How DISH scaled contact center and agent success with Amazon Connect

To make sure everybody is where they want to be. This is business 217, the breakout session. In this session, we will be discussing some key challenges that we will help you overcome with Amazon Connect within your contact center - the types of challenges that we anticipate and hear from customers often:

• Are you keeping up with the pace of innovation that's happening? How do we think about AI and the impact it can have on your contact center operations?

• Do you have challenges with scaling to meet your business needs?

• Are there individuals here that are looking to move from an on-prem solution to a cloud-based solution or maybe in the middle of it?

We will also cover challenges and how Amazon can help with training and retaining your agents. We will also talk about how we want to improve agent productivity so they can spend more time having deeper, more meaningful conversations with your customers.

My name is Steve Johnson. I am a Director and GM within Amazon Connect at AWS. Today we're going to walk through Connect and talk about how the aforementioned challenges can be overcome using Connect.

I'm excited to be joined by Dave Downing from Dish and Scott Stimson, the Principal Solutions Architect who has been working with Dish. They will come up later and tell you about their challenges in migrating to Connect and how they overcame those.

I recognize that as I look around the room, I see some familiar faces that are already Connect customers. I see partners here that I recognize as well. But we also have a good number of folks that are new to Connect, that are interested in learning more. So please bear with me for a few slides as I go through the benefits of Connect and give an overview before we dive deeper into the agent experience.

Amazon Connect is a single customer experience application offering one seamless experience for your entire team. At Amazon, we often get questions like "What software do you use to run your contact centers?" Back in the day, with the growth and scale Amazon had to move at, we built a lot ourselves. It wasn't until March 2017 that we began to take our industry knowledge and experience building our own contact centers to make it available as Amazon Connect.

Today there are over 50 different groups and subsidiaries within Amazon that leverage Amazon Connect to power contact centers. That includes subsidiaries like Zappos and Audible. Probably the most significant one is Amazon Retail - today Amazon Retail has over 100 contact centers globally with over 100,000 agents using Connect.

Let's review some of the key differentiators of Connect that people evaluate when considering using Connect:

• It is very easy to set up and configure - literally set up in minutes and agents can start taking calls
• Simple to administer - administrators can immediately start innovating with the graphical UI
• Great omni-channel solution across voice, chat, messaging, tasks - configured and managed in a single solution
• Built from the ground up with AI/ML to deliver real-world AI-powered results
• Enables you to measure, track, forecast and achieve business goals with real-time AI-powered analytics
• Scales quickly to meet cyclical business needs - ramp up and down quickly

In the first few years, Connect was seen as a disruptor - a new product with massive scale potential. Over the years we continued to expand and invest at a rapid pace - 95% of features still come directly from customer feedback that shapes our roadmap. We went from a disruptor to a visionary, and earlier this year were named a leader in the contact center space. While proud of that, we feel we're just getting started.

Customers have responded - here are just some that handle over 10 million contact center interactions per day. You'll hear from Dish today about their journey. Our strategy has been to innovate and launch improvements that address real-world problems and challenges customers face.

Let's look at some of the core capabilities of Connect. We provide a seamless experience across users, channels and data. We build Connect with four primary personas in mind, working backwards from the customer:

• Delight customers and save them time through great omni-channel experiences
• Empower agents to deliver great experiences with the right tools and data
• Enable supervisors and managers to focus on key areas needing their attention
• Help IT manage operations efficiently while innovating rapidly in a secure, scalable environment

Today we'll focus on the agent experience and how we reduce complexity to increase innovation speed. It starts with a unified agent application to reduce training time, burnout and cognitive load. We analyze interactions in real time and provide knowledge to reduce cycle time. Customer profiles deliver a unified view of data agents need. Case management brings contacts and info together. Step-by-step guides provide visual, clickable workflows to guide agents.

Many customers have shared success stories of efficiency gains with these capabilities. We provide options to meet you where you are - native capabilities out of the box, but also integrations with partners or custom builds.

Customer expectations rise every year - are you keeping pace with innovation? What do you want agents remembered for, and how can they improve? It's amazing what agents handle - navigating screens, finding solutions, trying to understand intent while multitasking. The cognitive load puts pressure on them. Experienced agents face repetitive contacts and lack upskilling opportunities. Onboarding new agents takes longer, causing frustration. Inconsistency results.

