CCW 2020: Rethink Your Contact Center With Amazon Connect

hello and welcome i'm pasquale demaio the general manager for amazon connect at aws i wish i was on stage with all of you at ccw but unfortunately things have gone a little differently this year perhaps you're a parent maybe you're engaged in digital learning or just trying to get through remote work i'm sure your life has changed dramatically as has all of ours perhaps also you're a contact center leader and you've had to deal with a lot of challenges at work potentially the transition of agents to work from home for example but with change comes opportunity and i hope you'll come with me on a journey as we see how folks in the contact center are learning to get by to change the way they think about things and actually then reimagine their businesses we've seen customers really go through three stages so as we move forward we are seeing that our customers who run these contact centers who are managing these challenges first just had to react they had to get their agents at home and they had to get them working quickly and then once they were there they wanted them to return to more of a semblance of being able to use the tools the skills they developed but it's not just getting back to business as usual obviously that's not going to happen for a while but in these cases where things change radically we do want to also take the opportunity to reimagine and customers are telling us that that's something that this is allowing them to do and we're going to walk through all the different ways that folks are changing the way they think about customer service and of course it wouldn't be complete if i didn't finish with a demo and for that i'll bring up my colleague yasser alhagan and he'll show you a little bit about how people are doing this when the pandemic started folks were scrambling aws got a number of inquiries about how we could help them migrate their contact centers customers contact centers were built on infrastructure that was difficult to enable and was hard to move up and down in terms of scale and just challenging to offer the opportunities to work remotely they were really stuck they needed to get their agents untethered and working from home efficiently quickly and safely aws and amazon connect helped their customers move their contact centers their operations often overnight it helped them both on board new agents and to scale their systems this allowed them to respond in ways and react faster than they could have on their own so for those not familiar with amazon connect maybe i'll just start and tell you a little bit about the basics it includes everything you would expect from a contact center skills based routing great voice and chat technology really high quality sound with the opus codec which makes it work great on the internet and then it also allows you to do call recordings and chat transcriptions for example the west virginia office of technology had a mandatory shutdown traffic was spiking for them their call center was inundated by the second week it was clear they weren't going to be able to keep up with what was happening their call of center receiving 77 000 calls it was overwhelming as people look to get unemployment insurance the phone systems they used were just not up to it but amazon connect was able to help them get going and 72 hours later they were seamlessly taking calls with no concerns with scale uk post office saw their use of skyrocket by 37 percent to 4 000 contacts a week customers need a way to communicate more effectively with them they took a look at amazon connect chat and saw they could use this to take on a lot of the communications that weren't being effectively handled on voice in just three weeks they were up and running and then after that they continued to innovate which allowed them to not just provide a quick solution for the customers but one that got better and better over time and finally i'd like to talk a little bit about one that i think is always close to my heart and that's food when margaritas mexican restaurant had to start suddenly taking tons of orders they just weren't expecting this and that resulted in heavily overwhelming their current phone system they were getting business signals and dropped calls by using amazon connect they were able to not only quickly get up to speed with something that had no scale problems but they also integrated quickly with their point of sale and other tools just one day of deployment that's an amazing outcome for folks and it's also great to see everyone got their mexican food these easy to use applications decrease the amount of effort it takes for agents to onboard and then increase the value to your end customer you may also be curious about how we reacted at amazon we get that question a lot we started by committing nearly four billion dollars in covet related initiatives we also reacted immediately for our employees with 800 000 of them worldwide it was incredibly important we support them with more than 150 process updates applying providing them with protective gear and building scalable testing coronavirus which is what we're working on right now early on on the covet crisis our customer service agents were told they could work from home we knew that was going to be incredibly important for their success and also for them to be able to deliver great customer service to our customers and with the size of the amazon businesses you can imagine we have a lot of customer service agents but it's not just about our employees it's also about the communities and we had to think carefully about that and one of the things we did was deliver 8 000 laptops to help seattle schools we provided food banks in 25 cities to help them serve 6 million meals and we hired over 170 000 folks to help how do we react for our customers well we're giving customers free mass at whole foods we've opened an hour earlier to help folks who are older be able to come in and feel safer