Our client, a customer engagement technology provider, designs its own Contact Centre as a Service (CCaaS) solution. With their commitment to driving impact and elevating customer experiences, the company decided to add AI-powered capabilities to their solution.

Business Challenge

Our client wanted to help their customers gain more visibility into their activities. Namely, they wanted to help them understand the “why” behind each contact center conversation and enable them to make better business decisions.

Traditionally, contact center management and QA personnel listened to selected call recordings to gain insight into the interactions their agents had with customers. This helped identify coaching opportunities and understand how to drive increased efficiency and revenue. However, this process only allowed them to cover a very small portion of the entire conversation landscape, leading to poor outcomes, compliance risks, and missed opportunities.

To meet the challenge, the customer decided to build an AI speech analytics solution. Since building it required AI expertise that they didn’t have in-house, the client hired Coherent Solutions as a suitable and reliable partner. Our team was tasked with delivering it within a tight timeframe.

AI-Driven Speech Analytics


The Coherent Solutions team dedicated 18 months to designing a next-gen speech analytics solution for the client’s call center portfolio. This innovative solution automates the process of analyzing call center interactions by transcribing every call recording from voice to text.

Leveraging AI technology, it can also analyze each transcript to summarize conversation content, identify key topics, and determine sentiment. The product can identify critical reasons for both customer satisfaction and customer churn by figuring out meaningful words and phrases that have positive and negative connotations and determining the emotional character of speech passages.
Additionally, AI-driven speech analytics provides actionable insights from individual conversations, such as uncovering genuine customer opinions on new products or services, identifying agent performance levels, and assessing the percentage of conversations meeting quality standards.

Based on the insights above, the solution provides aggregated analytics that help the business make better, data-informed decisions, improve service quality, reduce costs, and drive better call quality and compliance.

During the initial phase, we conducted AI-related research, developed prototypes to showcase AI model capabilities, and explored various options. Additionally, we crafted UX designs, validated them with the client, and refined the reference architecture.

Around three months into the development phase, we demonstrated the product to potential customers. Soon after, the MVP was launched and made available to customers.


The resulting solution fully meets the customer requirements in terms of functionality and usability. The solution comes with the following key elements:

  • Transcription Service: The solution converts daily call recordings into text using DeepGram. It comprises multiple serverless components provided by Azure functions, each handling a specific step of the workflow. A highly scalable and efficient event system facilitates seamless coordination among these components.
  • AI Analysis Service: Utilizing a combination of LLM and non-LLM AI models, our self-hosted AI analysis service extracts valuable insights from each conversation transcript.
  • Spokn AI Back-end: Serving as the backbone of the Speech Analytics solution, Spokn AI back-end employs a mix of .NET-based serverless components, APIs, Azure-powered event and messaging systems, and data storage solutions. It facilitates integration with external systems, orchestrates processing workflows, aggregates data, and delivers insights and analytics.
  • Spokn AI Web Application: Our Vue.js front-end application presents high-level data to end users, such as Contact Centre Managers and QA specialists, through a user-friendly yet sophisticated dashboard. This dashboard enables data-informed decision-making by providing an overview of key topics driving traffic in contact centers, conducting trend analysis of evolving customer needs, and facilitating comparison of current results with past performance.

All these components are hosted on and operated from the Azure cloud platform as a multi-tenant SaaS solution.


Satisfied with the outcomes, the customer enlisted Coherent Solutions to further enhance the designed solution. Overall, our partnership not only helped the client enhance service quality but also facilitated customer satisfaction, bolstered profits, and improved sales efficiency and compliance.

Looking for an experienced AI software vendor? Coherent Solutions is your ideal partner.

With our extensive expertise in AI services, we are prepared to augment your team and guarantee the prompt and budget-friendly delivery of your project. Reach out to us today to discover how our AI experts can help you from ideation to production.

Book Free Consultation

We Deliver

Streamlining Eyeglass Orders for a Leading Eyewear Retailer
Virtual Try-On Solution

Get a Free Consultation with Our Experts!

Simply fill out our contact form below, and we will reach out to you within 1 business day to schedule a free 1-hour consultation covering platform selection, budgeting, and project timelines.

This field is required. Maximum 100 characters.
This field is required. Maximum 100 characters.
This field is required.
Only digits are applicable.
Maximum 2000 characters.
* Required field

An error occurred sending your message.
Try again or contact us via

Message sent!

Here's what happens next:

  1. Our sales rep will contact you within 1 day to discuss your case in more detail.
  2. Next, we will arrange a free 1-hour consultation with our experts on platform selection, budgeting, and timelines.
  3. After that, we’ll need 1-2 weeks to prepare a proposal, covering solutions, team requirements, cost & time estimates.
  4. Once approved, we will launch your project within 1-2 weeks.