And all of this leads to again, higher agent turnover, it leads to inconsistent customer experience. So what do we suggest that uh how can i help our goal within the aging experience persona? And we, we, we we have agents become superhuman problem solvers. It's very aspirational. But first, what we try and do is we want to make sure that we provide agents with a unified workspace for experienced agents or new agents, the ability to, to learn and to use all of the tools in a single place out of the box. This unified experience provides a single place to find offers actions, answers doesn't require them to navigate all the different tools and applications.

How about using gen a i and how we can support, support those agents and mister click? But using genai, we'll talk a little bit shortly about amazon q that was announced and how we leveraged that within connect. But you can assist agents by automatically detecting the customer's needs, the customer's intent and you can deliver generated responses and generated actions for the agent to take with our step by step guides and third party application support and things we announced this week. You can bring together all of your work flows across the spirit applications and guide agents into an easy customer, uh easy to identify customer issues and recommend subs subsequent actions.

One of the things that's most powerful about this is when they're all used together. As i said earlier, we do our architecture and our approach is to make these composable in a way that you can use these individually and to meet you where you're at. But the power really comes to when these are all used together. So let's take a quick demo of what this looks like when it's all used together.

Nikki is traveling internationally and while using her travel card, it gets declined, she opens up her voyager travel mobile app and immediately starts a chat to see if she can increase her credit limit during her conversation and connects self service experience. It offers that she's a good fit to upgrade her travel card. She immediately asks to chat with an agent. Our agent John is able to easily preview the customer and case information for Nikki amazon. Q detects Nicki's question from the chat and offers strong the response and solution with specific information about the travel rewards card.

John is able to easily see how amazon q sources the answer from the enterprise knowledge base and he is able to paste the generated response from amazon q into the chat with one click. Nikki indicates that she would like to apply for the new travelers rewards card. And john sees the next best action for the credit card offer and is taken through a series of guided steps to apply for the credit card. The credit card application is successful and john has presented the information about the approval from guides and shares the news with nikki.

John wants to make sure nikki does not get declined for any transactions while traveling in france launching their homegrown credit card travel portal. He navigates to the application from the agent workspace and adds her travel information. The chat ends with nikki's issue being successfully resolved and john wraps up the call.

Now. True story. I came to find the location of the room about an hour ago, an hour before the session started and i was standing just outside these doors and i hope the individual's not in this room or if they are they they can remain anonymous. But there literally was a person having a conversation with their bank about their credit card being declined here in vegas this week. Didn't sound like it was going well ah as they as they did, but uh i digress but let's talk a little bit about the features that we kind of highlighted in that presentation.

There are three key announcements that we kind of made this week. Obviously, uh amazon q inside connect, we talked about how we can generate guides faster, the step by step guides that we jed earlier this year and then how we can unify your agent application or agent workspace by bringing in three p applications.

Amazon q, the broader sense of the definition that hopefully you heard from adams keynote ah earlier this week is that amazon q is a new generative a i powered assistant that is specifically for work and can be tailored to your business and have conversations and solve problems, generate content, take actions using the data and expertise found in your own company's information repositories, code base and enterprise solutions. And we were excited to announce that we had been working on something very similarly for a while and amazon q became part of connect. And what does that mean for connect? And what exactly does it do with this initial launch that we announce you might be familiar with the existing uh connect product called connect wisdom, which was an ml based uh recommendation, real time recommendation for knowledge based articles that is now amazon q and we are focused on generating real time responses, real time s actions, the agent should take real time solutions and all we see continue to reference those data sources that they came from to the agent need him. We take your data and we synthesize to generate these responses, synthesize all of your data to generate these responses.

We saw that in the demo kind of on the right hand pane of the glass or on the right hand side of the of the agent application. What you can see here is first we worked, we used generational a i to detect the issue, the customer issue. Second, we generate a response using amazon bedrock and large language models. Then we generate the, the next steps that we think the agent should take. And finally, we always make sure the agent is aware of if they needed to, to go find the source, if they needed to. We do this in real time through voice as well as chat and while the conversation is going on with your customer, amazon q will continue to provide those recommendations, but it pushes the previous recommendations down if they need to refer to them later, they're also able to manually search, should they wish to search in the box? They can find a particular response as well.

