The Need
From the moment they’re founded, tech startups must hit the ground running. During the early stages of a company’s life, being first to market – often with a Minimum Viable Product (MVP) – is almost the only thing that matters. This is doubly true when most of your customers are fast-moving tech companies too. If you can’t demonstrate that you’re agile enough to keep pace with their needs, you’ll quickly fall by the wayside.
For Coho AI, a company that builds a revolutionary product-led revenue platform for SaaS providers, speed is of the essence. Armed with a clear vision of the challenges SaaS providers face in converting free-tier users into paying customers, Coho AI founders were determined to get their solution into customers’ hands as quickly as possible.
Michael Ehrlich, Co-founder and CTO, comments: “Like any startup, we had limited resources in terms of both time and developers. We had to be 100% focused on building our product for delivery to customers, and we needed a tech stack capable of supporting that goal.”
The Challenge
Coho AI knew that as soon as its product took off, it would need to scale to dozens of customers and potentially thousands of users within months. A microservices architecture would make it easier to scale, but it would also be complex to set up and maintain—shifting the team’s focus away from developing new functionality and creating a significant DevOps burden.
“We had to find a way to be fast and efficient on the engineering side,” says Michael Ehrlich. “We couldn’t afford to waste time writing boilerplate code to configure infrastructure—we wanted to be totally focused on business logic.”
Ideally, the company wanted to find a platform that would abstract away all the complexity of servers, load balancers, databases, security, and integration. The platform would simply provide a reliable, high-performance environment for developing, composing, and running microservices—enabling the team to launch the product within a four-to-six-month runway.
“With Kalix we got to market 75% faster compared to other solutions we had considered.”
The Solution
Coho AI explored various options before encountering Kalix, Lightbend’s platform-as-a-service offering for building and deploying cloud-native applications and APIs.
“Speed and efficiency were our top priorities, and that’s where Kalix came into play,” says Michael Ehrlich. “We quickly realized that the ease of use of Kalix would enable us to develop much faster than we had initially believed possible.”
The Coho AI team started by using Kalix’s free tier to experiment with the platform on AWS. They quickly confirmed that Kalix would help them solve engineering problems much faster and more easily, without spending time on boilerplate code and operations.
“Once we got to know the technology, we signed up and just went from service to service, developing them and making them work together,” says Michael Ehrlich. “It really was that straightforward.”
He adds: “Lightbend’s documentation is excellent, and the Lightbend folks were very, very helpful. We even had a shared Slack channel so that we could ask them questions at any time. That level of responsive service really helped the project go smoothly.”
The Results
By building on Kalix and using their preferred cloud provider, AWS, Coho AI was able to develop and launch its flagship product within just two months—much sooner than the four to six months that the team had initially estimated.
“With Kalix we got to market 75% faster compared to other solutions we had considered,” says Michael Ehrlich. “The project proved that Kalix enables us to add functionality without writing too much code or spending time on maintenance.”
Since launch, Coho AI has successfully negotiated its next funding round and gained dozens of paying customers. As new clients come on board, the Kalix platform scales seamlessly, delivering excellent performance and effortless availability.
“Kalix is truly almost a zero-maintenance platform,” says Michael Ehrlich. “As we grow, we can continue to focus on new features and enhancements that meet our clients’ needs. We don’t need to actively think about DevOps, or worry about how our infrastructure is going to scale.”
