Written communication has been around for thousands of years, but we still look for ways to write better and faster. We went from carving stone tablets, to dipping quills in ink, to scribbling pen on paper, and finally to typing on keyboards and touch screens. In this world of supercomputers and big data, it feels like a matter of time until the next communication revolution.
Automation will be a game-changer for how we write. We need to find a way to use computers to not only write faster, but also to write richer content. It’d be amazing to have AI that learns what your audience likes and writes an amazing piece in your own voice… which I can then use to blog without doing any work!
Are you predicting that AI will write for us? Don’t chatbots already do that?
Yes they do, but when’s the last time you used a chatbot that was actually good? Don’t get me wrong — they’re great for automating things that are very repetitive and predictable, like cliche customer support questions — “How do I get a refund?”
However, they’re unreliable for conversations that go “off script,” which comprise the majority of our day-to-day interactions. A chatbot would not respond well to thoughtful prompts like “What’s your favorite thing about this product?” or “Why are you building feature X instead of feature Y?”
Chatbots aren’t good enough today, and they won’t be good enough to take over the world tomorrow. There will always be a need for humans to communicate personal thought and creativity, which is something that machines won’t replicate anytime soon. At the end of the day, full automation is extremely difficult and often not even desirable.
If it’s so hard to automate writing, how will AI enhance written communication?
We need a different type of tool — one that allows people to freely express what they want to say while automating the repetitive work. We need a tool that uses AI not to replace humans, but to supercharge humans.
The tech giants began to see the value of a tool like this. As recently as May 2018, Google released Smart Reply and Smart Compose in Gmail, which help users quickly insert short snippets into their emails. Facebook and LinkedIn/Microsoft have released similar features that suggest short replies in their messaging platforms.
It’s exciting to see the AI landscape shift in this direction, but these features are still limited to short, generic replies. There’s so much potential to add more personalization and complexity to these reply suggestions — which is exactly what we’re working on at Point.
What does Point want to accomplish?
We want in on the next communication revolution. We saw the foundation set by Google, Facebook, and Microsoft and have taken the next step by building our own engine: the Point API.
The Point API predicts what you want to write and suggests text for you to use. After plugging it in, you receive suggestions that you can just click to add to your writing. You can connect the API to any online communicator — Slack, Gmail, Facebook Messenger, a random iPhone app, your company’s internal messaging tool — you name it. The Point API also quickly updates suggestions on-the-fly when context changes. Your suggestions will be different if you’re texting your friend about an upcoming barbecue than when you’re pitching your product in a sales email.
The key component that separates us from other services is personalization. Point works closely with you and text you’ve written to build AIs that sound just like you. On the other hand, existing services use aggregate user data to create generic suggestions, which are accessible to everyone but are very simple and not always useful.
In short, we developed a new AI-powered system that predicts how you write in order to supercharge your communication platform. We envision a future where an AI automates mundane writing and suggests awesome content, yet still gives you the freedom to use your own voice. Will this type of tool truly be the future of written communication? We think so, and if it does, we’ll be there to let you know.