WEBVTT - Jeetu Patel Talks Evolving AI Trends

0:00:02.720 --> 0:00:09.600
<v Speaker 1>Bloomberg Audio Studios, podcasts, radio news. Okay, perfect transition to

0:00:09.800 --> 0:00:12.280
<v Speaker 1>get to our next guest, because we're talking about the

0:00:12.440 --> 0:00:15.240
<v Speaker 1>increased energy usage that we've seen a lot of that

0:00:15.280 --> 0:00:17.360
<v Speaker 1>has to do with AI. Check out all the news

0:00:17.360 --> 0:00:20.200
<v Speaker 1>about AI just in the last couple of days, Carol.

0:00:20.200 --> 0:00:22.400
<v Speaker 1>Lots of headlines. Apple rolling out this five hundred and

0:00:22.440 --> 0:00:25.520
<v Speaker 1>fifty nine dollars iPhone five hundred and ninety nine dollars

0:00:25.520 --> 0:00:28.920
<v Speaker 1>excuse me, iPhone sixteen eight with an AI bid for growth.

0:00:29.440 --> 0:00:33.040
<v Speaker 1>Microsoft unveiling AI tools that can create video game scenes

0:00:33.080 --> 0:00:36.159
<v Speaker 1>that would normally have to be programmed and animated by

0:00:36.159 --> 0:00:38.440
<v Speaker 1>a human. It's a model it built using data collected

0:00:38.800 --> 0:00:41.560
<v Speaker 1>from Xbox gamers and their controllers.

0:00:41.640 --> 0:00:43.640
<v Speaker 2>There's so much going on, and don't forget this comes

0:00:43.640 --> 0:00:46.080
<v Speaker 2>after the big news earlier this week. We talked about this.

0:00:46.200 --> 0:00:49.360
<v Speaker 2>Mira Muradi, the former chief CTO over an Open Ai,

0:00:49.520 --> 0:00:52.680
<v Speaker 2>launching a new AI startup. And then you had open

0:00:52.720 --> 0:00:56.160
<v Speaker 2>ai co founder Ilias a Skiver raising more than a

0:00:56.240 --> 0:00:59.520
<v Speaker 2>billion dollars for his startup at evaluation of over thirty

0:00:59.680 --> 0:01:00.600
<v Speaker 2>billion dollars.

0:01:00.720 --> 0:01:02.639
<v Speaker 1>Okay, lots of money coming into this and G two

0:01:02.640 --> 0:01:05.120
<v Speaker 1>Patel is on top of all of it. He's Executive

0:01:05.160 --> 0:01:07.720
<v Speaker 1>VP and Chief Product Officer at Cisco, where he's charged

0:01:07.760 --> 0:01:11.399
<v Speaker 1>with making sure that Cisco's portfolio works in the world

0:01:11.440 --> 0:01:15.440
<v Speaker 1>of AI. He joins us from Silicon Valley G two.

0:01:16.040 --> 0:01:21.039
<v Speaker 1>Where do you see Silica see Cisco's place in this

0:01:21.120 --> 0:01:21.960
<v Speaker 1>new world of AI?

0:01:24.160 --> 0:01:26.480
<v Speaker 3>You know the way that we think about this. Firstly, Carolynton,

0:01:26.520 --> 0:01:28.959
<v Speaker 3>thank you for having me on the show. And you know,

0:01:29.000 --> 0:01:31.200
<v Speaker 3>if you look at the way that the world is

0:01:31.240 --> 0:01:35.480
<v Speaker 3>evolving right now, it's it's moving really fast. But if

0:01:35.520 --> 0:01:37.959
<v Speaker 3>you look back at the gold Rush, think about Cisco's

0:01:37.959 --> 0:01:40.920
<v Speaker 3>position over here is we're going to provide the infrastructure

0:01:41.720 --> 0:01:45.160
<v Speaker 3>and the safety and security guard rail so that organizations

0:01:45.200 --> 0:01:48.720
<v Speaker 3>can actually use AI scalably in the market. And if

0:01:48.720 --> 0:01:51.160
<v Speaker 3>you think about a common sentiment that you have right

0:01:51.200 --> 0:01:54.240
<v Speaker 3>now in the market, and we've done a study to

0:01:54.280 --> 0:01:59.280
<v Speaker 3>actually prove this as well, there's an overwhelming majority of

0:01:59.320 --> 0:02:03.160
<v Speaker 3>companies where the CEOs are very excited about the possibilities

0:02:03.160 --> 0:02:05.080
<v Speaker 3>of AI, but virtually all of.

