Since the explosion of large language models (LLMs) began, companies have been racing to tackle the issue of overflowing inboxes by using artificial intelligence to sort emails and craft replies that mirror your personal voice. Email client Superhuman is rolling out an updated version of its auto-draft feature, designed to spot important messages and generate responses that feel less robotic. Superhuman has dabbled in similar territory before with tools like instant replies and follow-up auto-drafts. But many of those earlier attempts came across as overly enthusiastic AI sales pitches, so I rarely used them. The new auto-draft feature, however, feels markedly different. Over the past few days, after gaining access to the beta, I’ve sent several emails with little to no editing of the generated drafts. The app identifies which messages likely need a reply and composes a response based on the tone of your past conversations. It also offers two alternative variations you might prefer to send instead. In my testing, I saw drafts agreeing to embargoes on pitches to request more details, or confirming meeting times that I could fire off with minimal tweaks.

That said, the feature is far from flawless. By default, it often generated a positive response to a pitch or agreed to a meeting scheduled for after midnight. Luckily, I could quickly select another option from the other variations and send it off. The system learns from your behavior and refines its responses over time. For example, after the midnight meeting mishap, when someone suggested a similar time, the feature produced a draft stating that the timing doesn’t work for me. I receive thousands of emails each month, partly because AI makes it easier for others—like communications and PR professionals—to draft first messages. I’m not ready to fully hand over my inbox to AI, but this feature could help me respond to more people without having to type out lengthy messages. Users can personalize emails by navigating to Settings > Personalization and adding details about themselves and their role, along with uploading files or links for extra context. Superhuman co-founder Rahul Vohra noted that during the testing phase, 40% of auto-generated drafts were sent within one day, and 60% of those were sent without any manual editing. Vohra explained that earlier features like instant replies were built on older models such as GPT-3.5, which were less intelligent or had a smaller context window. With this new implementation, the company is using a mix of models. “Today, we are using a mixture of models to make this work. The actual writing is done by frontier models from both Anthropic and OpenAI. So we’re applying the maximum amount of intelligence and context to this that we possibly can to make the feature work,” Vohra said. Last year, Grammarly acquired Superhuman and then rebranded the company as Superhuman. Now, the company is building an assistant called Superhuman Go that spans platforms while carrying context over from one app to another.