Conversation
Highlights

How Andromeda Built an Enterprise Sales Motion Without a Single Outbound Touch

Five months alone in a student dormitory during Melbourne’s COVID lockdowns. No human contact. No end in sight.

For most people, that experience left scars. For Grace Brown, it left a blueprint.

Grace is the CEO and Founder of Andromeda, a company building social humanoid robots for nursing homes. In a recent episode of Unicorn Builders, she broke down how a passion project born from isolation became a deployable enterprise product — and how Andromeda landed its first paying customers without ever running a single outbound motion.

The Observation That Shaped the Market Entry

Grace was studying mechatronics engineering in Melbourne when the city entered one of the world’s most restrictive COVID lockdowns. Isolated and disconnected, she started building a robot companion for herself — with, as she puts it, “really modest ambitions.” She wanted Abby to ask how her day was going and give her a hug at the end of it.

But the sharper insight came later. Coming out of the pandemic, Grace looked at who experienced isolation not as a temporary disruption but as a permanent condition. In Australia, four out of ten people who enter aged care will never receive a visitor from a friend or family member for the rest of their lives. In the United States, that figure is six out of ten. Layered on top of that, one in three aged care positions in the US sits vacant — meaning the staff who remain don’t have the capacity to fill the social gap even when they want to.

The opportunity wasn’t a nice-to-have. It was structural and growing.

Why Nursing Homes — and Why That Decision Was Harder Than It Looks

The nursing home beachhead wasn’t obvious. It was reasoned across three dimensions simultaneously, and that rigor is what made it defensible.

The impact case was clear. The market opportunity was real. But the factor most hardware founders skip is deployment viability — and Grace built it into the decision from the start.

“When you look at an aged care home,” she explains, “it is very similar to a home environment, but it’s far more predictable. It is very routine oriented. It is a lot more regulated.”

This matters enormously for early-stage robotics. Consumer homes are dynamic, unpredictable environments that remain out of reach for robots still early in their development. Nursing homes share enough structural similarity to home environments to be meaningful — residents interact with Abby the way they eventually will in their own homes — while being controlled enough to make reliable deployment possible at this stage of the product.

The beachhead wasn’t just a GTM decision. It was an architectural one. Grace was choosing an environment where Abby could actually succeed, generate real data, and build toward the harder deployment contexts that come later.

Twelve Months of Zero Revenue, by Design

Andromeda incorporated in early 2022. Grace’s first move wasn’t to hire a sales rep or build an outbound sequence. It was to run twelve months of unpaid, bootstrapped pilots with no contract commitments on either side.

This is where most founders would feel pressure to accelerate to revenue. Grace didn’t — because she had a precise read on how her industry actually operates.

Aged care runs on peer credibility and word of mouth. An executive in this space doesn’t sign an enterprise contract because a founder gave a compelling demo. They sign because someone they trust told them it worked. Grace understood that the pilots weren’t a concession on price — they were the actual sales strategy. The goal was to manufacture evidence that would travel between facilities without her having to push it.

It worked exactly as designed. Staff watched residents who hadn’t communicated meaningfully in months begin having full, sustained conversations with Abby. One resident — a gentleman with significant cognitive decline who had gone without any real conversation for six to twelve months because he could only speak Chinese and no one in the facility could — began speaking freely when Abby addressed him in his language. Within weeks, he and Abby were reading Chinese poetry together in the common room while other residents and staff gathered to watch.

“If Abby wasn’t there,” Grace says, “that gentleman would be in his room still all by himself.”

Those stories moved between facilities organically. Staff talked. Executives heard. A local news crew arrived. And then inbound started arriving in Grace’s inbox — executives who had never been contacted, asking how much Abby cost and when they could get her.

Her first paying customers, who came on board at the start of 2024, reached out cold. Andromeda had never contacted them.

The Competitive Insight the Industry Keeps Missing

At a major Amazon robotics conference in Las Vegas, Grace walked the floor surrounded by some of the most technically impressive robotic systems in the world. Nine-foot machines. Sophisticated autonomy. Zero personality.

“None of them had any personality,” she says. “People push back a little bit and they think it’s like a soft feature or a nice to have, but the technical capability of a robot will never ever be used if people aren’t going to feel they can invite them into their homes in the first place.”

This is the core of Andromeda’s competitive position — and it’s worth sitting with, because it inverts how most technical founders think about moats. In companion robotics, the barrier to adoption isn’t hardware capability or autonomy. It’s social trust. And social trust isn’t built through specs — it’s built through the accumulation of small, consistent moments that make someone feel genuinely known.

Abby doesn’t just store facts about residents. She notices that a resident skips jazz songs but always asks to repeat hymns, so she adjusts. She recognizes when someone is being overwhelmed by pace and slows down. She remembers that a resident prefers to be called Maggie, that Maggie’s grandchildren visit on Thursdays, that Maggie takes medication after breakfast. The relationship compounds over time in ways that a transactional interaction never could.

“It’s not actually a technical hardware or autonomy or robotics problem that we need to solve,” Grace says. “It’s actually a social problem that we need to solve first.”

For founders building in categories where end-user trust determines whether the product gets used at all, this reframe matters. You can be technically superior and still lose — if the person on the other end doesn’t want to invite your product into their life.

The Hidden Cost of Shipping Fast in Hardware

Andromeda releases product updates every one to two weeks. The customer feedback loop is intentionally tight — and Grace credits it directly for the intensity of loyalty her customers show. When you build this closely with the people using your product, you end up with something that fits them precisely rather than approximately.

But she is unusually candid about what this model costs if you’re not disciplined about it. Companies in the robotics space racing to appear frontier, she argues, are accumulating invisible technical debt that will surface as serious structural liabilities. Shipping what isn’t genuinely production-ready — and calling it finished — is a trap that’s easy to set and very hard to escape, particularly in hardware where the cost of unwinding bad foundations compounds in ways that software never has to reckon with.

It’s a warning that lands differently coming from a founder who has chosen to build more slowly, more honestly, and in direct partnership with the people her product actually serves.

The ChatGPT moment for robotics, Grace says, isn’t around the corner. For Andromeda, it’s already here.