and have a better environment and we're vigorously combating price gouging on amazon.com now that we've moved past reacting it's hard to think about returning and we want your employees to turn to the office in a way that's both effective but also safe there are some organizations that are shifting to complete work from home strategy and others are taking a hybrid approach gardner's research suggests that 23 percent of companies will keep 20 of their workforce at home and that's a big change for folks it's going to be a lot of new sets of challenges for those folks working at home and also of course the folks who come back who need to feel safe how much room do they need between each other how do they keep track of that what supplies do they have so that they can stay hygienic and avoid any risks that are unnecessary at amazon the size and scale of our portfolio means that return to work is a complex task with 50 countries 400 offices 4 624 conference rooms it's daunting but it's something we're committed to we're addressing global supplies cleaning systems self-inspection audit processes and for our employees safety has multiple meetings of course the buildings need to be clean but there's more to it than just that we're protecting our people preparing the place and we're implementing process for them we've put protocols in place that enable thermal scanning provide hand sanitizer face coverings gloves we're also providing new signage so customers employees know what to do when they come to the building and how they should behave to keep each other safe and of course cleaning supplies everywhere easily accessible the employee experience starts before they ever enter the building now we're going to talk about how our customers are reimagining the customer service contact center customers are vocal about what they need they've been telling us they need to quickly stay agile take advantage of software as a service change based on their customers changing needs maintain a flexible infrastructure and demand that expands across many different contacts let's talk about reimagining intelligent automation with unforeseen spikes and custom demand and a limited number of agents it's key to house self-service automation using self-service automation for workflows has been practiced for a long time mostly offering frustrated interactive experiences where things were pre-defined our customers are telling us they need something better something to help move beyond static prompts and limiting workflows one of the best things a contact center can do is use high quality machine learning to improve their experience for example amazon poly which can turn text into lifelike speech amazon poly text-to-speech technology uses advanced deep learning poly is available in 29 languages for example spanish and 59 voices like penelope and ricardo and we recently introduced neural text to speech so things sound even better today we have natural sounding conversational speaking styles and to really show you how this works i'd love to bring yasir on thanks pasquale my name is yasser again and i lead our global solutions architecture team for business productivity applications as businesses think of new ways to interact with their customers they've realized the power of using data to create engaging and personalized experiences and a great text-to-speech engine helps bring that to life through our continued investment in machine learning amazon connect now offers a conversational voice style that anyone can use to create dynamic and personalized greetings simply by typing text into a text field our neural text-to-speech engine combined with our conversational voice style turns text into life-like speech that sounds natural and engaging i'll show you an example first with the standard text-to-speech then the conversational voice style i'm matthew one of the english voices you can use for text to speech and amazon connect this is my standard voice it's okay i guess but now i'd like you to hear how i sound with my conversational voice style i'm matthew one of the english voices you can use for text to speech in amazon connect this is my new voice with the conversational voice style i'm excited to help you create personalized experiences that delight your customers amazon connect shortens the cycle of creating self-service experiences from months to days or weeks using the contact flow editor in amazon connect you can execute scripts to interact with back office data and easily add an existing amazon likes chat bot to create an ai powered customer experience here's an example of what's possible hi nikki wolf i apologize that your flight was cancelled i can rebook you for the next available flight departing at 10 pm would you like to book that now no i'd rather rebook for the same time tomorrow is that possible yes that flight is available departing at 9am out of san francisco arriving in seattle at 11 45 a.m i can book you in c12c it's an aisle would you like me to do that yes thank you amazon connect is a true omni-channel contact center service giving you the ability to maintain context across channels with an open architecture and a rich set of apis the possibilities are endless that has enabled businesses all over the world to reimagine the experiences they create for both their customers and their agents an exciting example of that is what origin energy has done with amazon connect and apple business chat let me show you what's possible i'll start with the end customer experience when i'm looking to contact a business i often find myself searching for them on either apple or google maps when i click the call button notice how i'm now presented with both a call and a chat option this type of customer experience solves one of the common issues with chat discoverability the conversation is automatically started in the ios messaging app when i send this message it will be handled with amazon connect chat and the experience i created in the amazon connect