We recently released volume 1 of Adventures in AI, the world’s first AI comic book. It was a smashing success, and many AI Leaders asked us how and why we created it.
First, let me dispel a common misconception: the book was successful not because it’s a comic but because it tells an engaging and shareable story that helps AI Leaders grasp the value of AI. They get to the point without wading through unnecessary confusion to explore this abstract, complex technology. Comics can be a good medium for telling stories because they combine images with characters and dialog. But other forms of storytelling can work just as well.
A compelling story of AI can help others understand your vision and become evangelists for it. In this edition of Feedforward, I’ll explain why stories are a critical part of your AI strategy and share some case studies.
Before diving into the value of storytelling, let’s talk about the bigger picture. The real challenge of becoming an AI-driven company isn’t creating some fancy new model—it’s creating the organizational drive to adopt this technology. It’s changing habits and rethinking business processes. The world’s most amazing machine learning model won’t create a cent of value if business processes don’t change to take advantage of it.
Unfortunately, most people are intimidated by AI. They don’t understand it. Although you and I can read about an innovation in natural-language processing (NLP) and instantly imagine its transformative potential, surprisingly few people can make this leap on their own.
Change is hard. Big change is really hard. And becoming an AI-driven company is one of the biggest changes imaginable.
Let’s take the relatively straightforward example from volume 1 of Adventures in AI. The data scientist builds a cutting-edge NLP model to automatically read and classify documents for the legal department. Her working prototype gets good results.
Is the project done? Of course not!
The model won’t do anything until it’s deployed and used by the legal department. This is no small task. The department has to spend time and resources to give feedback about the prototype. Lawyers need training to use the model, and they must change work habits. Because this investment can make the department temporarily less efficient, leaders must be fully committed to making the project successful.
Other departments must be equally committed. IT needs to deploy and support the infrastructure. Finance needs to create sustainable budgets for the program. HR must recruit technical specialists and program managers. And on it goes.
So how can you get your organization’s buy-in for your AI vision? How can you motivate people to invest in it and change their habits?
In The Believing Brain: From Ghosts and Gods to Politics and Conspiracies--How We Construct Beliefs and Reinforce Them as Truths, Michael Shermer explains the science behind beliefs and decisions. We assume that we generate beliefs based on first gathering evidence and then drawing conclusions. But science suggests the opposite. Emotions drive our beliefs, and then we look for facts that support our preconceptions.
The implications of this research are clear: all decisions are emotional.
Like it or not, the key to getting organization buy-in for your AI strategy is generating emotional support.
Why are stories so effective? Once again, science has the answer. Stories cause our brains to release oxytocin, a chemical that affects attitudes, beliefs, and behaviors.
Paul J. Zak, founding director of the Center for Neuroeconomics Studies, has conducted numerous experiments on the effects of narratives on human behavior. It turns out that stories get our attention, provoke cooperation, and generate feelings of goodwill for the narrator.
Couple your AI strategy with a story, and you can literally change minds through brain chemistry.
Hopefully you now understand why your job is so challenging. Creating an AI strategy and getting your CEO to bless and fund it are critical first steps. They are also the easiest steps.
The bigger challenge is galvanizing the organization to sustain a multi-year transformational effort. Unfortunately white papers, plans, and presentations just won’t do it. Our most successful clients develop stories to supplement their AI strategies and continuously tell them to their constituents.
Explore some examples in the next section.
AI may be the “new electricity,” according to Andrew Ng, but decades ago, General Electric used stories to evangelize the “old” electricity.
In the 1950s, GE’s market for electrical appliances was limited by the number of homes that had electrical service. The company recognized that rural communities were more likely to invest in electrical infrastructure if residents understood the value of appliances like washing machines. So GE created a comic series called Adventures in Electricity to educate the public and generate enthusiasm for investing in electricity.
We love Ng’s electricity metaphor because it succinctly frames how AI will eventually be integrated into everything we do. Package the metaphor with some stories and you’ve got a winning combination for spreading your AI vision.
