Announcing Our Investment in Sable

Explaining what your product and service do is a core part of every company. The marketing associate invests in crafting and updating the right assets. The sales rep leads calls to identify a pain point, then showcases the solution five times a day. The sales engineer runs the same demo forty times a quarter. The deployment lead walks every new customer through the same setup. This is strange work -- high-judgment enough that companies staff top employees against it, repetitive enough that they spend their days saying the same things to different rooms -- and it is the one part of the software and AI business that never got cheaper. Everything the industry built to escape it (i.e. documentation, video libraries, in-app tours, chatbots) is a recording of the expert rather than the expert: it broadcasts, but it cannot look at your screen, your instance, your half-finished configuration, and cannot respond. Live expertise gets rationed by deal size. The enterprise buyer gets a solutions engineer; everyone below the line gets a docs page or a Loom recording. Context leaks across the handoffs in between -- SDR to AE to sales engineer to implementation to success -- and the buyer re-explains themselves at every stage. The gap between what the product can do and what the customer understands is the quiet ceiling on every business. Until now, the only way to raise it was headcount.

That ceiling was a reflection of the tools, not the work. And sable changes the available toolkit. AI has eyes and hands now: real-time vision to read the screen, computer use to click and type through it. So it can share your browser, move through different product interfaces, and respond to what you do as you do it. Full-duplex speech lets the agent talk and listen simultaneously, so it can be interrupted, absorb a change on the screen, and adjust mid-sentence. Not a recording of the expert -- the expert, at the marginal cost of the product it explains. What Aidan does every day was science fiction a year ago. Now, companies can put it in front of every customer.

Interactive Intelligence

The orchestration of real-time computer use, voice, vision, and video allows Sable to pursue a bold vision: building the first AI employee that can see, click, and explain a product in real time. A prospect or customer clicks a link, joins a live call, and Aidan - Sable's AI employee - is already waiting for them, ready to walk them through the product, answer their questions, maneuver in a shared browser, and drive the experience while customers watch and take over whenever they want.

The important part is that Aidan isn't narrating screenshots or replaying a scripted demo. It operates a live, fully functioning instance of the real product inside a LiveBox, a virtual machine built for shared, real-time interaction. Aidan can click into any corner of the product, react to what the user does mid-sentence, and pass control back and forth like two people at one keyboard. Sable calls this interactive intelligence.

Underneath sits the brain, a multimodal context graph that ingests a company's docs, recordings, and best demos into a single source of truth and grows sharper with every call. Once the brain exists, launching new use cases ( i.e. pre-sales, onboarding, new feature launch, partner activation) requires almost no work, and every interaction makes the whole system more capable.

Traction

Sable is less than a year old but already in production-with the world's fastest-growing AI companies like Notion and Decagon, as well as large public companies. The demand to hire Aidan is astonishing: Sable has an inbound waitlist of over 150 customers waiting to implement this new capability, before it has even launched publicly.

At Notion, what started as AI-led setup sessions for new users has quickly grown into custom agent activation, product demos, and global partner enablement. White-glove demos and onboarding, once reserved for the largest accounts, now reach customers and partners worldwide, in dozens of languages, with pricing tied to outcomes. These are the high-volume, high-judgment moments that many companies are looking to scale.

When the cost of expert, one-to-one interaction falls, the commercial functions built around its scarcity change shape. Companies staffed demos, onboarding, and enablement only for the accounts that justified a human; everyone else got a help center and a form. An AI employee that can run an unlimited number of live, expert sessions collapses that tradeoff. The long tail of prospects, customers, partners, and markets that were never worth a person's time can get the white-glove treatment too, in any language, at any hour.

8VC watched the same pattern work once before, at Outset, where moving qualitative research from 1:1 interviews to AI-moderated conversations at scale opened an entirely new operating model. Sable does this for the customer-facing motion, turning 1:1 human-led work into 1:many work led by an AI employee.

From Text to Interaction

For two years, interacting with AI mostly meant typing into a text box. That was a real breakthrough, but text is the narrowest channel as it pushes anything visual into the background and confines the exchange to a stilted, turn-by-turn rhythm. Selling, showcasing, and guiding users on how to use products were never problems that could be solved by text. They happen by showing and reacting in real time, and an AI that can see, click, and speak at once is finally built for that.

Over time, Aidan can become the primary interface between a company, its customers, and eventually their agents. Sable is turning the hard part - a living context graph and real-time interaction layer - into infrastructure any company can buy.

The Team

It takes unique technical firepower to push the frontier of multimodality - the kind Nim Ravid, Linda He, Leon Chen, and Itamar Rocha Filho and their team brought when they left Harvard research labs to start Sable. That depth shows up as speed at the frontier. The first time we met the team, they were already three steps ahead of the field, pushing advancements on real-time vision capabilities. As real-time computer use, vision, and full-duplex speech each became possible, Sable was first to make them reliable and put them in front of customers. That instinct for what's newly possible at the frontier, paired with speed of execution, is the core of our conviction.

Nim is a magnetic founder who earns the trust of sophisticated buyers by doing the work to understand their businesses from the outside in, with an exceptional ability and track record for bringing together the highest-caliber talent. Linda brings deep expertise in AI research and post-training from her time at Together AI and Tandem Health. Leon brings world-class engineering and deployment experience from SpaceX, DoorDash, and Google. Itamar combines applied AI research with exceptional engineering experience, having worked at Google and Meta and focused on Reinforcement Learning research at Harvard's Kempner Institute over the last two years. They've assembled a world-class team, which includes Math 55 alums, former quantitative traders, and International Math Olympiad winners. Together, they have already solved a set of hard technical problems the rest of the market has yet to crack: proactive vision, real-time browser use, and a shared interaction environment.

As models improve, new possibilities open faster than any roadmap can absorb. We are betting on the team most likely to seize those openings first.

Sable x 8VC

Today, Sable announced a $45M fundraise led by 8VC and Sequoia, and that Joe Lonsdale is joining its board. We are proud to partner with Nim, Linda, Leon, Itamar and the team as they launch publicly, grow their engineering and research bench, and bring their AI employee to the companies defining the next era.

To learn more, visit withsable.com - they are hiring!