Kazuya

Posted on Dec 5

AWS re:Invent 2025 - AWS & GenAI Power CDG's Telecom Software Modernization Journey (SMB202)

🦄 Making great presentations more accessible.
This project aims to enhances multilingual accessibility and discoverability while maintaining the integrity of original content. Detailed transcriptions and keyframes preserve the nuances and technical insights that make each session compelling.

Overview

📖 AWS re:Invent 2025 - AWS & GenAI Power CDG's Telecom Software Modernization Journey (SMB202)

In this video, Tony Stout, CTO of CDG, shares how his 55-year-old telecommunications billing company transformed from legacy data centers to a cloud-native, AI-native platform using AWS. CDG supports ISPs serving rural America to close the digital divide affecting 25-40 million Americans without broadband. Over 3.5 years, they leveraged AWS services including Bedrock, Connect, QuickSight, and Claude v4 to build agentic AI solutions. Results include doubling sales year-over-year for three years, 58% revenue growth, and attracting clients 5-10 times larger. Their AI innovations deliver 24/7 customer support chatbots, route optimization saving 25-40% fuel costs, automated scheduling increasing capacity 400%, and virtual agents for onboarding and troubleshooting. Tony emphasizes strategic planning, leveraging AWS partner network, fostering innovation culture, and focusing on customer-centric outcomes as key lessons learned.

; This article is entirely auto-generated while preserving the original presentation content as much as possible. Please note that there may be typos or inaccuracies.

Main Part

Closing the Digital Divide: CDG's Mission and Transformation Challenge

Good afternoon everyone. Did you know that one out of eight Americans still lack access to internet at home? Think about the impact of that. It's not around social media necessarily, but having access to telemedicine, working from home, and so on. We at Amazon have a set of values that we follow. We call them leadership principles. One of them is scale and success bring broad responsibility. This principle is really going to be the theme of the story today around social contribution, and I'm really pleased to introduce Tony Stout, CTO of CDG, which exemplifies this value.

So Tony, without any further ado? Thank you, Frank, I appreciate that. Well, good afternoon everyone. Today I'm going to tell you a digital transformation story about how a mid-55 year old Midwestern company is transitioning from a legacy scenario to an innovative disruptor and how we're partnering with our clients to help close the digital divide in rural America.

As Frank mentioned, I am Tony Stout. I am the CTO for CDG and CDG is an ISV that builds billing, subscriber management, and operational support systems for the telecommunications industry. We're headquartered in Champaign, Illinois, and our clients include small independent service providers that serve small town rural America and smaller suburban markets. We also provide carrier billing services to the largest service providers in America. So we have experience working with clients with a few thousand subscribers up to clients with many millions of subscribers.

First, I want to point out one of the key challenges facing our industry and really our country. Frank mentioned it earlier: there are 25 to 40 million Americans that lack access to fixed broadband. We know these underserved communities are falling behind both educationally and economically, and they deserve access to fast and reliable broadband. Our clients are building fiber and fixed wireless networks to these underserved communities, and CDG is supporting our clients by delivering industry-leading solutions that include subscriber acquisition, customer management, and workforce management, billing, ticketing, and many other services that form the back office system that these companies need to run their business.

We're also adding generative AI capabilities to help transform these often smaller clients so they can compete with their larger competitors. Three and a half years ago, CDG was also facing some major challenges. Our applications were running in legacy data centers on legacy infrastructure, and we wanted to modernize, scale, and transform our organization so we could be an innovation leader and move upmarket and sign larger clients and also obviously grow revenues.

We also needed to modernize our tech stack. That included eliminating technical debt, training our employees, and changing our culture to be more agile and experimental. We also needed to plan for the future and build a next generation cloud-native AI-native platform that could serve our clients for the next 15 to 20 years. We were faced with the question: how do we do this in a reasonable time frame and at a reasonable cost?

Partnering with AWS: Building a Next-Generation Cloud-Native Platform

So what did we do? Well, we partnered with AWS and we leaned into this partnership and leveraged a lot of their services. We leveraged the partner network to help us with things like architecture and implementation. We didn't have that expertise on staff. We also leveraged MAP funding to help accelerate our cloud adoption. Is anyone using MAP funding at all in the audience? We're addicted to MAP funding at CDG and we also utilize the enterprise support and incident detection and response teams to help our teams provide support for our production workloads as they improve their expertise via the AWS training and certification programs.

