Re-Founding Rhumbix: Our Journey to Becoming an AI-Powered Business

This past May, I attended the SaaStr Annual conference and I left with a message that fundamentally changed how I think about building Rhumbix. Jason Lemkin told the audience that while AI tools weren’t ready for prime time at last year’s conference, they are now. Speaker after speaker from major technology companies reinforced this message, sharing inspiring stories of how AI was transforming their businesses. But what really stuck with me was Jason’s challenge to every founder in the room: you need to re-found your business, regardless of where you are in your journey, or you face an existential threat of quickly becoming irrelevant.

That resonated deeply. As Co-Founder and CEO of Rhumbix, a workforce management platform for the construction industry, I realized we had an opportunity to fundamentally transform how we operate. Not just to build products faster or cut costs, but to reimagine what our company could become.

Six months into this transformation, I want to share what we’ve learned, what’s working, and why I believe every software company needs to be on this journey right now.

Starting with Real Impact

We didn’t begin with a grand AI strategy. We started by identifying workflows where AI could create immediate, measurable value. The results have exceeded our expectations.

Customer Intelligence at Scale: We now have AI note takers join every call with prospective and current customers. The transcripts feed multiple workflows that have transformed how we operate. First, we automatically generate meeting summaries with next steps and action items, saving each employee 2-3 hours per week. More importantly, we extract excerpts from these calls to build a voice of customer database that has become invaluable for refining our value proposition and generating marketing materials. We also route product feedback directly to our engineering and product teams, creating a continuous loop of customer insight.

Engineering Velocity: Our engineering team is now using generative AI tools to write over 40% of the code in recent development sprints. We’re also leveraging AI for code reviews, and it’s been remarkably effective at catching bugs before they reach production. The downstream impact is enormous. Our engineering team spends more time designing and building new features rather than firefighting production issues. Our customer success and support teams save hours per week they would have otherwise spent triaging bugs.

Customer Success and Support Transformation: We’re using AI to create documentation for new products and features, as well as weekly product release updates. We’re building a sub-agent for bug triage and automated fixes that’s already showing a 50% success rate at both identifying and resolving issues with zero human interaction. We’re also working toward automating resolution of customer support tickets, with a goal of handling more than 50% of Tier 1 tickets without human involvement.

The aggregate impact across the company has been significant. We’re saving 50-100 hours per week, equivalent to 2.5 full-time employees. And honestly, it feels like we’re just getting started.

The Challenges Were Real

I’d be lying if I said this transformation was easy. We’ve encountered large challenges along the way.

The Learning Curve: AI requires every employee to climb a learning curve, and finding time to learn something new while maintaining your daily responsibilities is genuinely difficult. The technology is powerful but not always intuitive, especially for team members without technical backgrounds.

Overcoming Skepticism: AI hallucinations are real, and they create healthy skepticism that must be addressed head-on. Some team members initially questioned whether the tools were reliable enough to trust with important work. We had to build confidence through demonstrated results and transparent conversations about limitations.

Constant Evolution: AI technologies are changing at a breathtaking pace. We’ve had to constantly iterate and re-evaluate whether we have the right tool for a given workflow. More than once, we’ve had to rip out and replace tools with better alternatives. That can be frustrating, but we’re building a muscle for adaptation that will be critical as the pace of technological change accelerates.

What I’ve Learned as a CEO

The biggest lesson from this journey is that CEOs need to roll up their sleeves and use these tools themselves. You can’t delegate AI transformation to a committee or a single department. You need to understand firsthand how these tools work, what they’re capable of, and what their limitations are.

During our recent, AI-focused Hackathon, I teamed up with one of the youngest members of our team to automate the generation of customer case studies from recorded conversations. Over 24 hours, we went through an intensely iterative process of trying something, failing, tweaking it, and trying again. By the end, we had auto-generated four to five case studies that were ready for some final design polish before being released.

Another breakthrough insight, shared by multiple teams working on completely different projects during the Hackathon, was the need to break larger workflows into a series of smaller workflows, each accomplished by a separate AI agent. For our first attempt on the case study agent, I literally uploaded 700 pages of transcript and said, “Go build a case study.” The results were mediocre. When we decomposed the problem into discrete steps, the quality improved dramatically.

We’ve been doing Hackathons for almost as long as Rhumbix has been a company, but the most-recent, AI-focused Hackathon was far and away our best yet. Watching non-technical team members build and ship functional AI workflows was inspiring and helped break through much of the skepticism that existed.

Another critical lesson: invest in training for your employees and yourself. Our initial efforts relied heavily on self-experimentation, which moved the ball forward but only incrementally. It wasn’t until we brought in an outside resource to advise the company that we started unlocking larger gains and thinking bigger about what AI could do for Rhumbix.

What’s Next

We’re still in the first or second inning of leveraging AI. My goal for our internal transformation is to scale Rhumbix to two to three times our current size without adding more than a few additional headcount. That’s an ambitious target, but the trajectory we’re on makes it feel achievable.

We’re also beginning to design and build AI functionality directly into Rhumbix products and services for our customers in the construction industry. The same transformation we’re experiencing internally can create tremendous value for the contractors and builders who rely on our platform.

A Challenge for Other Founders

If you’re a software executive reading this, my challenge to you is the same one Jason Lemkin issued to me: re-found your business. The AI tools are ready. The question is whether you’re ready to embrace them.

This isn’t about chasing hype or checking a box. It’s about fundamentally rethinking how your company operates and what becomes possible when you augment every department and team with AI capabilities. The companies that move decisively on this transformation will have significant advantages in velocity, efficiency, and innovation.

The journey isn’t always smooth, but it’s absolutely worth it. I’m excited to see where this takes Rhumbix over the next few years.

If you want to follow our progress, subscribe to our blog and visit us at rhumbix.com. I’d also love to hear from other founders and executives on similar journeys. What’s working for you? What challenges are you facing? Let’s learn from each other.

---

Zach Scheel is Co-Founder and CEO of Rhumbix, a workforce management platform serving the construction industry.