Enhanced Customer Experience through a Modern Data Platform

Executive Summary

InConcert needed to modernize its data platform to optimize customer experience. In order to address these challenges, we designed and implemented a multi-tenant analytical data architecture. We helped InConcert leverage StarTree’s Cloud to enhance its omnichannel contact center SaaS platform. StarTree provides strong foundations for scalable real-time user-facing analytics capabilities. Furthermore, the new Data Platform decouples analytical workloads from operational systems. This solution incorporated a centralized Data Lakehouse fully hosted on AWS with High Availability, real-time OLAP capabilities powered by Apache Pinot, and a semantic layer powered by Cube.dev to improve efficiency and flexibility. As a result, InConcert reduced costs and maintenance, improved data accessibility, and gained real-time insights, enabling faster decision-making and a superior customer experience. All this with zero downtime production migration. This was fully unnoticed by customers. Leveraging historical data storage for cost efficiency, coupled with the aforementioned enhancements, enables the advancement of Machine Learning and AI applications.

About the Company

InConcert is a trusted technology partner. Leader in Latin America and Spain in contact center solutions that support businesses through every stage of their customer experience (CX) transformation. Their comprehensive omnichannel platform integrates key solutions including cloud contact centers (InConnect), virtual agents powered by Agentic AI (InAgent), marketing automation and CRM (InFunnel), quality management and speech analytics (InSpeech), and workforce management (InTeam). InConnect centralizes communication across phone, chat, email, SMS, WhatsApp, and social media to streamline operations and enhance CX. InAgent enables the creation of intelligent virtual agents to automate and improve customer service. InFunnel automates marketing and sales workflows, allowing teams to manage leads, launch campaigns, create landing pages, and handle customer relationships across all channels.

The Challenge

With a clear goal to scale and enhance customer experiences, InConcert set out to modernize its data platform. Their existing infrastructure, while reliable for earlier demands, had begun to show limitations with scalability and cost-efficiency, especially as workloads grew in complexity. This led to high resource consumption for collecting, processing, and storing data in reporting tables. Additionally, reporting latency and access to real-time data became increasingly challenging as the system scaled, prompting InConcert to explore more modern, responsive architectures. The core InConnect system—responsible for both real-time operations and historical data queries—was being strained by reporting demands.

Furthermore, the lack of a semantic layer capable of unifying diverse data sources made it difficult to ensure consistent, governed analytics across tenants. They were also experiencing high expenses in storage. The identified problems mainly relate to:

• High resource consumption and storage use in report generation.
• Real-time supervision limitations due to inefficient architectures.
• Lack of integrated querying of structured and unstructured data sources.
• Scalability and performance challenges in CRM and Contact Center systems.

To support this strategic initiative, InConcert partnered with Mutt Data to co-design a scalable, unified data platform over AWS that could meet evolving business and customer needs.

The Solution

To tackle these challenges, we developed and deployed a multi-tenant data architecture hosted on AWS designed to expand the real-time capabilities of their omnichannel contact center SaaS solution, unlocking scalable, low-latency, user-facing analytics through StarTree Cloud powered by Apache Pinot. The goal was to build a Modern Data Platform by separating analytical workloads from operational systems. The core elements of this solution included:

• Deployment of a message bus to facilitate the democratization of events and enable the separation of information producers from consumers.
• Centralization of information.
• Centralization of access points, which allowed InConcert to streamline data delivery and simplify access across teams.
• Establishment of a centralized data lakehouse architecture aimed at unifying and generating data for reporting purposes, with considerations for latency and SQL interface accessibility.
• Implementation of a real-time OLAP engine using Startree for user-facing reports and operations.
• Integration of a semantic layer, Cube, to facilitate user-facing decoupling of engines and storage from logical consumption. This ensures flexibility in transitioning between technologies without depending on downstream user modifications. Cube was implemented as a universal semantic layer, providing a unified set of definitions for business metrics, dimensions, and logic that could be centrally managed and reused across hundreds of environments. This enabled InConcert to scale personalized data experiences without duplicating effort or infrastructure. Through Cube’s REST and GraphQL APIs, data consumers—ranging from internal apps to external SaaS clients—could securely and efficiently access curated datasets with consistent business logic, accelerating development cycles and reducing the need for engineering support.
• This transformation not only optimized performance and costs but also laid the foundation for future expansion and innovation.

How the Solution Works: A StarTree and Mutt Data Collaboration Story

Reporting and analytics responsibilities needed to be delegated to a separate component rather than the operational core to reduce the Inconnect system's workload. To achieve this, a message bus was introduced (MSK), enabling the decoupling of operational processes from reporting and analytics.

Next, data produced by the core component and residing in the message bus is offloaded to the Data Lakehouse hosted on AWS. This approach allows InConcert to consolidate both historical and operational data in a cost-effective, scalable environment that supports advanced analytics, AI-driven insights, and seamless integration with cloud-native services.

On top of the Data Lakehouse, the curated data is ingested into a real-time OLAP engine, enabling faster and more efficient analytics. Additionally, this engine is used for specific use cases that require direct ingestion from the message bus to power real-time dashboards.

