Using AI To Tell Better Stories

I have recently begun to work with a small, interesting company, called Significance Systems, who spent eight years developing a platform that uses Machine Learning and Computational Linguistics to help marketers become more effective at storytelling. What they do is really groundbreaking I think so I want to share some thoughts.

AI is becoming more prevalent in marketing, but it seems to be used more for increasing operational efficiency rather than marketing effectiveness. But I would propose that one of Marketing’s biggest challenges to being more effective and relevant is an empathy gap with their audience. Better storytelling could be the solution.

Storytelling is as old as mankind. It is the substance and matter of our culture and our search for meaning.

Today, on the internet, there exists a sea of stories that are interconnected by common topics, ideas, issues etc. These individual stories can coalesce into larger narratives that become a “cultural truth”, which help us form an understanding of our world - the past, the present and the future. It becomes a lens that we use to evaluate and make decisions.

Issues or topics like climate change, the European Union, trust in politicians, trust in brands, are all examples where narratives are shaping people's beliefs. Understanding how well these topics, and narratives around them, engage people, what is driving the narrative and what emotions the narratives elicit, will help governments, policy makers, cultural leaders, financial leaders, business and brand owners make better and more effective decisions.

For marketers, being able to analyze narratives that are important to a brand and breakdown who and what defines it, will help close the empathy gap to be more authentic and congruent with the audience.

So how can we analyze the internet to understand these narratives to support decisions?

Machine Learning and Computational Linguistics can be used to analyze online content on any particular subject in order to understand the depth and longevity of audience engagement in that subject. Then, by comparing it to many other subjects or topics, it is possible to categorize them to understand their significance.

Rather than looking at proxies for engagement such as Facebook Likes and Shares, which results in “cat videos”, AI can look at the substantive nature of engagement.

For example, does the narrative attract people’s cognitive focus enough that they are able to shape it, give it context and define what it means for them such that they could act on it in some way?

Just as we talk about turning data into information, then knowledge, content, and narratives work the same way. 

Using AI techniques to analyze content associated with a topic, we can see to what extent a narrative has evolved into something more complex and advanced. A piece of content is just like an individual piece of information. But as more content is created, knowledge can be formed and then wisdom – the narrative has more meaning for people. Eventually, this narrative becomes “vision” – a way to navigate the future. This evolution takes “cognitive focus” – in other words, substantive engagement. 

So AI can be used to help us to measure the substantive engagement in a meaningful way that up to now has not been possible. But there is more to engagement than how well evolved a narrative has become.

Another property of engagement is whether it is restricted to an individual network of like-minded people or does it manage to cross “tribes” of people and become worth paying attention to?

Has it broken through?

So there is an opportunity to classify narratives in terms of its ability to substantively engage people. We can then also identify what content is driving the narrative around a subject or topic, which channels the narrative is being discussed and the emotional affect of that narrative.

Most narratives do not engage people. This is typical of marketing campaigns. Some achieve no more than passing buzz. They may seem like successful campaigns on the surface, but they have not really modified people's thoughts or ideas. What matters more are narratives that have long lasting, deep engagement – but the key is to appeal to more than a narrow tribe of people.

For Brands, this kind of analysis can be used to get an orientation on the brand perception and equity within a market and its ability to disrupt the market in the future.

If a brand has deep, long-term engagement, then its audience has done more than just consume content about that brand, they have critically engaged with it and formed a deeper wisdom about the brand and it’s impact on the future. The brand has real meaning and value to people – whatever that may consist of. The more diverse the audience, the more power the narrative has to disrupt.

So this is an example of how AI is currently changing marketing for the better and to mitigate many of the risks inherent in the practice

• It can help a brand understand how deeply attached, bound or engaged their audience is with the brand and what drives that attachment
• It can help a brand build emotional empathy and congruence with an audience
• It can help a brand understand how well it is associated with an issue or activity that they want to get involved in e.g. if a brand wants to be involved with social issues, or whether they are planning a sponsorship deal
• It can help start-ups understand their audience more deeply, the significance of problems they want to solve and what is more important to them
• It can help media companies more effectively plan, improve and market their programming through insights into their audiences appetites and emotional reactions
• It can help news companies understand what resonates best with their audience
• It can help evaluate the real importance and impact of marketing activities - is it really engaging and changing people's brand perceptions
• It can help identify the most significant content that is shaping a narrative which can then inform the brand's content strategy so as to avoid poor generic brand content

Thinking about the recent Pepsi ad with Kendall Jenner that had to be withdrawn because of the backlash it received, whilst it’s impossible to go back in time, I think this use of AI could have saved them from making the errors they made. To begin with, they could have understood the narrative of the protest movement they were trying to insinuate themselves into. They may have seen that Pepsi had no association with the topic, and hence they needed to do more authentic work to build that association. They may have noticed the type and strength of emotion attached to the movement and which, hopefully, would have dissuaded them from getting involved, or at least been more attuned and congruent with it. And finally, they may have found someone more credible and authentic to the movement than Kendall Jenner.

Please do share your thoughts or questions. If you would like to know more about Significant Systems, please do get in contact. You can reach me at [email protected].

If you liked this post, perhaps you'd like to follow me on Twitter: https://twitter.com/metric_dos

Or find me on Google+ https://plus.google.com/u/0/+MeyrickDSouza

I hope you enjoyed this post, thanks for reading! That's all for now!