0:02:05.040 --> 0:02:06.480
<v Speaker 4>Them feel very underprepared.

0:02:06.560 --> 0:02:08.880
<v Speaker 3>In fact, the study that we did, we found that

0:02:09.000 --> 0:02:12.800
<v Speaker 3>ninety seven percent of the CEOs that actually we're pushing

0:02:13.000 --> 0:02:16.320
<v Speaker 3>AI projects that we're really excited about it only one

0:02:16.360 --> 0:02:19.360
<v Speaker 3>point seven percent of them felt fully prepared, and I

0:02:19.400 --> 0:02:22.120
<v Speaker 3>totally get it. I mean, this is a it's a

0:02:22.200 --> 0:02:25.000
<v Speaker 3>very fast moving market. People have to kind of think differently,

0:02:25.639 --> 0:02:28.480
<v Speaker 3>and so that's an area where we feel like, you know,

0:02:28.560 --> 0:02:32.280
<v Speaker 3>having preparedness and infrastructure, making sure that safety and security

0:02:32.800 --> 0:02:36.320
<v Speaker 3>don't become impediments for the adoption of AI, and having

0:02:36.400 --> 0:02:37.600
<v Speaker 3>the skills.

0:02:37.120 --> 0:02:41.560
<v Speaker 4>Gap get really addressed are the three.

0:02:41.360 --> 0:02:44.160
<v Speaker 3>Big challenges that most organizations have that we at Cisco

0:02:44.240 --> 0:02:46.880
<v Speaker 3>want to try to go out and help our customers

0:02:46.880 --> 0:02:47.919
<v Speaker 3>with CEE two.

0:02:47.919 --> 0:02:49.320
<v Speaker 2>One thing I think about is, you know, there's the

0:02:49.360 --> 0:02:52.799
<v Speaker 2>idea of move fast, break things, get it. There's also well,

0:02:52.840 --> 0:02:54.760
<v Speaker 2>wait a minute, we want to see where the dust

0:02:54.760 --> 0:02:57.240
<v Speaker 2>settles when it comes to AI a little bit because

0:02:57.720 --> 0:02:59.920
<v Speaker 2>we're at this point two more than two years in

0:03:00.040 --> 0:03:03.359
<v Speaker 2>and of just the euphoria and the build and the spend,

0:03:03.480 --> 0:03:05.760
<v Speaker 2>and now we want to see, okay, what do we

0:03:05.760 --> 0:03:07.800
<v Speaker 2>get for our money? Is there a little bit maybe

0:03:08.320 --> 0:03:11.639
<v Speaker 2>of that going on that there are executives, especially if

0:03:11.639 --> 0:03:14.000
<v Speaker 2>smaller mid sized companies, who are saying, we got to

0:03:14.040 --> 0:03:16.760
<v Speaker 2>kind of hold off until we really understand how do

0:03:16.760 --> 0:03:17.680
<v Speaker 2>we monetize this.

0:03:19.440 --> 0:03:21.680
<v Speaker 3>Yes, if you take a step back and broadly, like,

0:03:22.160 --> 0:03:24.160
<v Speaker 3>if we like to take a look at what's happening

0:03:24.160 --> 0:03:26.239
<v Speaker 3>in the market, there's only two kinds of companies that

0:03:26.280 --> 0:03:29.119
<v Speaker 3>are going to exist in the market. There's a first kind,

0:03:29.120 --> 0:03:31.239
<v Speaker 3>which is going to get highly dexterous with the use

0:03:31.240 --> 0:03:33.880
<v Speaker 3>of AI, and the second which is going to be

0:03:34.840 --> 0:03:37.200
<v Speaker 3>you know, struggling for relevance. Now, if you look at

0:03:37.240 --> 0:03:40.960
<v Speaker 3>the great ones, majority of them are getting slowed down

0:03:41.400 --> 0:03:46.240
<v Speaker 3>because of a few very foundational reasons. The first one is,

0:03:47.360 --> 0:03:49.480
<v Speaker 3>you know, safety and security gets to be a big concern,

0:03:49.560 --> 0:03:52.280
<v Speaker 3>especially for large regulated organizations. They want to make sure

0:03:52.280 --> 0:03:55.320
<v Speaker 3>that they tread carefully on that front. And as you