This is where StarTree entered the picture. Their StarTree Cloud, a managed Apache Pinot service, provides seamless synchronization between the Lakehouse and the message bus. StarTree/Pinot, in turn, supports both batch and real-time processing.

The introduction of Startree Cloud and the decoupling of analytical responsibilities from operational ones allowed InConcert to no longer require pre-aggregating data to generate reports, which also reduced the amount of duplicated data. By implementing this solution, InConcert was able to maintain a structured data pipeline following the Medallion Architecture which is quite a standard nowadays.

Finally, data is consumed by clients and tenants (internal applications and SaaS users) through a headless semantic layer powered by Cube.dev. This layer abstracts the physical data representation and engine complexity, exposing a rich set of APIs to interact with the data. It also enables custom models for each tenant while ensuring all data is sourced from the same table. The semantic layer and its definitions are more useful for Machine Learning and AI because they facilitate the interpretation of the data being read to extract useful information.

This flexible design allows for the creation of composite dashboards and reports that span multiple products and tenants, unlocking new strategic insights for multi-service clients.

The end-to-end batch pipeline is orchestrated using Apache Airflow.

Impact

Together, we implemented a modern multi-tenant architecture that significantly optimized operational efficiency, reduced licensing costs, and positioned the platform for long-term scalability. By optimizing data exchange through a message bus, the system now operates more efficiently, with reduced dependency on manual intervention. Additionally, the introduction of new features facilitated the development of machine learning models and improved reporting capabilities.

Centralizing all data into a single repository streamlined operations, enabling faster access to information for analytical purposes. With real-time data reporting powered by Pinot, decision-makers now access insights promptly. Moreover, there has been a notable reduction in responsibilities associated with the core transactional system, enhancing overall operational efficiency.

This foundation also enabled a unified view across multiple InConcert products and tenants, empowering customers with an integrated understanding of their operations, particularly valuable for organizations leveraging multiple solutions in the InConcert suite.

With Cube in place, InConcert dynamically serves analytics tailored to each customer's needs without compromising system performance or increasing operational complexity.

Its multi-tenant architecture, combined with centralized API delivery and logic reuse, drastically reduced development and maintenance overhead. Business teams gained greater autonomy through consistent, self-serve analytics tools, while the platform scaled to meet the growing demand for real-time, granular insights.

Additionally, the modernization significantly reduced the engineering hours previously required for supporting analytics and reporting issues. The system’s new architecture also lowered the cost and effort of delivering new analytics features, including massive data exports and on-demand data access.

Results

By combining StarTree’s real-time OLAP engine with Cube’s flexible semantic layer and a robust Lakehouse foundation, InConcert unlocked significant performance gains, operational cost savings, and greater data agility. Some of the greatest results of this project include:

• Data freshness improved, eliminating the need for pre-aggregated data and reducing storage requirements.
• Centralizing the system onto a single platform reduced both maintenance and incident-related costs.
• The migration was completed with zero downtime.
• The platform now features high availability (HA), ensuring a backup system is in place in case of failure.
• Regional contingency support is in place.
• Decoupled reports led to a notable increase in performance speed.
• Customers can tailor and create their own reports, addressing all their data requirements.

Wrapping Up

The collaboration between Mutt Data and StarTree has successfully reshaped how InConcert handles customer experience data.

By implementing a modern data platform, InConcert transitioned from a costly, high-maintenance on-premises system to a scalable, cloud-ready architecture designed for efficiency, flexibility, and real-time analytics.

Through StarTree Cloud, the solution ensured low-latency, high-throughput analytics, significantly reducing the strain on operational systems. Meanwhile, the introduction of a centralized Data Lakehouse streamlined data management, enabling seamless cross-product analysis, advanced reporting, and AI-driven insights. The integration of a message bus, a semantic layer, and orchestration tools further enhanced system reliability and adaptability.

As a result, InConcert now benefits from reduced costs, improved data accessibility, and real-time decision-making capabilities, laying the foundation for future growth and innovation. By combining cutting-edge technologies with a strategic approach, Mutt Data and StarTree empowered InConcert to deliver a superior customer experience while optimizing its internal operations.

Cube’s semantic layer was crucial in ensuring that analytical workloads were not only real-time and reliable but also reusable, secure, and fully adaptable across multiple tenants. Its role in abstracting technical complexity while enabling rich, customizable data experiences made it a key pillar of the solution alongside StarTree’s OLAP engine.

About StarTree

StarTree enables companies to build high-performing real-time applications on vast datasets via their fully-managed platform, Startree Cloud, powered by Apache Pinot. While there are many real-time analytics databases on the market, StarTree Cloud was designed for scalability while providing sub-second latency results when serving millions of users. StarTree Cloud can ingest a petabyte or more of data and perform fast aggregations at rates of 100,000+ Queries per Second (QPS).

About Mutt Data

Mutt Data empowers companies to thrive in an AI-driven future. As a trusted StarTree partner, we specialize in building modern data platforms with user-facing, real-time analytics. Our solutions integrate StarTree seamlessly into broader data ecosystems, leveraging the latest technologies and best practices in software engineering, cloud infrastructure, data architecture, data engineering, and machine learning.

Ready to unlock the full potential of your company? Schedule a call with us to discover how our expertise can elevate your capabilities.