We put this in beta at the end of july and we had a large number of customers participate in the beta and give us some wonderful feedback. Um we were excited by some of the responses that we have seen. There's certainly opportunities that we will continue to improve upon. But with it, they were seeing some results with regards to the time savings, it took not only their new agents but also their experienced agents. I like the for, for those of you again, i mentioned orbits earlier as a manufacturing uh manufacturer of uh irrigation systems, trager does the trager smoker and grills. Um and you can see that one of the things that we're hearing from customers more is less and less about a ht savings and more and more about how can i repurpose that a ht savings and actually provide a more premium customer experience. How can i leverage that? 10% 20% whatever the magic number is a ht savings and use that as an opportunity to not only resolve their issue but also potentially upsell them on a different offer or a different, a different thing. So we're seeing some of that within our uh customers as well.

Amazon q within connect currently available in these regions and currently free of charge until march 1st. So we ask that you take the opportunity to explore, experiment yourself uh and share your wonderful feedback because i'm sure you will on what we can do better, how we can improve and what's working. Well, as i mentioned earlier this year, we uh we g a step by step guides. I'm gonna, it's a quick refresher. Again, guides are used to enable you to quickly match customer issues. Do guided flows, take those steps, you can really standardize work across your agents, you drive consistency, easier to onboard, easier to train, easier to upskill. Uh as you go through the guides. And one of the pieces of feedback we got from customers was these were created through adding a new flow block within contact uh connect flows and it required some json to be added. And feedback was that's a little, little too much.

And so as i highlight here, the guide steps that we have, but our launch here was really specifically around our new no code u i guide builder. And what you can do now is you can simply drag and drop, create your own things. It doesn't have to be a technical person. It can be your contact center manager, it can be a supervisor, they can begin to create the workflows, they can begin to edit them. We all learn, we learn that the 10 step process maybe should, you know, maybe i forgot something, maybe i need to break something up into multiple steps. Now, you can iterate more quickly and you can begin to push these out to all of your agents. So that was the launch we had earlier today.

We have seen some positive feedback from customers as well. Again, you can see some improvements in productivity with experience. It's a health care solutions provider, helping hospitals and hospital groups and physicians. They rolled out guides in less than 90 days for over 100 agents and they saw productivity improved by 15% earlier this year.

You may have heard about an announcement within amazon retail about buy with prime that as a prime member, you can purchase on other retail websites if you will other market or other merchants websites. One of the benefits that they have is if you're a prime member, you can call and get customer support. Well, you can imagine if you purchase from different vent different merchants, our agents have to understand the rules, the policies of each of the individual merchants. They leverage guides to do that reduces the cognitive load for those agents allows them to contact, switch from one merchant to another merchant to another merchant. As those contacts roll in and agent workspace and guides was a key part of that success.

I asked earlier about folks that have spent some time going through with some agents and sitting with them and i'll quickly go through the this slide here where you sit with them if you've ever been called and they say please hold while i find that information and then you realize what's happening behind the scenes. Is there there there there's switching between tabs. They're like going through everything, finding order history, looking at crm data. What can i find about you? One of those hit? I wasn't even sure what that was that hid behind there. All these different systems that we asked the agent to go through.

Now there was a survey done a while back on average 8.2 i meet very few customers who raise their hand and say yeah, i'm less than eight. The value usually a customer that i meet are much higher than eight. That is an enormous amount of cognitive load for an agent. A ma burden put on training new agents and getting them up to speed. So by launching three p support within our third-party apps, you can now add in and indirectly into our agent workspace. Third party apps. We announced this kind of quietly a few weeks ago, we already have hundreds of customers have already integrated hundreds of applications and continue to experiment with that.

Here's been some positive feedback we've seen from some customers. And lastly, uh you can see some built one in five minutes around kind of aging analytics, things that they can do and bring that forward. And lastly, i'm gonna share another success story before we talk. Uh i'll transition over to dh as my time is up. But uh um unum helps provide affordable health care and access to disability life, accidental insurance, uh et cetera um through their workplace. And when they migrated over to connect and they migrated over to agent workspace and leveraging guides. They were able to deprecate their old custom desktop that they used to have and that they had previously built. And what were the results as you can see, connect enabled them to significantly decrease their agents on boarding time, reducing custom development. So those resources can be applied to elsewhere in the organization and overall it and contact center costs reduce.

