AWS Contact Center

CSC Generation transforms customer support with Amazon Connect


In this blog post, CSC Generation Holdings, Inc. (CSC Generation) shares why it moved to Amazon Connect to support customer service operations across its multiple brands. CSC Generation is a multi-brand technology platform that helps retailers such as One Kings Lane and Sur La Table transform into high-performing, digital-first brands. These retailers’ primary customers are individuals and families who are passionate about curating their perfect living environment. CSC Generation’s customer service operations offer various levels of support over voice, chat, email, and web ticketing to its brands. Each of these brands has its own support channels and queues, though some co-locate their offices.

To help retail brands become digital-first, CSC Generation needed a contact center solution that was flexible and offered built-in artificial intelligence (AI) capabilities to improve agent efficiency and self-service experiences for customers. As a result, CSC Generation’s development team built SnapBPO, its proprietary AI-powered BPO enhancement and productivity software. In a conversation with Andrew Templeton, Chief Technologist at CSC Generation, we delve into the challenges and opportunities that CSC Generation and its SnapBPO software seized by migrating to Amazon Connect.

What were the customer service challenges CSC Generation faced and the opportunities it seized by migrating to Amazon Connect?

The various CSC Generation brands and subsidiaries have different workflows, including interactive voice response (IVR) structures, for handling customer interactions. We knew we could encode the full breadth of these configurations and business rules into one contact center solution with Amazon Connect, and it was a huge opportunity to make future work more efficient across our brands. It is also easy to add sophisticated new features and experiences for customers via our SnapBPO software, like automating “Where’s my order?” type calls and providing real-time customer information.

It sounds like enabling customers to self-serve is important for CSC Generation. How have you integrated self-service capabilities into your operations?

In addition to providing self-service options to the customer facing websites for each respective brand, we enable our most common workflows to be serviced entirely by voice without number prompts in an IVR. This means that our customers can describe what they need help with in their own words. We use Amazon Lex, in addition to the SnapBPO solution, to answer our most common customer questions in complete sentences. The result has been that our average handle time (AHT) for “Where’s my order?” type calls is now under a minute.

You’ve also been experimenting with the potential of generative AI. Can you share how this technology has changed how you design and deliver self-service experiences for customers?

We no longer have to know all of the various ways a customer could phrase questions. Using generative AI, we can pre-generate lists of utterances in Amazon Lex, and use more robust, but slower, NLP models at runtime to read and understand inquiries from customers. For example, when a customer asks two distinct questions, like “Where is my order? And I forget, did I get the faster shipping on the towels?”, the SnapBPO self-service bot can identify each sub-inquiry and respond, “Your order from September 3rd will arrive tomorrow by post, and you did pick expedited shipping for the towels. What else can I help you with?”

Further, using Retrieval Augmented Generation (RAG), we are reindexing all of the information we have about products, policies for the business, and customer order history and behavior. This enables us to rapidly search contextually relevant information for a customer conversation in real-time.

With technology like generative AI-supported self-service experiences, naturally the element of customer trust comes into question. Can you explain how you are maintaining customer trust in your operations?

Amazon Lex classifies our sentences and transcribes phrases from customers, backed by all of AWS’s security and privacy controls. Our use of generative AI in this context is a lot more mundane than people realize. We’re primarily concerned with interpreting questions like, “What kind of sentence is this” and “Is the order number the first or second sequence of spoken digits?” By using AI in this way, and replacing PII data such as order number or customer name, we can leverage the best models without sharing sensitive information because we’re only sharing sentence structure.

The agent experience is also critical to CSC Generation’s customer support strategy. How did the migration to Amazon Connect impact agent onboarding and overall service delivery costs?

From an onboarding perspective, the single sign-on with Amazon Identity Center makes administration easier for our IT team. We just add an agent as a user to the application and it’s synchronized because of the Security Assertion Markup Language (SAML) integration. As we transition our call centers to the Amazon Connect agent workspace, our customer service agents will only need to familiarize themselves with one unified interface. Given that our agents occasionally switch between our different brands due to varying seasonality, this approach simplifies our workforce management. From a cost perspective, our telecommunications costs per agent are about 80% lower with our SnapBPO AI software and Amazon Connect than the solutions we were using before.

Can you share more details on your vision for a consistent agent experience across brands and how this can impact average handle time (AHT)?

We’ve unified the queue types and workflows by parametrizing our workflows on the particular brand’s policies. By using the Amazon Connect agent workspace, agents can see what the customer reached out about, how the IVR responded, and previous customer contact history. This context helps agents answer customer questions in a more personalized way and reduces AHT of agent-assisted interactions. Agents also have real-time access to internal knowledge bases to answer customer questions and can view and respond to emails using Amazon Connect Tasks directly from the agent workspace.

