85% reduction in time to develop and deploy debugging tool with Gemini 2.5 and Cloud Run
Easy-to-use services empower engineers to spin up new configurations quickly for rapid prototyping
Managed, cost-efficient services such as Cloud Run and Memorystore enable Quench.ai to focus on solving customer challenges
Large context window of Gemini 2.5 enables faster agentic AI development
Quench.ai used Google Cloud for a scalable, cost-effective AI platform, accelerating feature development by 85% and boosting customer productivity by providing instant access to internal data.
Quench.ai CEO and Founder Husayn Kassai is no stranger to building a successful business. His previous startup, Onfido, became a global leader in identity verification, making more than $100 million annual revenue before he and his co-founders sold it to Entrust in 2024 for a reported $650 million.
It was while scaling Onfido that Kassai experienced the challenges that ultimately led him to launch his current venture, Quench.ai. Despite Onfido's robust onboarding and handover processes, employees were taking longer than expected to get up to speed. Because information was siloed in different systems across the company, it was difficult for employees to quickly access the information they needed to work.
After selling Onfido, Kassai resolved to solve these challenges with the launch of Quench.ai. By using AI to give employees instant access to the information they need, Quench.ai enables businesses to overcome bottlenecks in their processes, increasing the overall productivity of their workforce.
The Quench.ai team moved fast as it built the platform, testing ideas and developing prototypes quickly to solve its customers' problems. This required scalable, performant, cost-effective infrastructure. However, as the company evolved, it became clear its third-party infrastructure didn't meet these criteria, and was unable to support the company's long-term growth.
As it searched for a new cloud provider, Quench.ai was impressed with the scalability, reliability, and speed of Google Cloud. What's more, with the Google for Startups Cloud Program, designed for scale-tier AI startups, Quench.ai could access credits to expand their infrastructure and grow their business without stretching resources. "We were impressed with the commitment of Google Cloud to building long-term relationships with startups," explains Kassai. "Having access to a dedicated account manager, even at an early stage, gave us confidence we had the right provider to scale with."
We were impressed with the commitment of Google Cloud to building long-term relationships with startups. Having access to a dedicated account manager, even at an early stage, gave us confidence we had the right provider to scale with.
Husayn Kassai
CEO and Founder, Quench.ai
As a startup, Quench.ai keeps its team lean and prioritizes speed and agility as it focuses on building the platform its customers want. The company doesn't have a dedicated team of cloud specialists, so its cloud platform must be easy to use and manage. With Google Cloud, the Quench.ai team can quickly and easily set up the services it needs to prototype ideas and develop solutions without bottlenecks. "Our engineers can typically find any answer they need by logging into the Google Cloud console," explains Kresimir Nesek, VP of engineering and product at Quench.ai. "This gives them the independence to spin up infrastructure and do their work without any handholding, which is very important for a startup."
Our engineers can typically find any answer they need by logging into the Google Cloud console. This gives them the independence to spin up infrastructure and do their work without any handholding, which is very important for a startup.
Kresimir Nesek
VP of Engineering and Product, Quench.ai
Cloud Run is very simple to use and lets you spring up new prototypes without the overheads of time or specific infrastructure knowledge. You don't have to invest in training to learn a more complex setup, you can just get moving straight away.
Kresimir Nesek
VP of Engineering and Product, Quench.ai
With fully managed services such as Cloud Run and Cloud SQL, Quench.ai can focus on developing its services without worrying about its infrastructure. Engineers use Cloud Run to easily deploy small, independent services for new features and rapid prototyping. For example, Quench.ai is using Cloud Run to develop a "handover agent" – a smart tool that makes it easier for employees to ask questions to get up to speed with their role when they've recently taken on a new job.
Building on Cloud Run makes it easy for the Quench.ai team to experiment and bring new features to life, without having to manage complex architecture. "Cloud Run is very simple to use and lets you spring up new prototypes without the overheads of time or specific infrastructure knowledge," explains Nesek. "You don't have to invest in training to learn a more complex setup, you can just get moving straight away."
Memorystore is another solution that enables Quench.ai to work quickly while managing costs. The team uses it both as a job queue for orchestrating complex data-processing tasks and as a cache. By storing results from previous, expensive operations, such as calls to large AI models, Memorystore prevents Quench.ai from having to make those same calls again when they re-run their experiments. This means the Quench.ai team can iterate more quickly and cost-effectively during prototyping and ultimately deliver a better service to their customers at a lower cost.
With Google Cloud AI services, Quench.ai is reducing its development time even further. With the help of Google AI Studio and Gemini 2.5, for example, Nesek was able to quickly generate a detailed blueprint for a custom debugging tool for internal use. Thanks to the large context window of Gemini 2.5, Nesek could add a large amount of system logs then ask Gemini to analyze the problem and design a tool to solve it. Nesek then used an AI coding agent to generate the code for this design before deploying the completed tool to Google Cloud as a secure Cloud Run service, complete with authentication and authorization through Identity-Aware Proxy.
Previously, this process would've taken teams weeks from start to finish. With Google Cloud, Nesek managed to complete it in just two afternoons of coding and two hours for production deployment — a time reduction of more than 85%. This acceleration ensures Quench.ai can rapidly diagnose and resolve issues, consistently delivering accurate answers through their platform.
Being able to use Google Cloud to design this tool with no prior knowledge, and deploy it securely, with no need to build the underlying infrastructure, is a new world of software development. We can build custom tools for our projects to accelerate development in a way we couldn't before.
Kresimir Nesek
VP of Engineering and Product, Quench.ai
"Being able to use Google Cloud to design this tool with no prior knowledge, and deploy it securely, with no need to build the underlying infrastructure, is a new world of software development," Nesek observes. "We can build custom tools for our projects to accelerate development in a way we couldn't before."
As Quench.ai looks to expand its use of Gemini 2.5 to give customers even faster access to relevant company data, it's also working on building agentic AI capabilities to fetch information from customers' systems in a smarter, faster way. This includes a sophisticated AI assistant that could actively search multiple unknown systems to find the right information to offer even clearer answers to customers' questions about onboarding and handover scenarios. "Having a large context window is helpful for this, as it allows you to send entire documents rather than trying to figure out which parts of the document are relevant," Nesek explains. "That's where we see Gemini 2.5 fitting in nicely."
For Kassai and Nesek, this rapid, innovative, and cost-efficient approach to development wouldn't be possible without Google Cloud. "The fact that we have never had to spend time worrying about our infrastructure has allowed us to focus on solving our customers' problems," Nesek explains. "At the same time, whenever we have needed a new service to build our product, we've never had to look further than Google Cloud."
The fact that we have never had to spend time worrying about our infrastructure has allowed us to focus on solving our customers' problems. At the same time, whenever we have needed a new service to build our product, we've never had to look further than Google Cloud.
Kresimir Nesek
VP of Engineering and Product, Quench.ai
Founded in London in 2023, Quench.ai is an AI workplace assistant which enables teams to search across all of their company tools and data in seconds, and run AI agents to handle what slows them down.
Industry: Technology
Location: United Kingdom
Products: Google Cloud, Cloud Run, Cloud SQL, Gemini 2.5, Google AI Studio, Identity-Aware Proxy, Memorystore