Monday Apr 1st, 2024

How Auto QA Works: An In-Depth Guide to OttoQa’s Process

Streamlining QA with Automation: How Otto Integrates with Your Existing Forms

In the quest for excellence in customer service, leveraging advanced technology to enhance quality assurance (QA) processes is no longer just an option—it's a necessity and one of the low-hanging improvements we can make with the AI of today.

Enter Otto, a pioneering Auto QA platform designed to seamlessly blend with your current operations, particularly those driven by existing QA forms.

This detailed exploration uncovers how Otto revolutionizes QA workflows by infusing artificial intelligence (AI) into the heart of your quality assessment, all while respecting and utilizing the foundational QA forms you've already perfected. 

This is "HOW" Otto works

Identifying Your Call Types

The first step is straightforward: we identify the different kinds of calls your team handles. For instance, retention calls are critical interactions where customers are considering leaving your service. Recognizing this, we categorize each call type to ensure Otto knows exactly what it's analyzing.

Setting Up Otto with Your Call Types

Once we know your call types, we make Otto aware of them. This is as simple as telling Otto, "Here are our retention calls, where customers might want to cancel their service." Otto takes this information and prepares to analyze these calls with the specific attention they require.

Capture v2

Crafting Tailored Questions and Contexts for Comprehensive Evaluation

Once we've mapped out the landscape of your call types, the next phase involves translating your quality assurance form into the proper context so Otto's AI engine can not only comprehend but analyze it with a high degree of sophistication.

Leveraging Your Existing QA Forms

First, we take your current QA forms—the exact sheets you use to assess call quality. These forms are crucial because they contain the specific questions you've already been using to measure the success of customer interactions.

For every question in your QA form, we create a specific prompt for Otto. This is where the context comes in. We use the details and criteria you’ve set in your QA forms to instruct Otto on what exactly to listen for in the call transcripts.

Example:Capture v3

After setting up your QA forms in Otto, the next steps in the process are where the real innovation shines. Here's a step-by-step breakdown of how Otto transforms a simple customer service call into actionable insights, leveraging advanced technology like Deepgram for transcription and utilizing powerful AI models such as ChatGPT 4 and Claude for analysis.

Converting Calls to Text with Deepgram

Once the customized form is ready, and a call is recorded, the first stop is transcription. Otto sends the audio file to Deepgram, a leading speech recognition service.

Deepgram's advanced AI quickly and accurately transcribes the spoken words into text, ensuring every nuance and detail from the call is captured in written form. This process takes just a few seconds, turning audio recordings into a format Otto can analyze in-depth.

Compiling Data for Analysis

Once Otto receives the transcript back from Deepgram, it organizes all the necessary elements for a comprehensive analysis and get us ready for its AI journey.

  • System Prompt: This is an overarching guide for Otto and the AI, setting the stage for what kind of analysis will be performed on the transcript as a whole. It doesn't focus on the specifics of each QA question but rather prepares the AI for the type of conversation it will analyze.

  • Individual Question Prompts: For every question in your QA form, Otto creates a specific prompt. These prompts are detailed instructions that tell the AI exactly what to look for in the transcript to answer that particular question. They are based on the context and criteria you've provided for each question, targeting the nuances of each call type.

Deep Analysis with AI

In our quest to deliver the most accurate and insightful Auto QA analyses, we've honed in on two cutting-edge AI models: GPT-4 Turbo and Claude 3 Sonnets.

Our experience shows that these models excel in dissecting and understanding the nuances of customer service interactions the best/ Here's a brief overview of how we incorporate them into our process:

Utilizing GPT-4 Turbo and Claude 3 Sonnet

  • GPT-4 Turbo: Known for its depth and comprehensive analytical capabilities, GPT-4 Turbo is our go-to for handling complex question analysis. We've tailored our prompts to leverage its strengths, ensuring that each analysis captures the intricacies of customer interactions and meets our stringent quality standards.

  • Claude 3 Sonnet: When speed without sacrifice on insight is crucial, Claude 3 Sonnet steps up. Its efficiency in processing and evaluating calls makes it an invaluable asset, especially for questions requiring swift yet detailed analysis. Like with GPT-4 Turbo, our prompts are specifically designed to harness the best of Claude's abilities.

GPT-4 Turbo and Claude 3 Sonnet then go to work, analyzing the transcripts against the prompts. These AI models sift through the text, looking for responses that match the criteria outlined in the prompts. Their analysis covers everything from sentiment and intent to the specific content of the questions.

Receiving the Data Back

Compilation: After the AI models complete their analysis, the results are compiled into a structured format. This includes scores for each question, detailed feedback, and insights based on the analysis performed by the AI.

Return: The compiled data is then securely transmitted back to Otto. This step is just as secure and swift as the initial transmission, ensuring no delay in getting the insights back to you.

Review and Action: Once Otto receives the analyzed data, it's ready for review. Otto presents the results in an easy-to-understand format, allowing you to quickly see how each interaction measures up to your standards and where improvements can be made.

Capture v5

The OttoQA Difference: Crafted Excellence in AI Auto QA

While large Contact Center as a Service (CCaaS) entities might offer a watered-down Auto QA solution as part of a broader suite, OttoQA stands apart.

Our singular dedication to Auto QA only allows us to channel our efforts and resources into perfecting this service. Much like a master brewer focuses on creating a unique craft beer, OttoQA concentrates on developing the finest QA technology in the market.

Without the burdensome overhead costs that weigh down larger companies, we're nimble enough to adopt and invest in the highest-performing Large Language Models (LLMs). This investment ensures that our platform not only matches but often surpasses human accuracy in scoring, delivering superior and consistent results.

We would love to show you what we are doing and how we can help you fully automate your QA for pennies.

Click here if you would like to schedule a personalized Demo: