We Rebuilt Everything In 2025. Now We're More Excited Than Ever.

If 2024 was building a plane while flying, 2025 was rebuilding that plane mid-flight. 

This year was one of the most transformative in recent tech history. AI capabilities that seemed futuristic in January became table stakes by the summer. And for those of us building in this space, the rate of change wasn't just something we observed, it was something we lived through every single day.

As we close out our third full year of Hoop, we wanted to share some behind-the-scenes stories from a year that challenged us, changed us, and ultimately led us somewhere we're incredibly excited about.

Starting Strong

We kicked off 2025 riding high. In December 2024, we'd just gotten Google's stamp of approval for our product in record time, and had big plans to scale in 2025. At that time, Hoop used AI to automatically detect and aggregate all your tasks across email, meetings and Slack (almost like a to-do list that wrote itself).

January brought milestones: we hit 1,000 trials in a single month. We launched "botless," meeting capture that eliminated the dreaded meeting bot joining your calls. When we shipped it to everyone in late February, the team celebrated with the energy you'd expect after months of grinding through audio fixes and transcript quality improvements.

The Honest Middle

By spring, we were shipping constantly. Multi-calendar support. Multi-Slack support. Shared lists. Auto-labels. Ask Hoop. Our velocity was incredible.

And yet.

Our PMF (Product-Market Fit) survey in April revealed that 31% of users would be "extremely disappointed" if Hoop disappeared, which was good, but not good enough according to Superhuman PMF rubric. More concerning: organic growth wasn't anywhere near where we needed it to be. We had built an acquisition machine, but not a sustainable engine. And inference costs for our product, though decreasing overall over time, were increasing per user because of Jevon’s law.

The more AI features we built into our product, the more inference each user required which made COGs more expensive. This is something most AI native startups will need to figure out long term (and why you see so many crazy funding rounds. It’s expensive to run AI products at scale!).

The Summer of Slackmageddon

In June, Slack changed their policies seemingly overnight for AI developers. Our product would no longer be able to use Slack data to create tasks for customers, a key differentiator. This, alongside unsustainable inference costs, was a wake up call that our current path needed to change.

While it was super painful to our loyal, incredible customers to shut down that product, it was unavoidable for the business.

We had to build something that customers loved, provided a clear business benefit in seconds of onboarding, and had unit economics that could scale over time. We spent time thinking through all that we learned building AI products, and decided on a clear path for the next product.

Hello, Hoop for Customer Support

From those ashes, a new Hoop emerged, AI-powered customer support for startups. While most legacy support platforms bolt AI on top of an ancient helpdesk codebase, Hoop is reimagining customer support with AI at its core. What does that actually mean?

We had learned how critical it is for user adoption to fit directly into users’ workflows. This new Hoop plugs directly into a support inbox (or helpdesk), automatically creating a knowledge base from existing context, learning the voice and tone of the business, and getting to work crafting responses to customer inquiries immediately.

It's like hiring an expert on your business to answer customer inquiries. Someone who doesn't need hours of training, works 24/7, and makes one person much more efficient by allowing them to get through inquiries faster than ever, without compromising customer service.

Within weeks, we added a chat component, using the same knowledge base to power on site support chat. Then we added a Stripe connection so Hoop could help answer billing inquiries. Then Zendesk and Helpscout integrations, so Hoop answers would appear directly in helpdesks. And we’re just getting started.

The pivot felt right almost immediately. Within weeks, we had our first design partners signed up. Businesses that were drowning in customer inquiries but didn't want to fully automate and lose the human touch. People who needed help but couldn't afford to hire a support team. Founders with a mandate to be efficient, but only legacy products with AI bolted on top as their options.

What We Learned

Reading through a year of weekly planning sessions (yes, we document everything), certain themes emerge:

The way we build products now is completely different from how we built them a year ago. The tools have changed. The capabilities have changed. What used to take weeks can now take days. The AI isn't just in our product—it's fundamentally changing how we create that product.

Initial assumptions can be completely wrong, and that's often exactly what leads to breakthrough insights. We built an excellent meeting task capture tool. But the world needed something else from us.

Having the right team with the right mindset makes all the difference.  When you're making gut-wrenching decisions over a holiday weekend, you need teammates who share your values even when you disagree on tactics.

Looking Ahead

We're ending 2025 building a different product than the one we started with. But we're ending it with more conviction, more clarity, and frankly, more excitement than we've had in a long time.

Hoop is now the easiest, most accurate customer support that fits directly into your workflow. We're already hearing from beta customers that it's changing how they run their businesses. That's the kind of feedback that makes everything, the pivots, the hard decisions, the starting over, worth it.

The rate of change in technology right now is unprecedented. For those of us building in AI, every week brings new capabilities and new possibilities. It's exhilarating and exhausting in equal measure.

Thank you to everyone who's been part of our journey this year: our users, our investors, our beta customers, and our community. The startup journey is hard, but it's always worth it with the right people beside you.

2026 is going to be a big one. We can't wait to show you what we've got cooking.