Klaus: Harnessing AI to deliver memorable customer experiences at scale

About Klaus

Founded in 2018 in Estonia, Klaus's AI-powered platform helps companies around the world run effective customer service QA processes, coach agents, and boost retention.

Industries: Technology
Location: Estonia

Tell us your challenge. We're here to help.

Contact us

About Revolgy

Google Cloud Partner Revolgy brings its customers cloud technologies adapted to their needs, with rapid time-to-value and smart solutions.

With Google Cloud, Klaus has built a highly scalable, AI-led platform that enables hypergrowth companies to review and optimize their customer service.

Google Cloud results

  • Enables customer service QA with AI and powerful data management tools at unprecedented scale and complexity
  • Connects with all standard customer service help desks in less than an hour for near-frictionless onboarding of new clients
  • Builds a leaner, more agile workflow: reducing developer reliance on DevOps specialists

Analyzes 10 million+ interactions daily in 100+ languages

Happy customers are repeat customers, but keeping them happy at scale can be a challenge. At a certain size, many businesses end up outsourcing their customer service handing the job over to trained specialists. Across the world, businesses are focusing more of their resources on customer service. In 2021, analysts at Grand View Research estimated the customer experience management market to be worth $8.79 billion and set to grow 18% every year between 2022 and 20301.
1 https://www.grandviewresearch.com/industry-analysis/customer-experience-management-market#

Even with help from third parties, however, customer service is an ongoing process, requiring quality assurance (QA) measures to maintain high standards. But the larger a company gets, the more complex and expensive QA becomes. That's where Klaus comes in, with an AI-led solution that simplifies QA for companies of every size. "We provide an end to end solution to help businesses understand their customer interactions," says Joonathan Mägi, CTO at Klaus. "Firstly, we use AI and statistical modeling to find the most valuable interactions, facilitate the review process and provide insights on 100% of conversations. Then on the other side, we provide coaching and other tools to act on these insights and continuously improve customer service."

Founded in Estonia in 2018, Klaus's unique approach to customer service management has found great success, winning the company globally recognized clients. Any platform requires a solid foundation, and for Klaus, that was Google Cloud.

A new paradigm for quality assurance

Traditionally, QA for customer service has meant choosing a small sample of recorded interactions—often as small as 1%—and then reviewing and scoring them. These scores provide the raw data that analysts can mine for insights into customer satisfaction, agent communication, and knowledge trends, as well as point to conversations that can be educational. However, the process is manual, often long and painstaking, and with such a small sample size, of limited value. As a business grows, finding the most valuable interactions to review becomes increasingly like finding a needle in a haystack.

By harnessing powerful data analytics and cutting-edge artificial intelligence, Klaus has built a QA platform that can ingest the totality of a company's customer interactions, rather than a small sample. The platform can then sift through them, teasing out the most valuable ones for review, and automatically score the most tedious and error prone categories for human evaluators.

Although AI is a fundamental part of Klaus, Mägi stresses that the company is not seeking to make AI the entirety of the platform. "The AI in our platform is there to help the human analysts, not replace them," he explains. "By replacing biased manual sampling with AI curated samples, and automating the basic scoring process, Klaus helps QA analysts to work through more interactions, more efficiently, and gain deeper insights."

Building the foundations of an AI-led platform

When it came to building its platform, Klaus chose very early on to build it on Google Cloud rather than another cloud provider, primarily because of its ease of use.

"With Google Cloud, we knew that we didn't need to worry about scalability," says Mägi. "But we loved how developer friendly it is. We didn't have to hire platform engineers or DevOps specialists. Our product engineers could do everything themselves."

Klaus runs a micro-services-based architecture orchestrated with Cloud Run and Google Kubernetes Engine (GKE), along with networking tools like Cloud Load Balancing and Cloud Armor for smooth, secure operations.

"With Google Cloud, we knew that we didn't need to worry about scalability. But we loved how developer friendly it is. We didn't have to hire platform engineers or DevOps specialists. Our product engineers could do everything themselves."

Joonathan Mägi, CTO, Klaus

The persistence layer, where the data is ingested, processed, and stored, is built with Cloud SQL for PostgreSQL and Cloud Spanner. "What's great about Cloud Spanner is that it takes care of the sharding of the data on its own," says Mägi. "That means we can dynamically scale up if, for example, we have to onboard a large customer, and scale down, so that we can keep costs as low as possible."

