, for our customers and therefore for ourselves and how we operate, was how to strike that balance between empathy and performance in communications in a way that honors this larger social context. The larger challenge is how to address this “new way of speaking” that seems to be emerging because of the crisis – a necessarily more mindful, context-sensitive, tonally-adjusted approach to branded communications – from a linguistic perspective. 

A language architecture for the new normal

Our focus is on translating the “new normal” into language architecture (i.e. concept categories, tags and metatags) that will feed our machine learning and train our algorithms to generate relevant output “like a human would.” In other words, how do we make sense of the world in AI & machine learning terms because there’s no machine that has learned how to speak at a time when not even humans know what to say. 

How do we make sense of the world in AI & machine learning terms because there’s no machine that has learned how to speak at a time when not even humans know what to say. 

Here’s a summary of what we’re doing to help our clients imbue that sense of context and respect when talking to their customers, and how to balance sincere human empathy with data-backed performance. 


Beginning in February, Persado creative teams began a systematic audit of client messages that had been finalized but not yet broadcasted. We called it a “mindfulness scan” — a human, in-depth look at all the words and messages to reassess how they can come across in the new context.

One message our team flagged had been written to go live during March Madness, the annual National Collegiate Athletic Association (NCAA) basketball tournament. The ink had already dried on the sale tagline, THE MADNESS BEGINS NOW. But in the context of COVID-19, the playful ad with its shouty caps and enthusiastic tone no longer seemed fun and on-point. Instead, it seemed scary.

The ink had already dried on the sale tagline, THE MADNESS BEGINS NOW. But in the context of COVID-19, the playful ad with its shouty caps and enthusiastic tone no longer seemed fun and on-point. Instead, it seemed scary.

Today’s context can do that to language. Messages that include innocent play-on-words suddenly seem sinister.

Put in the context of AI, machine learning has never had to interpret the tone or meaning of words in these kinds of situations, so injecting human understanding and sensitivity becomes even more resonant right now.

Mindful messaging don’ts — and do’s

Through that audit, Persado identified patterns — both positive and negative — that we evolved into mindful messaging don’ts and do’s. We highlighted some of those techniques in an early webinar along with a set of guidelines and began implementing them immediately for all client experiments. At the same time, the Persado development team began creating a mindfulness filter, or ‘mindful narrative’ as our teams call it, for the AI algorithm to help automate the process.

What’s new: Persado is conducting mindful language scans as a standard practice for all language experiments.

What’s next: Persado’s data science team is developing a filter for the algorithm that can automate and flag these do’s and don’ts for content teams.


Emotional language is the #1 driver of performance when it comes to most communications — to put it simply, people respond best to messaging that they can connect with. Persado’s AI is trained to constantly experiment with 15 different emotions, score these emotions based on response data, and link those scores to contexts, communication & audience types and so on, in order to continuously learn and improve performance based on the language that works. But at a time like this, Excitement or Urgency, for example, take a whole new meaning. Without getting too technical, if the social context is one of the parameters AI takes into account under normal circumstances to generate relevant content — along with the offer type, the brand voice etc — right now the social context is the single parameter that matters. 

Offer expiration-focused Urgency language can sound too alarming or insensitive even if in theory it is appropriate for the offer. Overly positive language, even if super on-brand, can come off as tone-deaf and drive customers away. 

To tackle this new challenge in terms of our AI’s “logic,” when it comes to picking relevant words and phrases from our database, we had to switch from offer- and brand-first to social context-first. Our experiments are now designed to explore Trust-based emotions more heavily — an algorithmic way of focusing on the human priorities of safety, health, and closeness with people we love (i.e. the “warm” side of our wheel of emotions). 

To tackle this new challenge in terms of our AI’s “logic,” when it comes to picking relevant words and phrases from our database, we had to switch from offer- and brand-first to social context-first.

Still, because our goal is to help our customers move beyond the “we’re here for you” and the “we’re in this together” messages, our Content Intelligence team is at the same time researching ways of conveying different emotions that can actually be relevant now, analyzing the different shades of Safety, Gratitude & Intimacy that have become extra resonant, and defining the ways these emotions can get connected to each brand’s voice. 

Verizon’s “We’re here. And We’re Ready.campaign, launched in the very beginning of the crisis, is a great example of being in tune with the times. It’s a message that conveys safety and confidence through the idea of competence — a rarely-used concept. Verizon features real engineers explaining in conversational language how they plan and prepare all the time to ensure the systems run as they need to. 

What’s new: Persado is working with customers to generate content that evokes Trust-based emotions in new ways, and fine-tune it to better engage their audiences.   

What’s next: As Persado’s AI runs more experiments with a new focus, the algorithm gains more insight into the variety of emotions and narratives that connect with customers as our classification of emotional concepts expands to feed our language database with more appropriate content and to sensitize our machine learning. 


We all saw it — the switch from “A message from our CEO” messaging to a more “Staying positive at home” tone. As the situation evolved from February to March and into April, brands have been exploring different ways of staying relevant and resonating with the new realities of their customers. The new-normal has translated into a new global language, a new way of speaking — speaking to customers who, at best, are social-distancing, working from home, and wondering what’s next. Every research project at Persado starts with gathering data & analyzing it, to understand the context and to translate that understanding in AI terms. In this case our Innovation team has been working on capturing those new nuances and concepts that have emerged, and classifying them in a structured way that will enable our customers to learn what works — what strikes the right balance between being mindful and getting results.

This particular project, condensed from a normal research period of months to a couple of weeks (because nothing is normal anymore), led to a core classification of four categories and multiple subcategories of language which we call Mindful Narratives. It’s a new series of tags and metatags that got added to our language architecture, to help us generate content that connects the dots between what drives performance and what makes sense right now, to understand what the right shade of this “new normal speak” is for each audience and each brand.  

One new idea is the concept of sanctuary embraced by brands like IKEA with its website tagline: Home. The most important place in the world. Another new idea is solidarity, communicated through CBS’s “We’re all in this together” public service announcement.

While these concepts aren’t entirely unprecedented, brands haven’t used them at scale, so we don’t know how well they engage customers. Persado is now capturing these new concepts and others as they emerge, and classifying them in a way that allows the AI to recognize and categorize them in language experiments. With data, we’ll better understand how these new ideas perform in certain contexts and industries. We’ll also understand their relative power based on how sanctuary-based messages perform in relation to solidarity-based ones, for example.

These insights enable Persado to help brands drive performance in a way that balances the need to be sensitive and empathetic with the need to maintain business health.

What’s new: Persado is running experiments with customers using the new Mindful Narratives concepts and assessing how well they connect with customers and represent the brand 

What’s next: We are continuing to identify and experiment with new ideas as they emerge in order to adapt the way we and our customers talk. We see this as an ongoing evolution, as communities around the world move toward the next version of normal.

Crises lead to big changes, but the changes we’re all facing right now are not some one-time, before and after event, but better viewed as a process of adaptation and growth. The way brands speak is going to change multiple times as we move through this crisis, because they have to. We have to revisit everything right now.

 “Don’t let a serious crisis go to waste,” as they say.


Subscribe to our newsletter for the latest news & information.