We also worked with the AWS telecom teams to learn more about emerging projects across the globe, and we leveraged the culture of innovation workshops which really educated and challenged our teams to think differently about the way we build and design software to become more agile, fail fast, and most importantly, to become customer obsessed. Most recently we have leveraged AWS Bedrock to be the foundation of our GenAI strategy. This has allowed us to become an innovative leader in AI innovation. We launched an agentic AI solution earlier this year built on Claude V4, and now we're working on agent-to-agent collaboration within our platform, and next year we'll be building A2A integrations with third-party partners.

As I'm sure you all will agree, we're just scratching the surface on use cases and problems that we can solve with AI technologies. CDG's plan is to build a network of agents with orchestration capabilities that will allow us to build virtual employees that can assist their human counterparts and eventually complete many mundane tasks autonomously, allowing employees to focus on higher value tasks.

Over the last 3.5 years, we've made significant progress migrating our applications into AWS and moving our clients there as well. We are well on our way to building this next generation platform that we call CDG Elements. We began our environment design and implementation in late 2022 and started client migrations in 2024. In Q1 of this year, we've completed the migration of our OSS BSS application and all of the clients using that application. We also launched our Elements platform this year and over the next 2 to 3 years we'll continue to enhance that next generation platform.

Next year we're going to be migrating a carrier billing application that actually runs on a mainframe. We're going to be moving that into AWS and moving those clients into the cloud over the next few years. By 2028, we anticipate fully decommissioning our data centers.

Growth Through Innovation: AI-Powered Solutions Delivering Real Results

So what are the results of all these activities? In a word, growth. We have doubled sales year over year for the last 3 years and grown top line revenue by up to 58 percent. Our newer clients are 5 to 10 times larger than our previous clients. By leveraging the scalability of AWS, we're able to attract larger clients. By leveraging the multi-region capabilities of AWS, we can provide high availability as our clients run 24/7 operations.

From an innovation standpoint, we're also leading our industry in cloud and AI adoption, and we are delivering industry leading AI solutions to our clients. A few examples include launching our first AI chatbot as I mentioned earlier. These virtual agents are persona based and trained with highly customized content, allowing them to become a subject matter expert in a particular area of our client's organization.

We're also now leveraging AWS Connect to add voice capability to these virtual agents to support outbound calling for payment reminders and scheduling reminders. In 2026, we will be evaluating Amazon Nova Sonic as our voicebot solution to support interactive two-way voice conversations.

We're also leveraging the Agent Core platform for building an onboarding virtual agent. It takes us about 12 months to onboard a new client at CDG. It's very labor intensive, and our goal is to reduce that by as much as 50 percent. We're leveraging this virtual assistant to help our clients self onboard onto our platform by providing helpful information about the process, links to interactive training videos and other services to help our clients work through the onboarding process. Once the client is satisfied with the configuration, these virtual agents can even configure our application for the client with the ultimate goal of radically streamlining the process.

Another initiative we're working on is AIOs. Working with our network equipment vendors, we're able to leverage real-time telemetry data from the network to identify trends and anomalies that might be occurring that could lead to service disruptions for their subscribers. Our goal is to surface these issues via automated tickets so our clients can get out in front of these network issues and resolve them before their clients are impacted. Our eventual goal is to send these tickets to virtual agents that can triage and troubleshoot these issues and resolve them before escalating to a human.

Another area we're expanding is in our data lake house and business intelligence solution. We launched with AWS QuickSight this year and next year we'll be adding QuickSight Q to support natural language processes with automated report generation.

I have a brief demo I want to share with you that shows some of these tools in action. The subscriber logs into CDG subscriber. The CSR agent receives a notification in the chat queue.

The AI chatbot understands the subscriber's sentiment and suggests appropriate replies to the CSR agent. The CDGQA agent delivers complete end-to-end test management. The agent performs the following tasks: analyzes requirements, generates test cases, and loads these test cases in BrowserStack. It initiates automated QA and captures results, creates Jira tickets for any defects, and reports them in BrowserStack.

