At CES 2024, we are showcasing how the world's leading brands are building with Amazon's devices and services, and leveraging the latest innovations in artificial intelligence to build the next generation of ambient experiences. This article is part of a series that showcases select brands on display at the Amazon Experience Area at CES 2024.
Until recently, having a personal assistant to help you with everyday tasks has been a privilege available mainly to the rich and powerful. But what if we could all have a personal assistant to help us as we go around our day-to-day, assisting with the challenges that life throws at us, while also helping us accomplish seemingly simple tasks?
As many of us know, the simplest of tasks can quickly get cumbersome on the Internet. Imagine that you want to reserve a table at a restaurant and email five friends the invite. It’s not long before your simple task starts getting long and complicated. First, you need to navigate to the restaurant’s website and make the reservation—filling in your details as you go. Next, you have to copy the details into an email or a calendar invite. Even if the website has an “Add to Calendar” feature, you have to look up the email addresses of your friends and send them the invite.
Here's where the new personal artificial intelligence (AI) agent MultiOn comes in. The agent can use the web to accomplish these tasks with the same degree of autonomy and ability as a person. All you have to do is tell the MultiOn agent what you want to do, and it can effectively break your task into smaller micro-tasks, doing each one for you.
MultiOn is an Alexa Fund portfolio company. The Amazon Alexa Fund was originally launched in 2015 to create momentum for voice and smart devices and show the art of what’s possible. Since its launch, the Alexa Fund has provided venture capital funding to fuel not only voice innovation but also areas such as artificial intelligence, hardware, and entertainment. The MultiOn technology will be on display at the One Amazon space during CES 2024.
Building the MultiOn AI assistant to take action
The MultiOn concept originated in March 2022. Omar Shaya, Co-founder of MultiOn, was studying business management and leadership at Stanford University. His experience in furthering the state of the art in artificial intelligence and focus on personalization algorithms influenced him to create a new product that could take care of mundane tasks while he focused on what he enjoyed.
“The vision is to democratize access to help,” says Shaya. “MultiOn aims to streamline digital life for everyone through an AI model that understands your needs and will get the job done for you.”
Shaya began working on the project in July 2022. Shortly after, Shaya met his co-founder, Div Garg, who was pursuing a Ph.D. in AI and reinforcement learning at Stanford and was thinking about the same problem. After the two teamed up, the project progressed quickly. In January 2023, MultiOn had its first limited launch as a browser-based extension. By November 2023, MultiOn had more than 30,000 people waitlisted for the next launch.
“We are building MultiOn quickly, but we are taking a careful approach in our public release,” says Shaya. “Our AI is taking actions and not just generating text, so we’re being careful around the safety and privacy aspects. We are making sure that the system is under control and that it’s not going to do something users don’t want it to do.”
The final product, a mobile application, is slated for release in early 2024.
AI for horizontal and vertical use cases
Shaya and Garg, along with a seven-person team of developers, designed MultiOn to address a specific gap in the market. People use artificial intelligence for basic tasks like turning on lights and listening to music, in addition to leveraging the technology to perform complex business tasks and predictions for highly specific tasks in select domains. However, what was missing was accessible AI to help with time-consuming, mundane tasks.
“AI often has a specific vertical use case,” says Shaya. “We’ve built an assistant to help with more ‘horizontal’ use cases.”
MultiOn is considering two paths to monetization on the consumer side: an ad-free subscription model and a free version monetized through affiliate access promotions. On the developer side, there are even more opportunities. Already, the MultiOn team has hosted several hackathons, during which developers used the MultiOn technology through an API to build for vertical use cases. For example, a developer could build a price comparison tool for a travel agent.
“Through this API, we’re saving developers potentially months of work so that they can focus on building what they’re really passionate about,” says Shaya. MultiOn plans to monetize the technology through a consumption-based model so that developers pay as they go for using the API.
Embedding MultiOn across the web
The vision for MultiOn is to integrate the technology everywhere.
“We made our technology very embeddable,” says Garg. “You can connect it with virtually any device or interface on the internet, like smartphones and smartwatches. This AI assistant will be available everywhere.”
For example, an airline company could embed the MultiOn API on its website to book the best, least expensive flights for users autonomously. The AI can accomplish online tasks many times faster than a person can, saving time and effort on both sides of a transaction.
As MultiOn grows and advances, the biggest challenge will be showcasing all that the agent can do and inspiring people to go beyond their expectations. “Our philosophy is to unite great technology with great design, with the user at the center,” says Shaya. “Whether they use it on the phone or some other device, we would like to give people back the time that they would otherwise spend doing mundane tasks so that they can focus on doing the things they enjoy and live a better life.”