We are at the beginning of an NLP revolution that will generate trillions of dollars in value in the coming decades. Most of Prolego’s clients embrace this future. They have tasked us with building applications to perform tedious tasks that no longer require human labor, such as reading and sorting documents.
But our clients are in the minority. Most companies still see NLP as an incremental technology rather than a transformative one. Russ and I have had dozens of conversations about NLP with stakeholders at these companies. We’ve recognized that most stakeholders have a hard time relating this abstract technology to their company’s operations. So we created Adventures in AI to help them understand NLP and the practical value it can generate.
Rather than describe a project, we told our story through two protagonists: JD Campbell, an innovation officer, and Lori Santos, a data scientist. So much of AI innovation is about relationships—building trust between teams. Our comic is a primer in the motivations, biases, emotions, and actions involved in those relationships.
One part of the comic has sparked a considerable reaction from readers. Lori’s outsized intellect is only slightly larger than her ego. She gets easily frustrated by the plodding pace of corporate innovation. Here’s a quick look at her frustration:
JD cares about Lori and recognizes the source of her irritation: she is bored. Rather dismiss her as an egotistical jerk, he looks for an opportunity to release her energy on a useful challenge.
The relationship between JD and Lori triggers an emotional response in our readers.
Many business people who read the comic are turned off by Lori and her attitude. We’ve received many emails along the lines of, “great story, but the data scientist is kind of a jerk. I found her ego to be an annoying distraction and can’t understand why JD suffers her.”
But many data scientists love Lori and constantly share reactions like, “OMG, I feel exactly like Lori!”
Of course, this tension was our goal. We wanted to trigger an emotional response through the relationship between these people.
Stories create emotions, emotions create beliefs, and beliefs drive behavior.
It really is that simple.
Lowe’s Innovation Labs follow a process for making stories a foundational part of corporate innovation. They partnered with the Savannah College of Art and Design to develop comics, videos, and images based on futuristic narratives. Even simple sketches can tell a story about a technology’s potential.
This video visualizes a mini-story of a family using augmented reality to shop at a store:
Stories help Lowes overcome the primary challenge for corporate innovation programs: sustaining a shared, long-term vision. Maintaining momentum for innovation through changes in leadership, markets, and more is an ongoing challenge.
Guess who has this same challenge? You, of course. Becoming an AI-driven company takes years of focus and investment. Stories can help you maintain your momentum.
You are a fan of science fiction. Maybe you grew up watching Star Trek, or A Wrinkle in Time was your favorite book. Heck, maybe you still love comic books. A love for fantastical stories goes hand in hand with a passion for AI.
You are one of those rare people who is excited about the potential of this technology. Whether you recognize it or not, you’re constantly telling yourself stories about how it will change your company. Start telling those stories at work to evangelize your AI program. Although making a comic is a blast—seriously, you will love it—any story you tell about AI will make your job more interesting.
So give it a shot.
Have a great example of storytelling about AI? Send it to email@example.com. I’d love to hear about your adventure.
To infinity and beyond …
May 2, 2021
Your goal as an AI leader is to get your teams to think like pros. You want them to strategically look for ways in which AI can lift the entire business instead of just solving a narrowly defined problem. Your team should constantly seek ways to advance the bigger vision of becoming an AI-driven company. In this issue of FeedForward, I’ll describe the difference between how pros and amateurs think about AI.
March 31, 2021
In this video Justin Pounders, Director of Machine Learning and AI Research at Prolego, breaks down natural language generation (NLG) into its most basic components and describes how you can begin building out these components in your business. (And, no, it doesn’t depend on GPT-3!) He describes how NLG depends critically on two questions (WHAT you want to say and HOW you say it), the types of data you can feed into NLG systems, and a development path for being able to summarize multiple sources of data in plain English.
March 30, 2021
Like most engineers, I hate tedious work. That’s why I love the idea of automatic machine learning (AutoML). As much as I want to love AutoML, it’s been incorrectly framed as a substitute for data scientists. This confusion arises from a misunderstanding of what actually happens in machine learning projects.