contact flow will begin i am presented with a list picker that i can interact with to select an option this is similar to response cards or a visual ivr i've selected a ride service demo and i'm now greeted by a chat bot because i've used this service before the system remembers me and greets me with a personalized greeting i'll ask to check for the status of my ride provide the trip number and the trip date our chatbot is powered by machine learning so it understands natural language like tomorrow next friday and so on and the same chat bot could be used for both voice and chat contacts i'd like to show you how amazon connect can seamlessly transition the conversation context if i escalate from a bot to a human so i'll ask to speak to someone about my bill you can even configure the chat bot to offer the customer the choice to switch channels i've now been put in queue to chat with an agent i'd like to show you what's possible by combining the amazon connect soft phone with the rich ai and machine learning services offered by aws to reimagine the agent experience on the left hand side of the screen you'll see the amazon connect contact control panel or the soft phone this is a webrtc based soft phone where agents only need an internet connection firefox or chrome and a headset it works perfectly in traditional contact centers or for work from home agents in the middle of the screen you'll see an area where we can use amazon translate to translate between languages this can enable agents to handle multiple languages for chat in real time on the right hand side of the screen you'll see a suggested actions area that is driven by our natural language enterprise search and indexing service called kendra powered by machine learning okay now back to the demonstration where i was put in queue to chat with an agent this is the incoming chat that was queued earlier when i answer the chat you'll see that the entire conversation between the customer and the chat bot has been provided to the agent right within the soft phone there's no integration or code to write to make this happen it's a native feature of amazon connect the agent now knows the exact reason for the contact and can be prepared to help the customer any other relevant data that might have been collected in the ivr or the chat bot can also be passed to the web application or crm directly from the soft phone amazon connect chat is natively asynchronous and gives you options for handling a customer or agent disconnecting or going away for a while let's say the agent disconnects the chat you can see the end customer getting notified and if they come back later or reply to the chat we can gracefully handle that reconnect them to the same agent or a different agent and still maintain the prior context now let's say this customer speaks spanish and decides to send a message in spanish to this english-speaking agent you'll notice that amazon translate converted the customer text to english and the key phrases were fed to amazon kendra which searched through the frequently asked questions to provide the agent with potential answers to the question the agent can then select an appropriate answer click on the thumbs up to help reinforce kendra's machine learning model for other agents and amazon translate will then convert the answer to the question from english back to spanish to the customer and those are just a few examples of how customers are using amazon connect and other aws ai and machine learning offerings to delight their customers and offer better agent experiences now i'd like to tell you a little bit about contact lens for amazon connect which can help you get a better understanding of your customers experience and gain customer insights and as of last month contact lens for amazon connect is now generally available for all amazon connect customers contact lens for amazon connect is a new set of analytics capabilities powered by machine learning that are integrated directly in amazon connect you can enable it with just a checkbox with contact lens you get new advanced search features in amazon connect where you can search for contacts based on words or phrases mentioned by either customers or agents or even sentiment scores you also get additional detailed contact analytics where you can see sentiment progression over time and the breakdown of customer agent and non-talk time another feature of contact lens is our automated contact categorization which i'll show you in the following demo we also have some exciting updates coming soon for contact lens including theme detection which will use machine learning to surface themes across many contacts to help you understand items of concern that you might not have even been aware of another upcoming feature is our supervisor assist feature which will surface contacts that need to be addressed in real time so you can take action to improve your customers experience before they hang up one of my favorite things about contact lens is how easy it is for businesses to turn it on and start gaining all these great insights you're also in full control as to which contacts to analyze let's take a look to enable contact lens you just drag the set recording and analytics behavior block into your contact flow if you don't already have it turn on call recording check the box for enable contact lens for speech analytics and set the language that's it there's no software to buy or tune no complex integrations and no licenses to purchase or manage contact lens can also automatically detect and redact sensitive data from both call recordings and transcripts and unlike traditional approaches for redaction contact lens uses a redaction engine powered by machine learning that has been trained on thousands of examples of pii data so it can intelligently pick out and redact things like names addresses credit card numbers and social security numbers contact lens can generate redacted and unredacted transcripts and redacted audio in addition