Artificial intelligence may not yet be able to replace Google search, but it can be useful in more specific contexts – for example, dealing with the drudgery of everyday tasks such as scheduling meetings. This is the premise of a modern startup, Skiswhich offers an AI assistant that you can loop through your emails to find the best time for everyone to meet.
Unlike other scheduling solutions like Calendly, Skej doesn’t require you to view someone’s availability to find a meeting time. In fact, if someone sends you a link to Calendly, Skej will scan the link to find times when you’re both available and then put the meeting on their calendars.
“I have never met anyone in my life who loved planning meetings,” says Skej co-founder and CEO Paul Canetti.
The Fresh York serial entrepreneur, who previously founded and sold the no-code app development platform MAZ Systems, also worked on another meeting startup called Bounce House. In this case, the service allowed users to pay to book blocks of time with professionals such as yoga or piano teachers.
The same founding team from previous efforts and others returned to work on Skej, including: Canettihis brother JustinChief Technology Officer Anindya Mondaland fourth co-founder, Szymon Baumer, who died of cancer three months after founding Skej in August last year. (The team has tribute page Simon on the Skej website, crediting him with creating “the basis of today’s product.”)
As Paul explains, Calendly is useful and has built “an amazing business,” he says, but he didn’t like posting about every free time he had. The only time he was really cheerful with planning was when he had a human assistant, like an EA. Unlike a technology platform, a human can easily understand the context of meetings and know whether to shuffle your calendar to accommodate someone significant, even if you’re scheduled to be busy, for example. This led to the idea of creating an AI assistant that could do the same.
To exploit Skej, you don’t need to download the app or visit the website – just add her email address to your conversation. Later, Skej will also have a phone number that he can add to text chats. This service currently works with any email platform such as Gmail, Outlook and others. It also currently integrates with other programs like Zoom and Google Calendar, with support for Outlook Calendar coming in the next few weeks
Using Skej only requires adding an email address to the conversation and then asking it to find meeting times in the response. For example, when TechCrunch scheduled an interview with Paul, he responded, “Skey, can you suggest a few hours that might work this week?” and the AI assistant emailed me options and a link to automatically connect my calendar to find the time. After responding to my preferences, Skej replied that the meeting had been set and added it to my calendar.
The system works because Skej’s user – in this case Paul – has allowed him access to his calendar. The skater simply sent an invitation from the calendar on his behalf.
However, if I clicked on the attached link, Skej could automatically book the meeting without having to repeat the booking. This last option works best for internal teams where multiple people need to meet to find a time slot that works for everyone in the group.
Under the hood, Skej uses a variety of LLM models, including those to interpret the language of email messages and then break it down into data fed into Skej’s proprietary system.
“Internally we call it the brain… and the Skej brain works like a scheduling engine, almost like a matching market,” Paul says. “This allows it to accommodate different people, in different time zones, with different views, different conflicts and different preferences,” he continues. “And I try to negotiate to find a match. Then… it displays the match or suggested times or data, and LLM helps you create a message that will sound natural once it appears,” notes Paul.
Skej also allows users to categorize different contacts to associate with different calendars, such as a work calendar or a personal calendar. Paul believes that over time Skej will be able to enable this type of categorization also using natural language. For now, there’s a more customary dashboard where you can configure your preferences and integrations.
One thing Skej doesn’t plan to do is build an app.
“It’s humorous, this is a question we get a lot from VCs…we think, ‘Well, you’ll eventually have an app, right?’” Paul says. But Skej, he says, is designed to “completely ignore tools that you already exploit and love, and that can be adapted to whatever workflow you’re already doing,” he explains.
“It doesn’t force you to exploit a specific app or a specific thing,” he adds.
Skej’s pre-seed investors include Betaworks, Mozilla Ventures, Stem AI, Spice Capital, Deftly.vc and Differential Ventures. The round didn’t exceed a million, Paul says. Skej’s remotely distributed team consists of three co-founders and two other full-time engineers.
The service, currently in public beta, is currently used by over 1,000 users. Skate is free for now and the team is gathering feedback, but will add a paid tier later.
