Moving to SF

We recently moved to San Francisco! We are thrilled to be in the Bay Area ahead of the 2025 Y Combinator Spring Batch commencing in April. If you know anyone we should get in touch with, please let us know.

As part of our move, Fernando ordered a bed frame, mattress, and dresser from Ikea. We’d planned it out perfectly so that his car would arrive carrying his tool so we could put together the bed frame and get his room all set up. Unfortunately, as is always the case, our perfect plan didn’t quite go as expected. We got a text from the guy hauling the car that the police had found a problem with the trailer, which meant that it would be arriving later in the week instead of Tuesday. Hearing this, we scrambled to the local hardware store to rent a drill before they closed at 6p only to find out that their one drill was already rented. What that left us with was “founder bed”, aka a bare mattress in the middle of our living room (image included below).

Why we’re so excited for agentic workers

There’s a shift happening in how work gets done—and it’s happening fast. Over the past year, we’ve seen a wave of new tools that don’t just help people work faster—they actually do the work for them. These are what we call agentic workers: autonomous AI systems that can take action, make decisions, and complete tasks on behalf of humans.

If you’ve seen companies like 11x, you’ve probably felt this future creeping closer. They’re offering fully AI-powered roles like sales reps or recruiters—and people are actually using them. This isn’t just hype. It’s a glimpse at a new kind of workforce: digital, on-demand, infinitely scalable, and designed to work alongside humans—not replace them.

And we think that’s a good thing.

Imagine a world where companies—especially startups—can scale without hiring armies of people. Where you can launch a new campaign, analyze a dataset, or recruit top talent without adding headcount or diluting equity. That’s what agentic workers make possible. They unlock productivity and remove the bottlenecks that slow teams down.

Wouldn’t it be incredible if founders and operators could skip the usual pitfalls of hiring too fast—or too late? What if they could validate product-market fit faster, run more experiments, gather insights, and make better decisions without burning runway? We see a future where businesses move with way more agility because intelligent, reliable agents are handling the busywork and freeing humans up to focus on what really matters.

To us, this feels like one of the most exciting shifts in work since cloud computing. And we couldn’t be more pumped to build the infrastructure that helps make it real.

The Problem

Right now, it’s way harder than it should be for companies to actually leverage agentic workers.

The ecosystem is still early—and completely fragmented. New agents are launching every week, but there’s no easy way to figure out what’s actually useful, what’s trustworthy, or how to plug them into your business without weeks of duct tape and custom code.

Even teams who are eager to experiment often end up hitting a wall. The process looks something like this:

• Spend hours digging through Twitter threads, Discord servers, and obscure marketplaces trying to find promising agents

• Struggle to integrate them into existing workflows or systems

• Cross your fingers and hope they’re actually working—because visibility is basically non-existent

• Try to justify the effort, but have no clear way to measure ROI or performance

• Worry about security, permissions, and whether these agents are accessing the right data—or way too much of it

For something that promises to save time, adopting agentic workers can still be a massive time sink, today.

And perhaps the biggest missing piece of all? Control.

Once you’ve got agents running, there’s no good way to manage them day-to-day. No central place to see what they’re doing, guide their behavior, or pull a human into the loop when needed. It’s like hiring a new employee with no email, no manager, and no performance reviews—and hoping for the best.

We’ve talked to dozens of teams—startups, scaleups, even solo founders—and they’re all saying the same thing:

That’s the gap. And that’s what we’re solving.

Our Solution: Controlling the Chaos

We’re building Beluga Labs, the orchestration layer for AI agents.

Think of us like Kubernetes for agentic workers—simplifying a fragmented, chaotic ecosystem into something powerful, reliable, and scalable. Instead of wrestling with dozens of disconnected tools, Beluga Labs gives businesses a single, streamlined platform to make agents actually work for them.

With Beluga Labs, companies can:

• Easily find vetted, ready-to-use agents
  Skip the hours of research. We surface verified agents with real reviews, so you can deploy with confidence.

• Integrate agents effortlessly with one-click onboarding
  No complex setup or dev cycles. Add agents to your workflows instantly—no glue code required.

• Track performance and ROI
  Get visibility into what your agents are doing and how well they're performing, through centralized analytics and logs.

• Maintain control and oversight
  Set granular permissions, audit decisions, trace behavior, and create human-in-the-loop workflows where needed.

• Control everything from Slack
  Whether it's adding an agent to a channel, requesting updates, or approving actions, you can manage your entire agent layer through a single Slack app. It’s the easiest way to stay in control without context switching.

Our goal is to make agentic workers not just accessible—but operational. To give businesses the confidence to experiment, scale, and move faster than ever before.

This is just the beginning. But we’re starting with the infrastructure that makes the future of work actually usable today.