1 00:00:00,690 --> 00:00:01,523 Now let's talk about 2 00:00:01,523 --> 00:00:05,150 how we can build chatbots with Amazon Lex & Connect. 3 00:00:05,150 --> 00:00:07,360 So Amazon Lex is the same technology 4 00:00:07,360 --> 00:00:08,780 that powers the Alexa device, 5 00:00:08,780 --> 00:00:12,220 so if you know Amazon as a virtual assistant called Alexa, 6 00:00:12,220 --> 00:00:13,540 you do, "Alexa, do something", 7 00:00:13,540 --> 00:00:16,120 and triggers some kind of recognition. 8 00:00:16,120 --> 00:00:17,750 And so Amazon Lex gives you 9 00:00:17,750 --> 00:00:21,210 that same automatic speech recognition, ASR, 10 00:00:21,210 --> 00:00:24,340 that will be helping you to convert speech to text. 11 00:00:24,340 --> 00:00:26,210 But there will be this added benefit 12 00:00:26,210 --> 00:00:28,270 of Natural Language Understanding 13 00:00:28,270 --> 00:00:31,630 to have an intent of the text that you're converting 14 00:00:31,630 --> 00:00:33,700 or the speech that you're converting. 15 00:00:33,700 --> 00:00:37,840 And the idea behind this is that, with this technology, 16 00:00:37,840 --> 00:00:40,150 we're going to be able to build chatbots 17 00:00:40,150 --> 00:00:42,020 or call center bots. 18 00:00:42,020 --> 00:00:45,370 So Amazon Lex doesn't just convert speech into text, 19 00:00:45,370 --> 00:00:48,410 it also helps understand what the intent 20 00:00:48,410 --> 00:00:51,940 and what the meaning of that text or that speech is. 21 00:00:51,940 --> 00:00:55,060 And so, on its own, Amazon Lex can be used 22 00:00:55,060 --> 00:00:58,130 as a standalone service, for chatbots and call center bots, 23 00:00:58,130 --> 00:01:00,320 which is good from an exam perspective. 24 00:01:00,320 --> 00:01:02,750 But then how do we have that call center in the first place? 25 00:01:02,750 --> 00:01:04,849 Well, you can create an Amazon Connect. 26 00:01:04,849 --> 00:01:06,450 And so Amazon Connect is a way for you 27 00:01:06,450 --> 00:01:08,520 to receive actual phone calls, 28 00:01:08,520 --> 00:01:10,120 and to create contact flows, 29 00:01:10,120 --> 00:01:12,030 and it's going to be a virtual contact center. 30 00:01:12,030 --> 00:01:14,890 So that means that you can have people calling you 31 00:01:14,890 --> 00:01:17,830 at a specific number, and you will be there 32 00:01:17,830 --> 00:01:20,350 to connect it into your inner services. 33 00:01:20,350 --> 00:01:22,170 Amazon Connect can also integrate 34 00:01:22,170 --> 00:01:24,440 with your customer relationship management system, 35 00:01:24,440 --> 00:01:26,720 so CRM, or AWS services. 36 00:01:26,720 --> 00:01:28,710 So let's see a flow of how that would work. 37 00:01:28,710 --> 00:01:30,510 So, say we have a user, 38 00:01:30,510 --> 00:01:32,860 and they want to schedule an appointment with us. 39 00:01:32,860 --> 00:01:34,090 So they will place a phone call, 40 00:01:34,090 --> 00:01:35,200 and they have to call something. 41 00:01:35,200 --> 00:01:37,570 So the first thing's that we do Amazon Connect. 42 00:01:37,570 --> 00:01:40,890 So remember whenever we want to connect our phone call 43 00:01:40,890 --> 00:01:43,370 with our infrastructure, we need Amazon Connect. 44 00:01:43,370 --> 00:01:45,240 So this is a visual tip, 45 00:01:45,240 --> 00:01:48,447 so we get the call each others on Connect. 46 00:01:48,447 --> 00:01:51,570 And then, from Connect, we need to recognize the intent 47 00:01:51,570 --> 00:01:54,160 of the caller, what is the caller asking? 48 00:01:54,160 --> 00:01:55,417 They would say, for example, 49 00:01:55,417 --> 00:01:57,490 "I want to schedule an appointment." 50 00:01:57,490 --> 00:02:00,360 And so, therefore, this stream of audio 51 00:02:00,360 --> 00:02:03,040 is streamed into Amazon Lex. 52 00:02:03,040 --> 00:02:06,290 And Lex, when we hear, "I want to schedule an appointment", 53 00:02:06,290 --> 00:02:08,690 uses machine learning to figure out, okay, 54 00:02:08,690 --> 00:02:11,600 what they want to do is actually schedule an appointment. 55 00:02:11,600 --> 00:02:13,460 And so, for this, there's a Lambda function 56 00:02:13,460 --> 00:02:16,090 that we would create to schedule this appointment. 57 00:02:16,090 --> 00:02:19,360 And Lex would be smart enough to recognize the intent, 58 00:02:19,360 --> 00:02:22,080 parse it, and then invoke another function with it. 59 00:02:22,080 --> 00:02:23,720 And our Lambda function, for example, 60 00:02:23,720 --> 00:02:25,070 could schedule that appointment 61 00:02:25,070 --> 00:02:27,260 into whatever system we want. 62 00:02:27,260 --> 00:02:28,860 So, from an exam perspective, remember that, 63 00:02:28,860 --> 00:02:30,780 any time you need to build chatbots, 64 00:02:30,780 --> 00:02:33,020 you think of Amazon Lex, and anytime you need 65 00:02:33,020 --> 00:02:34,940 to build a virtual cloud contact center, 66 00:02:34,940 --> 00:02:36,200 you use Amazon Connect, 67 00:02:36,200 --> 00:02:38,680 and these two things integrate together really well. 68 00:02:38,680 --> 00:02:39,930 So that's it, I hope you liked it, 69 00:02:39,930 --> 00:02:41,880 and I will see you in the next lecture.