2024 MIT Digital Technology and Strategy Conference: Lightning Talk - Lighty AI

RICHARD RABBAT: Hi. I'm Richard Rabbat. I want everybody to take a second. Close your eyes. No cheating. Close your eyes and think about your last week. And think about all the stuff that you decided to accomplish versus what you actually accomplished. Now open all your eyes. And think about what you could have accomplished if there was an AI companion that could help you. That could be your Chief of Staff, your admin, and that could help you achieve more for yourself.

I'm Richard Rabbat. I'm a PhD from MIT. I was most recently VP of Product at Twitter. Rebuilt the whole data and ML platform. When you look at AI, most of the systems that you look at-- Claude, ChatGPT, all these systems, they've been trained on a lot of data that's available publicly. So they can give you a lot of information. We've seen some amazing startups that are helping you retrieve information. But a lot of these systems are actually not aware of how you work. How your company organizes itself. What workflows actually work.

Meanwhile, everybody is under pressure to adopt AI because there's an immediate need for bringing more productivity into the workplace. So Lighty is an agentic AI. So basically, it's an AI that does things. And what it does is, we build an enterprise-specific LLM that can help you with all tasks, including onboarding, planning, coaching, scheduling, task assignments, tracking your OKRs. And we also do it in a personalized way so that the AI actually speaks the language that you want and converses with you in a way that makes you feel comfortable.

Here's an example of the AI basically helping you with scheduling, a task that actually ends up taking a lot of interruptions of your time. Each of these interruptions typically takes you about 20 minutes to-- it takes you about 20 minutes to recoup from that interruption. In this case, basically, Lighty just takes over the conversation, negotiates on your behalf, available time, and then sets it up on your calendar. And it's all basically powered by a knowledge graph that captures you, your relations with other people, and the work that you're doing.

And the next thing that happens is when you do this, when the other person gets the calendar entry done, they actually have a conversation. They've actually created the relationship with the AI. So it has a vital impact as well. Here's your daily planner. Typically you go into your day, and you're like, my goodness. I had like a bunch of meetings that I should have moved around. But it's too late. And here, by Lighty coming and, like, giving you the information ahead of time, you can say, I'm just going to cancel. You use natural language and Lighty will just do it.

And then at the end of the day, all of us are missing that coaching capability that we used to have. In some cases, you're assigned an executive coach. You find your own mentors. But Lighty here basically looks at what you're spending time on and helps you reorganize against the goals and objectives that you said in this case. I said that I want to spend more time on deep work and Lighty gives me recommendations by looking at what my calendar looks like and giving me insights into what I should do-- so basically, blocking more time to do deep work, canceling some meetings that are too tactical, et cetera.

And then finally, when you look at every meeting that you get into, when you leave that meeting, some people have taken notes. They remember some action items. Lighty is much more comprehensive. It makes sure that all the action items that were spoken about are actually captured. And then there is follow through on the on these action items. And there's typically a bias when somebody is taking notes because they just care about the things that matter to them. Lighty is more comprehensive. Takes down all the pieces and makes sure that they get assigned and people progress on these tasks.

We're also doing a production deployment with a platform partner. So it will be available to hundreds of thousands of companies in that case. We have a number of pilots that we've established, and we would like to work more with companies that want to really improve how their teams work and how they collaborate to be able to deliver more for themselves, for their teams and for their organization. Thank you. We're outside, and we'd love to connect with you. Mikey, our CEO, and myself will be there. Thank you.