With the data ingested and processed, Klaus can run it through its own proprietary AI components, which form the core of the platform's offering. These are fed with data pipelines built with BigQuery, Cloud Composer, Cloud Dataflow and elements of Vertex AI.

For Mägi and his team, the most welcome aspect of Google Cloud has been the transparency of the platform: "When I talk to our engineers, they are consistently impressed with the visibility that we get into the system with tools like Cloud Monitoring and Cloud Trace," he says. "It means we can debug and optimize our services very quickly."

With such a complex ecosystem that was always evolving as the company grew, Klaus wanted to simplify its billing as much as possible, which is why the company engaged the services of Google Cloud Partner Revolgy. In addition to helping Klaus optimize its infrastructure costs, Revolgy kept the company abreast of the latest developments with Google Cloud and its products. "Every project we do on Google Cloud, we orchestrate the billing through Revolgy so it's all nice and simple," says Mägi. "What's more, they have a lot of experience and contacts with Google Cloud and they can pass that on to us, whether it's best practices or being aware of new features coming out. It's been a great collaboration."

For Revolgy, the partnership with Klaus has been mutually beneficial. "Working with Joonathan and the Klaus team has been a pleasure from the very first moment. We knew how important the customer experience is for our partners and considering we had in front of us an extremely skilled Google Cloud team, it wasn't difficult to align on common goals," says Vlad Birsan, Chief Sales Officer at Revolgy.

"It is not your typical vendor-customer relationship, but rather a balanced collaboration," explains Stepan Kaiser, Cloud Services Lead at Revolgy. "Joonathan very often provides valuable feedback to Revolgy and Google Cloud and helps us improve our services."

"What's great about Cloud Spanner is that it takes care of the sharding of the data on its own. That means we can dynamically scale up if, for example, we have to onboard a large customer, and scale down, so that we can keep costs as low as possible."

Joonathan Mägi, CTO, Klaus

A platform designed for scale, complexity, and agility

With Google Cloud at its core, Klaus has built a platform that integrates powerful analytics with AI to help hypergrowth companies solve complex customer service challenges at scale. "Google Cloud has enabled us to operate at massive scale and complexity compared to traditional QA platforms," says Mägi. "We ingest and analyze more than 10 million customer interactions a day, in more than 100 languages, and those numbers are growing every day."

As impressive as the platform is, Klaus has worked hard to reduce friction as much as possible in onboarding new clients. "Unless they are running a particularly complicated customer service setup, most companies can be up and running on Klaus in less than an hour," says Mägi.

A key part of Klaus's success is due to how quickly its developers are able to iterate on Google Cloud and bring out new products and features that adapt to clients' individual needs. "We invested heavily into our machine learning pipelines early on, which means that there is not much of a delay between our AI research, model development and product implementation. Thanks to the flexibility of the pipelines we can offer custom data handling for specific customers that would be impossible with conventional ETL," explains Mägi. "We could offer our clients robust solutions much more quickly as a result."

"Google Cloud has enabled us to operate at massive scale and complexity compared to traditional QA platforms. We ingest and analyze more than 10 million customer interactions a day, in more than 100 languages, and those numbers are growing every day."

Joonathan Mägi, CTO, Klaus

Although it deals with very large clients, Klaus itself is still relatively small, with fewer than 70 employees. As a startup that is still in its own phase of extreme growth, every resource is precious, and Mägi would rather spend time and money working on core business problems than hiring DevOps specialists. With Google Cloud, he explains, he hasn't had to: "At this stage, our developers can take care of their own needs on the platform. Google Cloud has allowed us to stay lean."

Klaus is currently looking ahead to another year of rapid growth, after closing a $12 million round of Series A funding towards the end of 2022. As well as drumming up new business, the company is also seeking to expand the capabilities of its platform. "Right now, Klaus works on customer interactions that have already happened," says Mägi. "We're in the early stages of a project that will use all insights that we can generate to assist customer service agents in real time. That's really exciting to look forward to."

Tell us your challenge. We're here to help.

Contact us

About Klaus

Founded in 2018 in Estonia, Klaus's AI-powered platform helps companies around the world run effective customer service QA processes, coach agents, and boost retention.

Industries: Technology
Location: Estonia

About Revolgy

Google Cloud Partner Revolgy brings its customers cloud technologies adapted to their needs, with rapid time-to-value and smart solutions.