0:03:55.400 --> 0:03:58.520
<v Speaker 3>have things like deep seak and you know, the cost

0:03:58.560 --> 0:04:01.000
<v Speaker 3>of models gets to be lower and lower, and the

0:04:01.000 --> 0:04:03.880
<v Speaker 3>cost curve goes down precipitously, you're going to have more

0:04:03.880 --> 0:04:06.320
<v Speaker 3>and more models that are out there. And these models

0:04:06.320 --> 0:04:09.760
<v Speaker 3>by definition are unpredictable, and we've found that it's pretty

0:04:09.800 --> 0:04:12.480
<v Speaker 3>easy to jail break these models, and so that safety

0:04:12.560 --> 0:04:14.200
<v Speaker 3>security concern is a real one.

0:04:14.880 --> 0:04:16.360
<v Speaker 4>The second area is that.

0:04:16.279 --> 0:04:19.680
<v Speaker 3>People just aren't quite equipped for going out and making

0:04:19.720 --> 0:04:23.000
<v Speaker 3>sure that the infrastructure is ready for AI. And then

0:04:23.040 --> 0:04:25.360
<v Speaker 3>third one is this notion of skills, and if you

0:04:25.400 --> 0:04:28.960
<v Speaker 3>think about skills, I personally believe that staying on the

0:04:29.040 --> 0:04:32.320
<v Speaker 3>sidelines is not going to help you go up the

0:04:32.400 --> 0:04:35.440
<v Speaker 3>learning curve. So you have to jump in and you

0:04:35.480 --> 0:04:36.960
<v Speaker 3>have to jump in the fray and make sure that

0:04:37.000 --> 0:04:41.000
<v Speaker 3>you're starting to take some projects and learn from those

0:04:41.040 --> 0:04:45.000
<v Speaker 3>projects and have these practical experiences as a company so

0:04:45.040 --> 0:04:47.200
<v Speaker 3>that you get a better feel, a better intuition of

0:04:47.200 --> 0:04:49.360
<v Speaker 3>what's happening in AI and what it's going to take

0:04:49.400 --> 0:04:52.200
<v Speaker 3>to succeed. But just staying on the sidelines, I think

0:04:52.279 --> 0:04:55.080
<v Speaker 3>is a really bad strategy that's going to let others

0:04:55.120 --> 0:04:58.279
<v Speaker 3>get ahead of you and you'll.

0:04:57.680 --> 0:04:58.279
<v Speaker 4>Be left behind.

0:04:58.839 --> 0:05:02.160
<v Speaker 1>One thing love to hear from folks in your position

0:05:02.320 --> 0:05:05.120
<v Speaker 1>who are interacting with customers each and every day is

0:05:05.440 --> 0:05:09.480
<v Speaker 1>try to get an understanding of where demand is, who's buying,

0:05:09.560 --> 0:05:12.560
<v Speaker 1>what house spend is. When it comes to it, what

0:05:12.560 --> 0:05:16.600
<v Speaker 1>can you tell us about the environment right now, So there's.

0:05:16.360 --> 0:05:19.440
<v Speaker 4>No shortage of demand signal on AI right now.

0:05:19.480 --> 0:05:22.120
<v Speaker 3>In fact, what's happening is there's a lot of money

0:05:22.200 --> 0:05:25.760
<v Speaker 3>in from different parts of it that are going into AI.

0:05:26.240 --> 0:05:29.000
<v Speaker 3>And then the question becomes, what are the use cases

0:05:29.040 --> 0:05:30.839
<v Speaker 3>where AI is being used really well? So if you

0:05:30.880 --> 0:05:34.560
<v Speaker 3>think about some of the you know, really highly adopted

0:05:34.640 --> 0:05:39.880
<v Speaker 3>use cases right now in large companies software coding and

0:05:39.920 --> 0:05:44.359
<v Speaker 3>software development, in twenty twenty five, you will actually be

0:05:44.440 --> 0:05:50.760
<v Speaker 3>able to have, you know, a meaningful amount of your

0:05:50.800 --> 0:05:53.279
<v Speaker 3>code for a mid level engineer that'll actually be done

0:05:53.279 --> 0:05:55.640
<v Speaker 3>by AI. And so that's an area that's going to

0:05:55.640 --> 0:05:59.680
<v Speaker 3>be hugely you know growing. The second area that we've

0:05:59.720 --> 0:06:02.360
<v Speaker 3>seen a fair amount of movement and momentum on is

