Computer-assisted writing technology has come a long way from the first spell-check software. Today, there are tools that can check that your writing is on brand, uses the right tone, is easy to read, varies in vocabulary, and doesn’t include bias. And that’s just a shortlist of what’s available.
Powering these types of writing assistants is artificial intelligence. AI can analyze large amounts of data to determine patterns and scan for correct usage—saving writers and editors time and helping them catch errors.
To get more insights into AI-powered writing tools, Content Science Review spoke with Amy Cuevas Schroeder, director of content at the AI writing assistant technology company Writer (formerly Qordoba).
SCHROEDER: There are a variety of AI writing assistants to choose from now, and we worked with an analyst to vet the benefits of each of them by use case. Content marketing specialists use AI writing assistants to make sure their writing is engaging and written in a consistent brand voice across platforms, including websites, blogs, and social media. In addition to editing, some AI writing tools help with SEO strategy.
I’ve been producing content in pretty much every format for 20 years—starting in journalism and magazines, then for advertising agencies, then for companies like Etsy and Minted, and now for Writer. I think of Writer as helping to scale my brain, or at least my copy editing process, so that I can focus on creating original content. A good AI writing assistant goes beyond catching spelling and grammar mistakes; I like Writer because it helps identify wordiness, passive voice, and helps me make sure that I’m staying on-brand and consistent with our messaging.
Lisa Young, a content strategy manager for Twitter, says Writer is their “content guardian angel,” and I think that’s a great way to think about an AI writing assistant.
SCHROEDER: At Writer, we use AI technology mainly for editing—to make sure our content is on-brand, using our latest and greatest messaging, and written in a clear voice and tone that our readers will understand. With the rollout of our new diversity, equity, and inclusion technology, we’ll be even more confident that our language is inclusive. We’ve launched a new feature I’ve been using internally in beta for a while called “snippets.” Snippets allow you to easily insert chunks of brand messaging, such as a common call to action or value proposition message, into anything you’re writing.
SCHROEDER: We recently created a timeline that gives a snapshot of AI’s evolution in the last 50 years. It’s interesting to think back to 2007, when Apple welcomed autocorrect to mobile OS. In the last decade, machine learning, natural language processing (NLP), and neural networks have made major strides, allowing for the next-generation of writing assistants to be AI-based.
Advances in deep learning (i.e., neural networks) meant that massive corpuses of well-edited content could train a grammar model on what good grammar looks like—no complex rules necessary. Further advances in machine learning made setting up and deploying these models into products that were easy for users became easier, too. We have been the first to move from grammar to other models: what is healthy communication? What is on-brand communication? And we’ve applied the same machine learning technology to give people those kinds of copy editing suggestions.
SCHROEDER: We foresee AI-powered technology continuing to help people at work and in their daily lives. We believe AI should help people, not replace them.
At a time when Slack, Zoom, and the comments/chat fields of just about every B2B tool have quickly become the way we talk to our co-workers, sometimes we—and our thoughts and ideas—don’t come across as intended. We’re realizing that now more than ever, individuals and companies need more control and insight into what their teams are writing. Consumers are paying close attention to brands’ voices and messaging. Advances in AI and NLP mean there are now automated solutions for this. The future of communications at work will include an automated double-check before you hit enter or click send, that what you’re about to communicate is what you actually want to say.
SCHROEDER: Like any new technology, the challenges come down to data and people. If you are a data scientist or machine learning engineer interested in the future of human communication, get in touch! On the data side, it’s how machines learn, of course. We’re obsessed with customer privacy and don’t use any of our customers’ data for training, so it does mean a constant effort to build datasets that we can use to build models.
SCHROEDER: Try to clone yourself! Creating a solid styleguide is an important first step in being able to use AI in your work. If you don’t have a styleguide, or need to refine it, check out styleguide.com to create a styleguide from scratch or see examples of other organizations’ styleguides. Writer allows you to incorporate your styleguide and quickly make sure that all your content is one-brand and hitting the right tone.
Content that uses emotive language performs nearly twice as well as purely factual content. Learn more in this guide from Acrolinx.
Learn why one page is rarely enough to rank for competitive topics and how to build a content cluster that positions you as an authority in this MarketMuse whitepaper.
Make better content decisions with a system of data + insight.