So hopefully this background and understanding of kind of how connect can help you solve. Some of those challenges came through. Um with that uh since my time is up, i'm gonna turn it over and uh we'll hear dish's journey and how uh how dish handled this. Please welcome uh dave and scott.

Though we offer a variety of customer service options from 24/7 phone and chat support to web and mobile self-serve to in-home installation and service, just a number of different channels for our customers to interact with, what's most important is not what we provide, it's why we provide it.

We're committed to providing our customers with the best possible experience, whether it be viewing television, automating your home, talking to loved ones on your mobile phone or interacting with our call center agents and in home technicians, we strive to make your experience memorable. And I'm excited to say as a result, Dish Network has been ranked number one in customer satisfaction among cable and satellite television providers nationally for six consecutive years by J.D. Power.

This award is a testament to our commitment to providing our customers with the best possible experience, but we couldn't do it without some great technology like AWS. So I'm excited to share our story about a to Zs connect today with you. But I'd, I'd like to take a step back and, and start with why and it comes down to customer experience.

CX is more important now than ever. Customers expect to be able to interact with us in a variety of different channels, whether it be phone, mobile, email, chat, social media, you name it and we in turn have to respond and be able to provide a consistent and positive experience across all channels.

Although there's a lot to unpack there, we were looking for specific ways to do the following:

1. Scale our systems to provide a consistent experience regardless of load, personalize our customers, interaction with us across multiple channels, providing contextual awareness

2. Increasing our rate of delivery so we could test and learn quickly

3. Doing all this while efficiently managing our cost

And call center optimization was a key component of this. Now, we saw a number of opportunities within that space. We looked we were looking to leverage AI to automate tasks such as call routing and customer service inquiries so that our agents could actually focus on conversations and building that relationship with our customers.

We were looking to use cloud based call center solutions to provide more scalability and flexibility versus our on prem solutions. We're looking to implement omnichannel customer support to allow customers to contact us through a variety of channels and receive consistent and personalized experience.

We wanted to speed up our time to market allowing us to quickly and inexpensively test our theories before we dive in and we wanted to collect and analyze customer feedback to determine what's working, what's not and develop insights into what we could be doing better.

Finally, we wanted to move to a pay-go model. We wanted to get out of the business of paying those high watermark licensing costs. So we partnered with AWS and focused on the Connect platform.

So how exactly did we complete a 15,000 agent migration across eight organizations following a sixth JD Power award? It was no easy feat at Amazon. Every idea starts with a working backwards process where we envision our end state and we work backwards to achieve it. The working backwards process is how we ensure customer obsession, maintaining Dish's high bar of customer experience while envisioning a new solution and clarifying the benefits of the idea, continually evaluating the approach throughout and iterating if and when needed.

Our team set an ambitious goal, working backwards to migrate to Amazon Connects cloud based contact center platform on a condensed timeline to achieve this goal. We took a one team approach where AWS and Dish came together formed cohesive teams across lines of business and organizational units.

We started with a pilot or a proof of concept, migrating a small number of agents on an internal use case to test the technology. Next, we held interactive demos and workshops, learning sessions allowing teams to be comfortable with the technology. A key benefit highlighted during this was the API driven integrations where surround applications that are needed to support the contact center integrate natively into Connect.

We took a data driven approach using the rich metadata and analytics and reporting capabilities powered by Amazon Connect to optimize and enhance the agent and customer experience throughout the design cycle. Following our well-architected framework, we built a secure, cost performant and operationally excellent contact center in the cloud.

In the end with careful planning and collaboration across organizations, we successfully migrated 15,000 agents to Amazon Connect within two weeks. The enhanced flexibility, scalability and customer experience have exceeded our expectations.

Let's dive into how that's possible. About this time last year we were in planning phase. So stepping back a bit, we'll dive into our internal use case here in a couple of short slides. But in business speed matters, we started with our internal pilot utilizing a small agent pool to test and validate features, theories the enterprise contact center have come to expect to deliver on powerful customer experiences.

From here, we took additional use cases and brought them to life, validating additional features such as agent bargin and within a few short months, the evaluation was completed. Our initial estimates to complete this migration was approximately 12 plus months to ensure all feature parity is there.

