AN INTERVIEW WITH STORYFIT FOUNDER AND CEO, MONICA LANDERS
Once a science fiction fantasy, artificial intelligence could be the biggest technological breakthrough since the internet. StoryFit, an Austin-based technology company, was created to introduce AI directly to the publishing, television, and film industries. By using scalable, dynamic metadata and deep analysis of written content, StoryFit uncovers narrative trends and insights that inform market and sales strategies. Ultimately, its goal is to use AI to help producers battle the challenges of today’s entertainment market. To learn more about how artificial intelligence can impact the industry, we interviewed StoryFit (previously Authors, Inc.) founder and CEO, Monica Landers.
Q: HOW IS AI USED TODAY?
Considering that most applications of AI before 2010 were academic, it’s truly amazing how widespread it is today. It’s nearly impossible to avoid it at this point. Practically every business sector is using it some way to speed up decision making, identify fraud, anticipate consumer preference, support customer service, scheduled flights, etc. It’s everywhere. Some of the most successful applications of modern AI are: Siri, Alexa, and Google Home Arguably the most ubiquitous instance, each of these is an AI superpower in your pocket (or on your shelf). Amazon Their algorithmic AI used to anticipate consumer wants and needs has led them to become one of the strongest economic powers in the private sector. Financial Trading All those fast and secure trades on the Chicago Mercantile Exchange and New York Stock Exchange? Yep, that’s AI. Data Security The need for AI that thinks as fast if not faster than hackers is needed more and more every day. Most malware uses borrowed code–only between 2 and 10% change between iterations. Companies like Deep Instinct and Kaspersky Labs use AI to train learning models to predict the files that will hold that borrowed code and prevent security breaches and data theft.
Q: HOW IS AI USED IN PUBLISHING?
Although many industries broadly employ data analytic tools, publishing’s use of data has mostly restricted itself to book sales and distribution. In publishing, big data’s promise has been only partially met. Until now. Artificial intelligence (AI) has wide-reaching applications and implications–so much so that it can be understandably difficult to see the forest for the trees. This technology has real potential to soothe some of the most common aches of publishers and producers of narrative media. AI holds a lot of future payoff for publishers who implement it early and often. Since it can learn, the real issue is what question do you want to tackle first? At StoryFit, we see direct and immediate impacts to discoverability, audience targeting and engagement, recommendation engines, efficiency, and acquisition risk assessment. Specifically, AI in a publishing and production environment can perform the following: Metadata & Discoverability Data integration Reader Analytics Improve Audience Targeting Increase Audience Engagement Predict mass audience (bestseller) appeal
Q: HOW RELIABLE IS AI ANALYSIS VERSUS HUMAN ANALYSIS?
AI is as good as the data it’s fed and the direction it’s given. In this case, AI is created through a process of machine learning, whereby it basically uncovers and then replicates automated decision-making based on the parameters it’s delivered. Once the algorithm is well-trained and directed, then yes: AI is reliable. For one thing, it’s consistent. It doesn’t get tired, need to take a break, or get distracted by personal reading preferences. Humans are amazingly creative, observant, and resourceful. But at a certain point, our brains just can’t process any more data any more quickly. Meanwhile, a computer using AI continues doing the task at hand, not changing its analysis due to fatigue, overwork, boredom, or a headache.
Q: SEEMS LIKE THERE WERE CAVEATS IN THE PREVIOUS ANSWER, SO DOES AI MAKE MISTAKES?
A: Though machine learning algorithms mathematically minimize bias and error-inducing variance, some data has a strong bias, like salary or occupation. Looking at a set of text with those elements, it’s difficult to “eliminate” that bias. The best thing to do is to review the results, make note of the bias, and either inform or correct for it (depending on the task). Just like the humans that write the programs, when given incorrect/incomplete information, AI will make mistakes. It’s not magic, and that’s why the diversity and quality of all the data is important to the process.
Q: HOW DOES AI ACCOUNT FOR THE EMOTIONAL CONNECTION A READER MAY HAVE WITH A LITERARY WORK?
We have to create rules for the computer to recognize the emotion because of course it can’t connect the same way humans do. Those rules start with natural language processing and sentiment analysis which are able to infer the emotions different linguistic styles, circumstances, diction, and perceived emotional arc evoked in readers. At a glance, sentiment analysis is fairly straightforward: text is analyzed and, using natural language processing, each part of the text is categorized positive–happy statements–or negative–sad/angry statements. Within each language, words can be determined positive (elated, kiss, jump), negative (smash, kill, cry), or neutral (the, a, road). Take these altogether and graph the results, and you can see the emotional arc of a text mapped out in a physical form, called a sentiment map. This is just the beginning, from there we can group the language, with the surrounding action, and the thousands of other features we measure to infer how a human might connect to a particular emotional arc. We can then check our work with real humans!
Q: REVOLUTIONARY LITERARY WORKS CAN SOMETIMES BE UNCONVENTIONAL. HOW DOES AI PREVENT A GOOD/GREAT, BUT UNCONVENTIONAL STORY FROM “SLIPPING THROUGH THE CRACKS?”
