Five events in a single day.

We may have just pulled off our greatest NY Tech Week Unicorner experience yet. We hosted three incredible firesides (full recordings linked):

Immediately following, we brought together 30 professionals to a beautiful terrace for an AI Champions dinner, and capped off the evening with an afterparty, featuring a live performance by the one and only Ian Asher.

Missed us in NYC? Don’t worry. We’re about to pull off our most ambitious event yet in SF on Wednesday, June 24.

If you’re coming to Figma Config, join us for our experiential festival turned afterparty: Welcome to the Stratosphere.

Food, drinks, and carnival games on a rooftop festival. Immediately followed by a live performance by Cheat Codes.

In the meantime, enjoy today’s article in the Highlight… on Highlight.

Ethan and Arek 🦄

Highlight sees everything on your screen, understands your workflows, and where and how you work, and uses that information to coordinate how you and your team collaborate by proactively predicting next steps. The core feature is suggesting actions, such as drafting an email on Outlook, updating a document on Notion, or configuring Claude Code to run unit tests. In addition, it will provide contextual reminders, like “In your 1:1 last week, there was a follow-up you had stated would be ready by Friday. Is this done?” Highlight is platform and tool-agnostic, meaning it observes how you work and collaborate no matter which tools you use. It is expanding to bigger use cases as a cross-team solution, expected to be ready this summer.

Check it out: highlight.ai

Highlight offers Free, Pro, and Enterprise tier plans, all of which have a base subscription price followed by memory-, storage-, and usage-based pricing. Each tier gets a certain amount of credits and storage, and users can pay for more should they run out. The Pro version offers premium and flagship AI models, premium support, instant action items, and more, while the Enterprise plan includes everything in Pro plus additional team-specific customization, custom SLA and security compliance, and more.

Highlight was spun out of a startup called Medal.tv in 2024. Medal is focused on video game streaming and capturing moments of users playing with friends. There, an AI assistant called Highlight was born as a research project, exploring how LLMs could interact with and interpret what was happening on the player's screen. The interest grew as the use case seemed to expand beyond just video games. For the project to scale and bring on engineers, the team realized it was best for it to be a separate entity. 

It’s the two co-founders of Highlight, Sergei Sorokin and Joshua Lipson, who have helped it go from a research feature within Medal.tv to a product of its own. While Lipson served as CTO and later Chief Architect at Medal.tv, Sorokin was the VP of Product at Discord for 8 years. After leaving Discord, he knew he wanted to work in AI, but specifically wanted to help AI make humans better at what they do.

The duo was fascinated with a problem they called the “coordination tax” of manual effort required to coordinate everything from different software tools to teammate to teammate communication. That shared obsession with reducing coordination tax is what shaped Highlight's focus as an intelligence layer built around how people work across every tool and conversation on their desktop.

Highlight’s goal is to eliminate what it calls the coordination tax. Despite the number of modern, innovative AI tools today, there is still a mundane process of moving information that needs to be coordinated. While existing tools are good at doing a single task in a single setting, such as summarizing a meeting, fewer are good at coordination. This could be, for example, the sequence of first recapping a meeting, then sending the notes to the right stakeholders, and finally creating follow-up action items in a shared workspace.

Highlight is addressing this by creating a shared intelligence layer for teams. First, it observes how users operate and collaborate on their desktop and starts suggesting “drafts” of actions with certain tools the user uses. For example, it might suggest follow-up emails and updating a document. It is then up to the user to decide whether they want to accept that draft, with flexibility allowing for editing and small modifications, but the basic groundwork is generated proactively by Highlight.

One particular advantage Highlight has is that it’s model- and tool-agnostic. Many tools exist as a wrapper around a model like ChatGPT or Claude or focus on automation around one specific software, but Highlight is flexible and compatible with the setup the user already has. This adds flexibility, as it’s not necessary to adjust behavior or adopt a specific software, making the barrier to onboarding Highlight significantly lower.

And, there are macro trends that heavily favor Highlight. One in five workers now operates remotely, making seamless team coordination harder than ever. Meanwhile, AI adoption is accelerating fast, with 90% of global companies already using or actively exploring AI, and 92% plan to increase their AI investments over the next three years. In this scenario, the coordination tax is only getting steeper.

After speaking with Sorokin, he acknowledged two challenges that Highlight has to navigate. First, is how to balance information at a team level while keeping privacy at the forefront. Highlight’s team is actively keeping this top of mind as it builds its team-focused solution. The approach already baked into the individual product gives a sense of its privacy policy, with neither Highlight nor its LLM partners training on user data, and fully encrypting all messages in transit and at rest. 

Second is that Highlight often requires context about how someone works, so suggested actions don’t come up until enough context has been provided. As a result, the experience doesn't deliver its full value right away, and the impact or “wow” moment for users isn’t immediate. But this is something that can and should be improved as they continue to acquire more users and enhance its decisioning layer. 

Aside from recognizing the strong experience and credentials displayed by Highlight’s founders, we also think the team is forward-thinking and adaptive. What started as a spin-off of a video game streaming startup has grown into a personal assistant for individuals managing their own workflows and is now being extended to how teams think, coordinate, and act together. Highlight is expected to release its team-focused desktop AI solution this summer, building off of its individual offering.

The early signs of this team-focused solution are positive, as there’s a growing waitlist, and the most successful users accepting about 2/3 to 3/4 of drafts actions. If that release goes well and continues its early momentum, Highlight will be the tool that eliminates the coordination tax among teams and individuals.