When we started Anon in 2023, we were pioneers in an emerging frontier. Browser use agents were barely a concept, and the technology was so nascent that most people were only just starting to imagine AI actually navigating websites like humans do. Yet we saw the wave coming and started building authentication infrastructure for what we knew would become a critical need.

By 2024, our timing proved prescient. There was an explosion of agent building companies tackling computer use, particularly in the consumer sector. Companies like Adept, Rabbit, MultiOn, Humane, and Minion raised massive rounds promising to revolutionize how people interact with the web. The market was validating our thesis, and we had a successful launch that attracted nearly 70 customers eager to build the future.

But after a year of watching how our technology was actually being used, we discovered something that would reshape our entire business and foretell the struggles many of these companies would face.

What We Learned from Our First Year

Our startup customers were using our authentication layer to build agents that scraped social media profiles and booked flights online. These were impressive technical demonstrations, and they generated consistent monthly growth. But as we dug into the data, two critical realizations emerged:

First, there was a fundamental difference between automating a task and automating a job. Our startup customers were focused on low-value tasks with minimal marginal value—one-off activities that saved minutes. Meanwhile, our enterprise clients were doing something completely different: they were automating entire jobs. Jobs are repeatable workflows that save millions, not minutes. This distinction would prove fatal for many consumer agent companies who were building incredibly complex browser automation technology to solve problems that weren't worth the engineering investment required to make them reliable, while million-dollar problems were hiding in plain sight in the enterprise.

Second, because we were only providing one piece of the whole browser automation stack (the automation) we could not control the outcomes. Our authentication technology enabled our customers to “open the door” to third-party web applications, but we couldn’t control what they did on the other side. Their failure was our failure.

The Hard Decision

These two realizations pushed us to make one of the hardest decisions a startup can make. Despite having nearly 70 paying customers generating consistent monthly growth, we knew that being distracted by startup experimentation vs focusing on enterprise automation would dilute our impact.

So we made the call to churn these customers and focus on:

1. Enterprise-level jobs with high marginal value
2. The whole browser automation stack, including authentication, navigation, extraction, and transformation

The Evolution (Not the Pivot)

Here's what's crucial to understand, however: we didn't throw away what we built. The core authentication infrastructure—the technology we spent years perfecting—remains our foundation. Multi-factor authentication, CAPTCHAs, session management, secure and persistent logins. All the complexity that breaks traditional automation is still exactly where we excel.

What changed is how we package and deliver that technology. As we engaged deeper with enterprise clients, we saw three common paths companies were taking to try to automate high-value workflows, and they all had fundamental flaws:

• They could build human operations teams → expensive and impossible to scale.
• They could build their own browser automations → brittle solutions requiring constant engineering maintenance.
• Or they could use AI "computer use" software → great for demos, unreliable in production.

By expanding our solution to handle the entire workflow—authentication, navigation, extraction, transformation, and data delivery—we created one platform that actually works, replacing all three broken options. We evolved from providing a tool to delivering complete business outcomes.

This strategic focus has already proven itself: our contract values have grown over 80-fold as we've moved from selling components to solving complete workflows. When you're automating a million-dollar manual process, the value is undeniable.

Where We Are Now and What's Next

This evolution has transformed how we think about product development. We're no longer building tools and hoping customers find valuable use cases. We're identifying high-value, repetitive workflows and eliminating them entirely. We've moved from automating tasks to automating entire jobs.

By Q3 2025, we'll release Anon 2.0, letting your team create new automations without writing code. You'll have complete autonomy to build and deploy agents for any authenticated workflow. But you don't have to wait—we're already helping enterprises automate their most critical processes today.

The lessons of the past year have shown us something profound: the real opportunity isn't in making flashy demos or automating simple tasks. It's in tackling the complex, messy, authenticated workflows that companies have resigned themselves to throwing human labor at forever. These are the processes that compound in value when solved; the ones that transform entire industries when automated.

That's the world we're building at Anon. One authenticated workflow at a time.