The relentless flood of emails is a modern-day challenge, and for years, companies have sought to tame the overflowing inbox with the promise of artificial intelligence. From categorizing messages to drafting responses, the goal has always been to make AI-powered solutions sound as natural and personal as possible. Now, email client Superhuman is rolling out a significantly enhanced version of its auto-draft feature, aiming to revolutionize how we interact with our inboxes by creating AI replies that truly sound like you.
Superhuman has previously ventured into AI assistance with features like instant replies and follow-up auto-drafts. While these were innovative for their time, they often felt a bit too rigid or, as some users described, like an overly enthusiastic AI salesperson. The good news? This new iteration of the auto-draft feature feels remarkably different and genuinely effective. After getting early access to the beta, I’ve found myself sending generated drafts with minimal to no editing, a testament to their improved quality.
This sophisticated new feature intelligently discerns which emails demand a response and then crafts a draft based on the unique tone of your past conversations. It doesn’t stop there; to give you ultimate flexibility, it also generates two additional variations you might prefer to send. This approach ensures you always have options that align with your communication style, even if the first suggestion isn’t quite right.
My personal experience with the feature has been largely positive and incredibly efficient. I’ve seen drafts agreeing to embargoes on press pitches, confirming meeting times with just a quick glance, and even politely declining requests for authored posts on TechCrunch. The ability to send these with minimal or no edits has been a significant time-saver, streamlining tasks that previously required manual typing.
Of course, no AI is perfect, and this feature still has its quirks. Occasionally, it might default to an overly positive response to a pitch or suggest a meeting time in the middle of the night. Thankfully, the availability of alternative drafts means you can quickly select a more appropriate response and send it on its way, maintaining control over your communications.
Crucially, the system is designed to learn from your interactions, refining its suggestions over time. For instance, after correcting a draft that proposed a post-midnight meeting, the AI began generating more sensible options when similar times were suggested again. This adaptive learning is key to the feature’s long-term effectiveness, making it a more personalized assistant with every use.
For someone like me, who receives thousands of emails monthly, often boosted by AI-generated first drafts from others, this feature is invaluable. While I’m not yet ready to fully hand over my inbox to AI, this tool significantly helps me respond to more people without the burden of typing out lengthy messages. It offers a powerful middle ground, boosting efficiency without sacrificing human touch.
Understanding Your Tone: How Superhuman’s AI Works
Personalizing your AI-generated emails is incredibly straightforward within Superhuman. Users can head to Settings > Personalization and add specific details about themselves, their role, or even include files and links for more context. This deep customization ensures the AI drafts replies that truly reflect your professional persona and current commitments.
The impressive effectiveness of this new auto-draft feature isn’t just anecdotal. Superhuman’s co-founder, Rahul Vohra, shared compelling statistics from the testing phase. He revealed that 40% of all auto-generated drafts were sent within one day, and an astounding 60% of those were sent without any manual editing whatsoever. These numbers highlight a significant leap in AI-powered communication quality and user trust.
The Advanced AI Powering Superhuman’s New Feature
The superior performance of this latest auto-draft feature is largely due to a significant technological upgrade. Vohra explained that earlier versions, like Instant Replies, were built on older models such as GPT-3.5, which had less intelligence and smaller context windows. This new implementation leverages a far more advanced and sophisticated AI architecture, marking a substantial leap forward.
Today, Superhuman employs a sophisticated blend of AI models to achieve its impressive results. “We are using a mixture of models to make this work,” Vohra stated, elaborating that “the actual writing is done by frontier models from both Anthropic and OpenAI.” This multi-model approach ensures the maximum possible intelligence and contextual awareness are applied, making the feature remarkably effective and human-like.
Beyond the Inbox: Superhuman’s Broader Vision
Beyond individual email responses, Superhuman’s evolution continues at a rapid pace, especially following its acquisition by Grammarly and subsequent rebranding. The company is now developing Superhuman Go, an ambitious assistant designed to operate seamlessly across various platforms. This intelligent assistant will carry context from one application to another, promising an even more integrated and productive workflow for users.
Superhuman’s new auto-draft feature represents a pivotal moment in the quest to make AI email replies genuinely useful and human-like. While it doesn’t eliminate the need for human oversight, it significantly reduces the cognitive load of managing a busy inbox, allowing users to communicate more efficiently and effectively. It’s a compelling demonstration of how AI can enhance, rather than replace, personal communication, pushing us closer to a truly intelligent email experience.
Source: TechCrunch – AI
