9.12.24: MIT STEX Demo Day - Lighty AI

RICHARD RABBAT: And welcome to the start of Exchange. Thank you so much for having me. I will be speaking about Lighty AI. Let me share my screen. So I'm Richard Rabbat. I was previously VP of Product at Twitter. I've exited my previous startup that I sold to Snapchat, and I've spent a number of years at Twitter, Google, Zynga, and other places, 20 years experience and a PhD from MIT co-founding team, as all MIT alums, as well.

And the problem that we decided to address is the fact that with AI in general, most of these models have been actually trained on public data as opposed to capturing the unique DNA of the enterprise itself. So these AIs that you use today, are not trained on your private data. They're just trained on publicly available data. Meanwhile, everybody is under pressure to adopt AI because of productivity improvement potential.

And Lighty AI is an agentic AI that allows a company to capture the DNA of that enterprise. So our LLM is very specific to the enterprise and is powered by a knowledge graph. And that knowledge graph is all the data that is collected that is private to the company emails, wikis, , documents, calendars that allow our AI to be much more effective when it comes to supporting the organization.

And the things that we talk about, in terms of workflows, are things all the way from the beginning, from onboarding a person, to planning and coaching. Because these days we're in a situation where people are looking for help to actually improve their productivity.

Scheduling is a very important task because typically, people are very interrupted all the time. And our AI helps you with scheduling, task assignments but also tracking of your objectives and key results. All of this is based on personalized interactions so that the AI really speaks to you, as an individual, the team, and the organization.

Here's an example of the AI agent at work. In this case, it is helping with scheduling a meeting. Instead of me constantly looking at my calendar and trying to offer some times, the AI will just look at my calendar, propose it to the other party, schedule it, and then put it on the calendar. And I'm not interrupted. Every interrupt today, actually costs us about 20 minutes of trying to get back in the zone. And Lighty basically, does away with all these interruptions.

And it's all powered by a knowledge graph that is private to the enterprise. This is something that we're very excited about where on the day before your next day, Lighty sends you an email about, like, a reminder of what your plan is for the next day and highlights some issues.

In this case, I'm looking at my schedule, and it looks like I'm all set. If I need to make some changes, I will ask it-- like in this case, I want to cancel a meeting. I'm going to cancel the daily leadership sync. And Lighty does it. I don't have to get out of the flow, and it just does it for me.

And then finally, as Lighty learns about you and your goals, it looks at how you can actually be better at how you manage your time. In this case, I've told Lighty that I want to spend more time on deep work, and then it looks at where I'm spending my time. And then it tells me that, in this case, I should actually spend more time on blocking times for my deep work and other things, as well.

And then finally, we all come out of meetings, and we've had like a lot of action items that we sometimes forget. What Lighty does is actually capture all the transcripts and turn them into action items and follow-ups so that you're effectively conducting your meetings and the post-meeting work that you're doing.

We're looking for additional pilots. We already have some companies, enterprises, and mid-market. In this case, we would like to work with more enterprises to help teams accomplish more and deliver on their goals and objectives. We propose a two-month pilot where we identify specific KPIs to deliver on and really solving about each [INAUDIBLE] enterprise's unique problems.

Thank you, this is my email. I'd love to have more conversations with you as we progress through this demo day.

SPEAKER: Thank you, Richard. Let me ask a couple of questions. Again, a reminder, post your questions in the Q&A. And I'll start with one. Some companies are still trying to figure out where best to use AI? So how would you work with those companies?

RICHARD RABBAT: So we actually spent some time working with the companies, asking about what problems they are trying to solve. Because AI is not about putting a technology and just making it available, so you can say, I'm using AI. It's about solving specific problems.

In this case, we look at specific ways of improving productivity, where we see that a company-- like, one of the teams struggling to like meet their goals and objectives. And then we work with them and we bring Lighty to help with improving that productivity.

SPEAKER: And you stay on-- what's the--

RICHARD RABBAT: Yeah so it's a platform that becomes available. At first can be used by a team and then we usually extend it to other teams as people get to see the value and then they can onboard to the platform, as well.

SPEAKER: Great, and then another question, can you combine it with other communication apps like, Teams and others?

RICHARD RABBAT: Yes so this is like what's really great about Lighty. Lighty is a platform that connects with multiple ways like Teams, Outlook, Slack, and other systems, Notion and whatever you use because we know that communication happens and happens in different ways. And we adapt to each company's unique way of communicating.

SPEAKER: OK, Richard, thank you so much. And again, you can engage further with Richard in the chat and in the breakout rooms.