However, multiple factors encouraged our customer to challenge our teams to do it on a shorter timeline. Planning to support the size of contact center required the flexible capacity a cloud based contact center such as Amazon Connect provides. The ability to scale up, scale down and scale in when needed without having to pay agent seats was extremely valuable.

Our teams engaged in discovery workshops, ensuring and diving deeper into all necessary integrations, workflows and agent customer use cases to ensure the migration plan was feasible. And through the amazing work of our technical team, some of who are in the audience today, we were able to successfully migrate 6000 agents in a single go in one day.

6000 agents went live on Amazon Connect and 15,000 agents within two weeks. This was a monumental achievement and a true testament to the hard work and dedication of all of our teams involved and the results speak for themselves.

Post migration, Dish is now able to handle 25% more calls per hour. In fact, on the day after go live, the contact center handled two times the concurrent call volume.

Let's take a moment to step back to where we started. The initial pilot focused on solving for agents that were no longer working in centralized office locations where previously call center leaders lacked visibility into specific agent issues such as power, hardware headset, internet or larger environmental and regional issues.

Throughout a shift, our goal for the pilot was to enable agents to self report and/or self serve by reducing agent and employee conversations by 25%. On top of the rich data provided with Amazon Connect, we built dashboards allowing the workforce team to make decisions based on this data in real time. And we were able to exceed our initial goal by achieving an average decrease of 60% of live calls since the launch, driving major efficiency improvements for contact center operations teams.

The leadership team appreciated the flexibility of pilot changes so much that they didn't want to let it go. So the pilot was expanded into a full migration.

In the first month on AWS, we were able to take seven times the call volume that we were scaled for in our on-prem solutions. That's pretty impressive. We're now also in a position to be able to take more volume as our business grows without a significant investment in infrastructure.

AWS releases regular features with no downtime which allows us to pick and choose what we enable and when and deliver on our schedule. The platform makes tying call routing, IVR, call recording, transcription and redaction relatively seamless. Instead of investing time in integration and troubleshooting, we could spend time developing capabilities.

We were able to speed up the onboarding of new agents cutting about a day's worth of work out of our workforce management team. We saw a reduction in the abandoned rate resulting in more customers being helped. We also saw a reduction in after call work with the improved capabilities to handle after call work Connect provided us.

Connect provided us with multiple ways of routing, some of which we haven't implemented yet, so we've got complete flexibility there to be able to figure out how to route our calls. Our business partners were able to make changes like messaging much quicker without engaging IT.

We saw reduced infrastructure administration - the stuff that we used to spend a lot of time on before like patching, upgrades, maintenance was now being handled for us. We began to pay for what we used instead of a named or concurrent usage model, which meant we could manage our costs a lot more effectively.

And finally, AWS has an extensive catalog of services already available that we can tap into in the marketplace, which allows us to seamlessly integrate, innovate, develop and deploy new capabilities.

So we've had a lot of great outcomes, but we feel like we're only at the beginning. I'm excited to share with you what's next for our contact center. As we focus on improving the agent experience, we want to equip agents with features that allow them to resolve customer issues quickly and accurately on the first contact.

So we're looking at things like integrating with knowledge bases using the AI/ML or Q capabilities for easy answer lookup and meaningful recommendations to the agent screen, pops with customer history and account details, really surfacing relevant and meaningful data to make the agent's job easier and finally embedding our own custom CRM into one screen using Agent Workspace to drive efficiencies and navigation across the customer's information.

By streamlining these interactions across these resources, we aim to empower agents to set them up for success.

We also want to improve the customer experience by focusing on our north star, a true omnichannel experience to give the customer the option to switch between channels seamlessly while carrying context leveraging AI to better serve them through self-serve. And most importantly, moving more towards a predictive support model so they don't have to contact us.

We also want to improve efficiency by further integration of our account management tools and CRM with the Connect platform and Agent Workspace, feeding back data from actual experiences like screenshots, transcriptions, usage to drive insights and improve the next experience.

We're all in the same situation as many of you - improving customer experience is a never ending journey. We don't always get it right either. But having a partner like AWS that provides technology that can help us test and learn, implement and scale to handle volume quickly, definitely improves our odds.

Thanks for your time today and thank you to the Dish team for sharing their journey. I hope you found this session valuable. There are additional breakout sessions on relevant topics that I encourage you to attend. And please join us at the networking reception this evening to continue the conversation.