The goal of machine learning or artificial intelligence is not to replicate human creativity, but instead, amplify it. At StoryFit, our AI is not built to proscribe quality or replace the editor. It empowers them with more data than they could ingest in a lifetime, distilled into actionable insights. The benefit of a computer doing the analysis is that it can discover the latent aspects of a novel. So, while authors may have their own identifiable style, there are underlying components of a novel that a computer can see better than humans. Books hold undeniable cultural weight– and they are also data. Unlike consumer data that, even formless, has proven itself so valuable, text has inherent rules and formations. And each deviation from one rule has its own set of rules, deviations, and acceptable actions that AI can learn, compare, and predict, just like humans can. When book clubs discuss a possible connection of a white lily to the heroine’s inevitable death, they are discussing the pattern identified in a semi-closed system. They are performing one tiny fraction of the computational potential of an algorithm trained on literature.
Q: HOW DOES AI LEARN ON ITS OWN? HOW DOES IT KNOW WHAT TO “READ”?
Essentially, we provide the data for the model to train the AI. So, we tell it what to read. This is the crux of machine learning. If done improperly, you get a useless algorithm.
Q: WHAT IS YOUR COMPANY DOING TO MAXIMIZE AI CAPABILITIES?
Our artificial intelligence service is specifically built to solve the most common challenges publishing and entertainment companies face in delivering the right content at the right time and to the right audience. We bring deep content and audience insight to editorial development, marketing, and sales while reducing risk, costs and delays.
StoryFit integrates data, elemental story analysis, and market research to improve audience targeting, determine market fit, and mitigate investment risk. Everything you need to quickly and painlessly make informed decisions, edge the competition, and drive consumer engagement. Publishers can take advantage of this in three applications: StoryFit Intelligent Metadata StoryFit uses AI to craft metadata that increases discoverability, engagement, and sales. Supplying metadata is proven to enhance discoverability–accurate metadata further increases both discoverability and sales for new and existing content. AI scans the text of a story and extracts the words, phrases and terms that describe the story’s content, impact, and sentiment and pulls the key elements important to readers–character types, conflict types, setting, era, and more–for your keywords. These words are compared to and filtered through dictionaries, topic models, Google trends, and best industry and Amazon practices to determine their effectiveness to drive consumer discovery. The software further analyzes, tests and updates the keywords for improved performance. Intelligent Metadata generated by StoryFit™ AI provides our customers with the right metadata, optimized for accuracy, performance, and that conforms to major distributor/resellers’ requirements. Intelligent Metadata allows publishers to give each title more room to succeed and breathe life into classics or titles that just never found their audience with optimized keywords that respond to sales performance and market behavior. StoryFit Content Analytics Our content insights and analytics enable publishers and studios to learn their stories’ secrets, support decisions, pitches and process with data built from industry bestsellers and veteran wisdom. StoryFit™ reads your text in seconds and measures it against key points that have proven correlations to commercially successful outcomes. The metrics result from analysis of thousands of high-performing titles and allow publishers and producers to find new titles with the best chance of success and better position existing investments, identify better comps, and kick-start marketing. Included in the content analytics are market readiness ratings, style insight, story arc, faceted comps, accessibility index, and predictive audience profiles. StoryFit Custom AI StoryFit AI can learn and adapt to its environment; by listening to your business problems and analyzing your specific data—be it customer profiles, sales history, web behavior, and more. — StoryFit can tailor a solution just for you. By applying our algorithms to your proprietary data and publicly available information you can create a tactical maneuver that amplifies your strengths and allows you to move faster than your competitors. Just some ways publishers can design their own AI with StoryFit are to: Superpower your own recommendations engine Rapidly identify important performance indicators and categorize content to suit your marketing goals Use own top performers as a rubric to find new material and uncover potential powerhouse titles in your backlist Surface new audiences for old content, and new content for your loyal audiences Identify manuscripts that match your unique editorial tastes and generate comps from your own library
Q: HOW DO PEOPLE USE STORYFIT?
StoryFitTM technology is artificial intelligence employing natural language processing, machine learning, and proprietary algorithms to tailor solutions for the entertainment industry, specifically book publishers and movie and television producers. StoryFit is used by book publishers, book publishing marketers and consumer engagement specialists as well as movie producers and television studios. Our goal is to provide robust artificial intelligence that brings data to decision-making and drives consumer discoverability. This year, we’ve begun pilot programs of our metadata service with over 15 publishers and content insights with movie studios, too. In 2016 the Book Industry Study Group released a study that found titles with keywords filled out in their ONIX messages saw 34% higher sales than those without–and the overwhelming majority of titles available on the market either don’t have them or don’t have keywords that are optimized to facilitate consumer search and product merchandising. It’s early days, but our metadata customers are reporting positive revenue gains for the titles that utilize our metadata. Despite all this evidence, most publishers don’t have a successful keyword program in place for new titles, let alone for backlist fulfillment. But with AI-generated keywords, suddenly they can enrich the metadata of their entire backlist in just a week. Our technology continues to power the StoryFit marketplace, where we offer subscription packages that include content analytics, which customers report using to negotiate with agents, plan targeted Facebook and Google Adwords campaigns, and more effectively communicate editorial direction with writers.
Q: WHAT DO YOU SEE AS THE FUTURE OF AI?
Moving forward, those who adopt AI will be able to move faster and be more dynamic in an ever-changing marketplace. Complex problems will be tackled and opportunities for growth--and even creativity--will emerge. Who wouldn’t want technological advancements that make life better? Whether it’s making a task less tedious, a moment more enjoyable, or something more accurate, it’s in our nature to improve. Art and Science aren't exclusive. And likewise, human ingenuity and creativity will be enhanced by AI.