to the original audio from call recordings you can then control access to the redacted and original versions using amazon connect's security profiles another feature of contact lens is the ability to automatically assign category labels to your contacts with a built-in rules engine in the rules page i can configure the categories assigned to contacts based on keywords and phrases mentioned in the contact some customers are using this capability to drive consistency in their call dispositions and automate call scoring here's an example i want to label contacts where the agent correctly greeted the customer so i can create a rule looking for the following phrases or words in the first 30 seconds of the call where the agent was the speaker i can also create complex categories by combining sets of words or phrases in different parts of the conversation or by speaker with contact lens enabled i get advanced search capabilities right in the amazon connect contact search page for example you can search for contacts where certain words or phrases were mentioned by either agents customers or both you can also search for contacts based on either customer or agent sentiment you can even find contacts that start or end with a given sentiment for example you can look for calls that started with a poor customer sentiment but ended positively another parameter you can search on is non-talk time often times having long non-talk time can be indicative of a poor customer experience or an agent training opportunity okay let's use some of these parameters to hone in on some specific calls i might be interested in in this environment we have a demo scenario based on a ride sharing company i want to find calls where a customer might have complained about a late ride which i know is a common complaint and would likely have a negative customer sentiment so i'll search for contacts where the customer mentioned late ride i'm also interested to see call examples where agents handled this well and were able to turn the customer around and end on a positive note contact lens normalizes sentiment scores with a range of minus 5 which would be very negative to positive 5 which would be very positive so i'll look for calls that started with a sentiment score of less than 0 and ended with more than 0. this will show me calls that started with a negative customer sentiment and ended with a neutral or positive customer sentiment i see there's a matching contact so i'll click it to get to the contact details page this is where i can see all the details for this specific contact on the top left i see the contact summary with things like the unique contact id the start and end times of the contact the contact duration and which agent handled the contact on the top right i can immediately see things like sentiment progression the breakdown of the customer sentiment over the entire call as well as the percentage of time the agent and customer spent speaking and the non-talk time over here we have the category labels based on the rules we defined earlier based on these labels it looks like the agent correctly welcomed the customer and correctly ended the call it also looks like contact lens correctly labeled this contact with ride issues given that the ride was late this contact categorization saves me time because now i can quickly tell if this is a call i want to dive deeper into without having to listen to the entire audio first this is the enhanced audio player where we get a quick visual representation of the call audio areas of the call that have negative and positive sentiment will be highlighted in red or green i can also see the audio amplitude so over here i see there's an area with an increased amplitude and negative sentiment let's click and see what was going on you'll notice that auto scroll will get me to the exact area of the call in the transcript let's take a listen i get it things happen but you guys run a business and i'm the customer right if you're gonna be late just tell me i can't just show up to a pitch when i want that's not how it works does this happen often do i need to use a different service so clearly the customer was upset and frustrated you'll notice that contact lens provided accurate transcription and punctuation next i'd like to show you how sensitive data redaction works redaction was enabled on this contact and because the user i'm logged in with doesn't have access to the unredacted information the pii data is redacted in the transcript as well as the audio we can easily listen to the associated audio by clicking on the timestamp and clicking the play button let's listen to the audio where the pi data was mentioned right so i see your 11 35 a.m pickup from this morning is that the trip you'll notice the audio has been redacted as well and the areas containing pii data were automatically replaced with silence this is important so that the overall length of the audio for the original and redacted files are consistent the cool thing about this is that these call recordings are stored in your aws s3 bucket so you can easily access the raw audio files if you needed them for other use cases as a tool mentioned in his overview theme detection is coming later this year to contact lens here you can see contact lens uses natural language processing to automatically detect and underline words and phrases from the call transcript that indicate call drivers or reasons for the customer's outreach for individual calls as part of theme detection contact lens will identify common patterns of call drivers across multiple calls using machine learning as we look at the rest of the transcript we can see that the agent acknowledged the problem apologize to the customer offered a refund and a future credit which is all within the company guidelines the customer is happy with the result and the agent ended the conversation with have a great day now it's great to have all these insights such as transcripts sentiment scores and categorization