For one of our brands, we started using Amazon Connect Customer Profiles and Cases inside the agent workspace to easily track and disposition incoming customer inquiries. Our agents have a full view of the customer and their short or long-running inquiries to better support requests such as checking order status, reporting damaged items, price adjustments, order cancellations and returns, and general product inquiries. We look forward to quickly enabling these capabilities in Amazon Connect for our other brands.

Unifying technology, the agent experience, and data seems to be a big theme for CSC Generation. How have you seen value in unifying customer service data, as well as margin reporting, across CSC Generation’s brands?

We‘ve begun storing data in an internal data management platform that operates as a schema registry partitioned by brand. For example, One Kings Lane and Sur La Table each have their order schemas registered with the data management platform. Generative AI is used to unify the disjoint schemas for the same logical type (e.g., order line items) into a standard format. By unifying this data, an individual brand can selectively access information from all brands under the CSC Generation umbrella so that a chatbot or an agent can tailor offers to customers based on their purchases at other brands.

By consolidating data from diverse sources, we have a clearer perspective on early customer interactions, allowing brands to identify which strategies and touch points resonate most with their audience. These insights are crucial for evaluating advertising effectiveness, understanding customer preferences, and for stakeholders interested in a thorough analysis of SKU profitability.

In addition, our call center operations have significantly improved by unifying data. This new approach allows us to address urgent issues more effectively, handle customer queries, and enhance service quality. Ultimately, this new approach will enable the transition from being just a cost center to a potential source of revenue generation.

What is the impact of these customer support back-end changes on customer loyalty and additional sales?

Our favorite example is with one of our furniture brands. We started a paid customer loyalty program for access to enhanced sales, early access to new products, and a cash back account. When interacting with the customer service team, customers who are deemed likely to convert for the loyalty program are made the offer and convert at mid-double digits for the paid program, and then those same customers have double the average lifetime value of customers in the same cohort.

Going forward, how does CSC Generation and its SnapBPO software intend to use Amazon Connect? What is the plan for expanding generative AI to more use cases?

We’re excited about the potential of generative AI and plan to expand our use of the technology with Amazon Connect. Our vision extends beyond using generative AI to suggest responses for agents in chat and voice channels; we aim to also use it to authenticate users by transitioning from traditional security questions like “Where were you born?” to dynamic, transaction-based inquiries such as “What did you purchase 3 days ago?” This change elevates the customer experience and tightens security by requiring answers tied to each individual’s unique, recent transactions. Embracing unstructured data is the key to gaining a deeper understanding of your customers.

We’re happy that we finally found a solution for our SnapBPO software and call center and customer service operations that is reflective of the CSC Generation mission to pursue excellence in and standardization of different operational domains within our portfolio companies.


Amazon Connect provides CSC Generation and its AI-powered SnapBPO software the ability to take advantage of the frequent releases and enhancements that Amazon Connect offers and to leverage AI and automation to not only improve customer service, but also to reduce costs and drive growth. By integrating self-service experiences and leveraging built-in capabilities, like AI, to help agents be more efficient, CSC Generation has been able to transform their customer service operations, making it more efficient and effective.

Learn how global businesses are using Amazon Connect to improve customer and agent experiences.

About CSC Generation Holdings, Inc. (CSC Generation)

CSC Generation is a multi-brand technology platform that acquires and transforms retailers into high performing, digital first consumer-centric businesses. CSC Generation stands out from the rest of the retail industry with their strong marketing intelligence, supply chain and distribution channels, outstanding customer support, and experience with brick-and-mortar stores and online and catalog sales.

Andrew Templeton

Andrew Templeton is a seasoned technical leader with P/L experience, specializing in the design and execution of innovative, AI-driven automation projects across various industries including construction, fleet logistics, and e-commerce. He has notably led initiatives in customer service automation and quality issue detection, and has developed advanced engines for source code refactoring in TypeScript, along with establishing dynamically scaling data integration platforms. His technical acumen is extensive, with expertise in TypeScript, Python, Ruby, and client-side frameworks like React and Vue, complemented by advanced certifications in Amazon Web Services.

Chris Phan

As a Technical Account Manager specializing in AWS technologies, Chris channels his passion for technology and frontend development into providing a dynamic blend of technical expertise and client support. He’s dedicated to helping customers navigate their unique challenges and achieve strategic successes through personalized guidance and continuous education in the ever-evolving realm of cloud-based web solutions.