This website requires JavaScript.

Announcing AutoQA: Score Support Conversations Automatically

Apr 7, 2023
Scale your QA by 50x with our KlausAI-powered solution. AutoQA scrutinizes your every support interaction for automatic, actionable reviews – 100% coverage guaranteed.
Klaus AutoQA dashboard.

In today’s highly competitive business landscape, making the right decisions is critical. Yet, many organizations rely on shockingly small sample sizes when evaluating support interactions. 

Today, we are thrilled to announce AutoQA for Klaus — an AI-powered force to shake up and automate how you manage your support quality. 

Get started today 

On average, only 2% of support interactions are manually evaluated. Unfortunately, that’s not enough to provide a comprehensive understanding of your support quality and make the right improvement decisions. 

Moreover, relying only on human reviewers can lead to biased results, which can be detrimental to your business. These are often caused by subconscious prejudices or preconceptions, ill alignment of how to interpret rating categories, or lack of calibration to keep evaluators aligned. 

On top of bias, loss of focus or lapse in human judgment make it easy for errors to happen – imagine reading through a long conversation thread, trying to catch spelling errors. 

Add small sample sizes and evaluation processes prone to errors into the mix, which can lead to skewed data that can’t be trusted. 

But fear not; there’s a better way to evaluate your support interactions: AutoQA.

Previously, we dipped our paws into the many possibilities of AutoQA, laying a strong foundation by deep-diving into support interactions and automating the discovery of conversations that provide valuable learning opportunities.

By enriching your help desk data and introducing advanced AI-powered functionalities like

  • Sentiment analysis for finding conversations where the customer displayed delight or frustration
  • Spotlight for discovering conversations that are critical to review
  • Conversation Insights and its powerful data-mining capabilities to make sense of your help desk data

into your day-to-day QA processes, we’ve built a robust base for expanding Klaus’ AutoQA arsenal and advancing towards unprecedented 100% coverage. 

We are now thrilled to take automated quality assurance in Klaus to the next level, providing a revolutionary approach to make it more efficient, accurate, and smarter than ever before. 

How Klaus’ AutoQA is changing the way you do QA: 

  • Standardize your QA processes and achieve 100% coverage by automatically scoring every agent and support interaction across multiple categories and languages.
  • Klaus’ AutoQA acts less like an assistant and more like a coworker. No model training is required with our plug-and-play solution. Simply step in and see the magic in action. 
  • AutoQA data is exposed to Klaus’ powerful dashboards, so you can put your laser-sharp focus on improvements and setting your team up for success.

100% coverage is no longer a far-fetched goal

As companies strive towards 100% coverage – the holy grail of conversation reviews – it is no surprise that achieving it solely on human powers continues to be a wild goose kitten chase. 

But it’s no longer impossible, thanks to Klaus’ AutoQA. 

By leveraging the powers of AI, AutoQA automatically scores support interactions across multiple categories and languages without requiring additional resources. 

Not only does it eliminate the issues related to manual conversation reviews, with AutoQA, you can:

  • Handle huge ticket volumes in less time without missing hidden gems of information
  • Ensure all your agents and categories are automatically covered
  • Improve efficiency and your team’s overall workload

But there’s more to AutoQA than speed and effectiveness. It provides a scalable solution, delivering increased accuracy and the ability to identify potential errors that may be missed by your human reviewers – ensuring that your support quality is always top-notch.

Identify areas of improvement, eliminate compliance risks, and provide targeted coaching to your support reps. 

Trust that with AutoQA, your interactions are being consistently evaluated and your team delivers the high-quality support experiences customers expect and deserve.

To each their own scorecard categories. Or is it?

When it comes to customer service quality, the definition and interpretation can differ across support teams. 

At Klaus, research is the basis for every action taken. Before building our AutoQA solution, we analyzed the scorecards of more than 1000 customers to understand what quality means for contact centers. 

As a result, we identified the most commonly used categories used in millions of reviews and analyzed how reviewers interpret them. Additionally, we determined the most popular rating scales for each of these categories to mimic and automate the processes used in manual QA.

Most widely used scorecard categories

Hungry for more insights like this? Download our Customer Service Benchmark Report here.

Our research revealed that some categories are highly specific to individual businesses, such as product knowledge or following internal processes, requiring customized solutions to evaluate them accurately.

However, many categories are universal and can effectively be evaluated automatically. For example, good spelling and grammar are essential to all businesses, as is opening and closing a conversation. Most companies follow similar rules and guidelines for these categories, thus making them well-suited for automation via AutoQA. 

Manual proofreading for grammar is a thing of the past. It can be time-consuming and tedious, where many mistakes slip through the fingers. Instead, we should focus on automating grammar using powerful grammar detection tools.