labels at your fingertips right inside of the amazon connect user interface but this data is so rich and useful that getting it programmatically can unlock all sorts of additional use cases for your business well you can now have your cake and eat it too contact lens will automatically take all the data it's generated and upload it to your s3 bucket in a human readable json format the json file will include all the same items you see in the ui such as identified categories conversation characteristics sentiment breakdown and turn-by-turn transcription including exactly who said what and at what time stamp in the call and if redaction was enabled you'll see that the text is redacted in the json output and you'll even have the exact timestamps for the redaction the json file will also include some items you don't see in the ui yet such as interruptions including whether the customer interrupted the agent or the agent interrupted the customer and talk speed like average words per minute for both agent and customers i hope you've enjoyed this brief demonstration of contact lens for amazon connect back to you pasquale running contact centers requires a number of disparate tools historically it's no easy task to reliably work together and get them to work the way you want to in a flexible environment but you know your contact center needs to be flexible moving forward amazon connect is built from the ground up to be a hundred percent cloud-based to enable that flexibility for starters it's got paid to go pricing so you don't have to worry about capacity planning or complex telephony integration additionally that brings about the next topic which is scale you want to be able to scale up and down because as we've seen probably the last few months it's very hard to predict what your customers are going to need they might have an emergency one day and then two days later they might have a lot less usage it's nice to be able to know your system can handle either of those completely fluidly with no worries and of course all of this is only interesting because it can integrate very easily with all the other tools they have so for example if they have existing systems whether that be workforce management crms we often have a pre-packaged solution for them for example salesforce zendesk aspect if they don't have that it's easy to build and our customers have told us they often get up and running in as little as a day and when you think of crms you might think of salesforce and now amazon connect is not only their preferred contact center solution it also powers a new service cloud voice product and just recently salesforce amazon connect officially launched service cloud voice service cloud voice allows for natively integrated phone and digital channels in a powerful tool that's seamless to use agents can get real-time ai powered call transcription and help with all the things they need to give great customer service and for supervisors they can view calls in real time and help their agents perform even better here's how a few of our many customers have been able to reimagine customer service and support spokeo is a popular people search information service in just two weeks they fully deployed contact center they migrated all 75 agents and started quickly accepting over 50 000 contacts a month with amazon connect that was easy supercare health is a leading provider of in-home ventilation therapy their on-premises contact center had 300 agents and two voice gateways two application servers two media servers a recording server a reporting server they moved to amazon connect to avoid high maintenance licensing costs unplanned downtime and to reduce agent onboarding time john heincock also a salesforce customer use contact lens for amazon connect they can use this to automatically understand and assess every call that happens in their contact center using powerful ml they can see why customers are calling look for themes and address problems or challenges to help train their agents to provide even better customer service in the future so today we've talked a lot about how customers have reacted to covid you've heard a little bit about how people are starting to return to work such as amazon and now we're talking a little bit about how we're reimagining the contact center for the future using amazon connect and of course it's great to see it all in action but there's still a lot of work ahead of us we don't want you to feel like you're alone we're here to help first we have experts in amazon connect to help you our customer enablement team includes folks who can provide aws trainings certifications we have one-off project help for example a tool called aws iq lets you get off the shelf solutions that you only pay for which are up and running and doing what you need them to do we can also work with you to create appointment bots and normalize your data these aws iq projects are really powerful and if you're a partner we want you on this journey with us partners can come on board with the amazon partner network the apn and it's a global program with new folks coming on board from both sis and from isvs to build software all the time consulting partners software builders we welcome you all we love you to help deliver great products for our joint customers together accenture and voice foundry are great examples of this as is deloitte lots of folks have come on board thank you for spending your time with us today stop by our virtual booth and learn more about amazon connect also enjoy the rest of the conference i really look forward to seeing you hopefully next year in person you

Channel: Amazon Web Services

Views: 2,833

Keywords: AWS, Amazon Web Services, Cloud, AWS Cloud, Cloud Computing, Workforce, Continuity, Amazon, Connect, contact lens, customer experience, contact center, customer engagement, ml-powered, analytics, machine learning, artificial intelligence

Id: pARyuIZpXz4

Length: 28min 33sec (1713 seconds)

Published: Fri Sep 04 2020