The AI agent LLM utilizes the CDGMCP server tool to retrieve serviceable products through the Forum API. These are a few of the tools that we have been building over the last year leveraging AWS services. I want to talk now about some of the results of these initiatives.

One of our clients in Indiana had the objective of doubling the number of installations that they could schedule in a given month without adding additional scheduling agents. They leveraged our automated scheduling processes and were able to double and eventually increase scheduling by 400% without adding additional employees. In many cases, the subscriber is scheduling the orders themselves as part of the workflow process of purchasing the broadband service.

We also launched an AI route optimization solution as part of the AWS Partner Network. That solution evaluates our clients' daily jobs, looks at the calendar, sees the jobs scheduled for the day, knows that each job has a specified start time, and knows the starting location of every technician. It schedules the jobs and assigns them to various technicians to optimize routes and fuel consumption. In many cases, this is driving 25 to 40 percent savings in fuel consumption and drive time, allowing our clients to add more jobs onto the daily calendar without adding additional resources.

Our new AI chatbot, based on Claude v4, is providing 24/7 by 365 customer support, which reduces call center volume and allows our clients to focus their call center resources on higher value tasks like outbound selling and selling faster broadband packages. We are also leveraging these AI agents to assist the commercial sales teams of our clients in building bills of materials for commercial sales proposals as well as performing troubleshooting steps for tier one support agents with recommended next actions based on historical trends.

We can also transcribe call center recordings, rate call center agents, and perform sentiment analysis on the conversations. Additionally, our open architecture allows us to create real-time integrations with the network equipment vendors so we can automate provisioning. This allows our clients to provision new clients without manual provisioning and switching between multiple systems.

Looking Ahead: Roadmap for 2026 and Lessons Learned from Digital Transformation

As we think about 2026, which is right around the corner, what are some of the new initiatives on our roadmap? We are going to be launching a QA virtual agent that can perform code reviews and run automated test scenarios against our source code. We also intend to improve our onboarding virtual agent as we look to further drive down onboarding times and efficiency. We will be implementing a network of agents solution next year with a number of POCs planned for agent-to-agent integration with third-party partners.

Regarding customer impact, our solutions are really starting to impact our clients. Our top 10 clients are adding tens of thousands of subscribers every month, which is impressive considering these are rural markets and in many cases competitive markets where they are competing against much larger competitors. This growth is allowing them to focus their energy on expanding their network, which is helping to close the digital divide we talked about before.

We are also expanding our new Elements platform next year to include a new visual workflow builder. We are launching a new point of sale solution and a new platform automation designer that allows our clients to create automated actions based on certain conditions.

Additionally, we're integrating AWS Kiro into our development teams to further accelerate our developers' efficiency and reduce time to market for new features. So as you can see, we've been very busy over the last couple of years and we've learned a lot. We've had a lot of success, but we've also made a lot of mistakes and I'd like to share some of those with you.

As you're thinking about a large digital transformation, I would recommend you plan strategically and start with clear business objectives. At the end of the day we're trying to solve business problems. It's easy to get lost in the technology, but if you can take a large complex project and break it down into small phases, it really helps with better control and reduced risk. Also leverage partnerships. The AWS partner network is full of partners that can help with just about anything you might need, whether it's architecture, cost optimization, or security. That really helps accelerate success.

Don't forget about the enterprise support team either. That really helped our team provide almost immediate support on our production workloads. At CDG we're trying to foster a culture of innovation. We're an older company. Our employees weren't necessarily on board day one, but they're starting to adopt new technologies and embrace some of the changes that we're pushing and driving, including building and executing a solid plan around generative AI.

I believe that the successful companies of the next two to three years will be those companies that have a solid plan and they execute well on generative AI, and that really will drive a competitive advantage. Lastly, focus on impact. You start with a clear problem definition. Measure your outcomes. Take a customer-centric approach to make sure you're focused on what matters most, which is the customer, and that will help drive real business value to your organization and to your customer's business.

I hope these lessons learned and this transformational story can assist you in your journey. Thanks for your attention. If you have any questions, I'll be to the side of the stage for the next 30 minutes or so. Stop by and say hello.