Say goodbye to the old way of doing things and embrace the power of AI. With the help of automated grammar detection, we can enhance our communication and streamline our processes, freeing up time to focus on more creative and strategic tasks. It’s time to welcome a new era of efficiency and productivity.

Andre Tättar
Andre Tättar
Machine Learning Researcher at Klaus

Klaus’ AutoQA has built-in benchmark knowledge of what categories represent and what scales to use. This way, your team can focus on where the communication nuances leave machines uncertain and human evaluation is needed.

The underlying scientific and technological ability of AutoQA is multifaceted. We use methods from naive but ultra-transparent phrase matching and catching textual signs of neural coupling to orchestrated switching between OpenAI-s API and multiple self-hosted large language models (LLM) – making up KlausGPT. With it, we are adding more automatically scored categories at a pace not imaginable in the pre-LLM world. 
Mervi Sepp Rei, Phd
Mervi Sepp Rei, Phd
Head of Machine Learning and Data at Klaus

Klaus’ AutoQA currently offers 8 different categories to speed up your QA processes and identify errors potentially missed by your human reviewers: 

  • Grammar & Spelling
  • Solution
  • Tone
  • Empathy
  • Readability
  • Greeting
  • Closing
  • Comprehension

New ways to customize AutoQA to your needs

There are times when one size does not fit all. We’ve now incorporated more flexibility and customization options to tailor AutoQA to suit your unique ways.

  • Ignore list for Grammar & Spelling

    We understand that not every word or phrase fits neatly into a predefined grammar rule. To address this, we’ve introduced an Ignore List feature. Now, you can list the words and phrases so AutoQA understands which ones are a-okay with you.


  • Accepted greetings and closings

    You can now define the acceptable ways to greet or say goodbye to your customers. Once listed, AutoQA will detect whether the approved greeting was used in a conversation.

Global reach with multilingual AutoQA

Klaus AutoQA is designed to be multilingual, making it a truly global solution.

Many contact centers operate in multiple languages, and having equal reviewing capacity in all of them can be challenging. We have transcended the language barriers by utilizing the most advanced large language models. KlausGPT-powered AutoQA categories are all multilingual. 

This ensures you have a comprehensive view of your interactions, regardless of the spoken language.

Parts of our advanced AutoQA features may use OpenAI’s GPT language models, which, while offering potentially superior generative performance compared to other LLMs, may face adoption limitations due to strict security and compliance regulations.

To address this issue, we have developed KlausGPT. Our secure solution enables companies to leverage the power of specifically tuned language models for scoring customer service conversations, while not sacrificing their data security and privacy measures. With KlausGPT, businesses can leverage the recent advancements in AI without risking their security and compliance.

Joonathan Mägi
Joonathan Mägi
CTO at Klaus

Full QA visibility into 100% of your conversations

With Klaus’ brand-new AutoQA dashboard, you can access a complete overview of all these automatically scored interactions. 

Klaus AutoQA dashboard - 1

This level of visibility is unprecedented and enables you to identify patterns and trends that may not have been visible before.

In addition to valuable data points, get an automatically compiled overview of your top- and low-performers, with an option to easily create Coaching sessions with the quickest improvement potential.

Klaus AutoQA dashboard - 2

Automated coaching insights

Automating conversation reviews and tracking performance KPIs is only the primary step toward improving your customer service quality.

Putting the learnings into practice will take your support quality’s excellence to the next level. Recurring coaching sessions are what tie it all together. 

AutoQA - Coaching session view

All relevant insights gathered through AutoQA are sprinkled across Klaus’ coaching capabilities for even better actionable insights. 

Easily spot changes in performance metrics, praise top performers, keep an eye on team members needing a helping paw, and monitor the impact coaching sessions have on your team’s results. 

Klaus’ AutoQA features will be released in phases, so stay tuned for more news soon!

The future of AutoQA is now

You don’t need to settle for inadequate sample sizes and biased results. 

Klaus’ AutoQA solution provides a more efficient and accurate alternative to the manual review process, ensuring that 100% of the conversations and all participating agents are evaluated in a standardized and objective manner. 

While AI technologies are evolving at an unprecedented speed, changing how you work won’t take away your job. Instead, it will become more engaging and fulfilling as you will have more time to handle more intricate cases that require the ultimate touch of human intelligence.

By utilizing AutoQA, support teams can easily pinpoint improvement areas and provide their agents with targeted coaching and training. Ultimately, this results in better support quality, improved customer satisfaction, and a more productive support team.

Make the right decision for your business and customer service with AutoQA, the next generation in support conversation reviews. 

Never miss an update

By subscribing you agree to Klaus' Privacy Policy and would like to get educational content to your email.