The future of words: augmented writing
In 2014, I co-founded Textio to build a company around a novel and potentially industry-changing idea: what became known as augmented writing.
The light bulb of inspiration came in the summer of 2014 as part of a lengthy and sprawling conversation with Kieran Snyder about the history of writing software (something I’ve been involved in my entire adult life.)
We realized that all of the writing software in the post-typewriter era was missing something essential. There’s software to help you decorate text—to help format it and make it beautiful. There’s software to help you share your writing, and to collaborate in the cloud with others.
But there had never been software to help you actually write the right words.
Suddenly we realized that, for the first time, the advent of machine learning and easily-available cloud computing meant that creating this kind of software was possible. We could build software that would help you write the right words based on data.
It felt like a huge opportunity, and so we founded Textio together to build this new category of software—something we eventually came to call augmented writing.
Bringing augmented writing to life required not just inventing new technology but also reimagining the fundamental user experience of writing. Things that had seemed so important in the old era—italics, fancy styles, print preview—were suddenly replaced by new, more important concepts such as guidance and data visualizations.
Now, just a few years later, augmented writing is seemingly everywhere, and even the biggest players such as Google and Microsoft are following down the path we created. The concepts we dreamed up during the lazy summer of 2014 seem poised to redefine what people expect whenever they’re writing words in the coming decade.
A closer look
Although Textio has grown beyond 100 employees now, we started out with just a few people. So, in our earliest days, I designed the entire user experience and wrote the frontend code myself. Coming from my previous job leading a team at Microsoft larger than our entire company is now, this was tremendously fun.
I had the chance not just to invent so many new things and play with new interaction concepts, but also to actually write the code and ensure that all the little delightful details I cared about felt exactly right.
At its heart, Textio analyzes your words as you write, telling you how well they will work. If you are writing an email, for instance, Textio can tell you how likely someone is to read it or respond positively to it. Textio knows this because its technology has been trained from the outcomes of over 1 billion pieces of writing.
The Textio Score encapsulates all of this machine learning and natural language processing into an easy-to-understand 0 to 100 score. We tried lots of other visualizations, but the simple, single score model proved key to people understanding and accepting Textio’s advanced guidance.
Another really important aspect of augmented writing is the notion of being able to tell who will respond to your writing. For instance, will your writing attract more people who identify as men? Will it attract older people but turn off people in their 20s?
Understanding who is likely to respond based on the language you choose helps people to write more inclusively—to write things that will appeal to everyone.
Designing the right visualizations to illustrate these concepts was tricky and took a lot of iteration and testing. In the end, the solutions we implemented have worked well and have started to be duplicated by newer augmented writing companies.
One of the really interesting challenges in designing Textio’s user interface is that it can plug in on top of other places people already write, like Gmail, Outlook, or LinkedIn.
Building the UI with this scenario in mind is a complex. We consider which things translate well as a “layer” on top of the existing interface, and which need to be clarified, simplified, or changed.
We weren’t willing to compromise on the quality of the user interface, so we’ve put a lot of sweat and ingenuity into perfecting this hybrid experience.
More information
The best way to learn more about augmented writing is by visiting the Textio web site. You’ll find a bunch of great resources about what augmented writing is and how it works.
Augmented writing is just the most recent revolution in the never-ending march forward of communication software. Let’s stop writing like it’s 1995 is a short piece I wrote that puts this paradigm shift in context of what came before (and what’s coming next).
It’s been a long path from typewriters to terminal-based word processors to the the 1990s formatting nirvana of WordPerfect and Microsoft Word to the cloud era of Google Docs collaboration. If you’re interested in learning more about the history of writing software (with lots of fun pictures), I think you’d enjoy this quick read: The dawn of the augmented writing era
I’ve also done a few podcasts in which I talk about different aspects of augmented writing and the process of founding a company around it. You might enjoy them!
Lean Startup Podcast: How Moving to a Startup Can Be Rocket Fuel for Your Career
a16z Podcast: The Product Edge in Machine Learning Startups