0:06:02.440 --> 0:06:05.279
<v Speaker 3>this notion of a contact center agent. Each one of us,

0:06:05.640 --> 0:06:08.080
<v Speaker 3>Caroly and Tim, we all probably will spend about four

0:06:08.080 --> 0:06:11.400
<v Speaker 3>to six weeks of our lifetimes on hold waiting for

0:06:11.880 --> 0:06:14.600
<v Speaker 3>some agent. When you call a company and they need

0:06:14.640 --> 0:06:17.120
<v Speaker 3>to make sure that they actually solve your problem. That's

0:06:17.160 --> 0:06:20.440
<v Speaker 3>an enormous waste of time for humanity at large. And

0:06:21.240 --> 0:06:23.480
<v Speaker 3>I think there's going to be ninety to ninety five

0:06:23.520 --> 0:06:27.159
<v Speaker 3>percent of these calls will be addressed with a voice

0:06:27.200 --> 0:06:29.159
<v Speaker 3>agent that is going to be an AI agent, where

0:06:29.200 --> 0:06:31.040
<v Speaker 3>you it'll sound just waiting.

0:06:32.720 --> 0:06:34.520
<v Speaker 1>Really, I'm really still waiting for that to happen.

0:06:34.560 --> 0:06:37.200
<v Speaker 2>One that actually works and is I don't know, like

0:06:37.240 --> 0:06:39.680
<v Speaker 2>I do feel like we've all gone through automated.

0:06:39.240 --> 0:06:43.680
<v Speaker 1>Systems and representative representative et representative.

0:06:44.120 --> 0:06:44.719
<v Speaker 4>That's how I do.

0:06:45.680 --> 0:06:47.000
<v Speaker 3>That's exactly right.

0:06:47.920 --> 0:06:51.680
<v Speaker 2>So is it really going to be a very different experience.

0:06:53.080 --> 0:06:54.800
<v Speaker 3>Well, I think one of the challenges that we've had

0:06:54.839 --> 0:06:59.159
<v Speaker 3>if you think about, you know, calling an agent, and

0:06:59.200 --> 0:07:02.440
<v Speaker 3>when you have something that's an automated agent, the reason

0:07:02.480 --> 0:07:04.440
<v Speaker 3>it sounds so frustrating is that it sounds like a

0:07:04.520 --> 0:07:08.960
<v Speaker 3>robot and you can't interrupt it. It doesn't understand tonality,

0:07:08.960 --> 0:07:12.400
<v Speaker 3>it doesn't understand expression, doesn't understand the emotion that the

0:07:12.440 --> 0:07:15.480
<v Speaker 3>customer might be feeling. And so what you're now starting

0:07:15.520 --> 0:07:18.640
<v Speaker 3>to see happen is these voice agents are getting so

0:07:18.640 --> 0:07:22.240
<v Speaker 3>sophisticated that you can interrupt its speed. It speaks at

0:07:22.240 --> 0:07:24.520
<v Speaker 3>the speed that a human would be able to speak.

0:07:24.560 --> 0:07:27.720
<v Speaker 3>You can redirect it, you can throw curveballs. It'll understand

0:07:27.800 --> 0:07:29.800
<v Speaker 3>your tone if you're frustrated. It's not going to ask

0:07:29.840 --> 0:07:34.040
<v Speaker 3>you silly questions when you're frustrated about things that are

0:07:34.080 --> 0:07:36.000
<v Speaker 3>like how's the weather, how are you feeling, because that's

0:07:36.040 --> 0:07:38.920
<v Speaker 3>not what the customer might be might might want to

0:07:38.960 --> 0:07:43.080
<v Speaker 3>talk about. So those kind of dynamics will fundamentally change

0:07:43.560 --> 0:07:45.480
<v Speaker 3>the take rate and adoption of this.

0:07:45.560 --> 0:07:46.520
<v Speaker 4>So in the next two.

0:07:46.440 --> 0:07:51.400
<v Speaker 3>Years, I assert that it'll probably be about eighty to

0:07:51.440 --> 0:07:55.120
<v Speaker 3>eighty five percent of the inbound calls that come in

0:07:55.200 --> 0:07:57.040
<v Speaker 3>will be able to be handled by an AI agent,

0:07:57.080 --> 0:07:59.559
<v Speaker 3>and the ones that aren't handled by an AI agent,

0:08:00.040 --> 0:08:01.880
<v Speaker 3>the AI agent will be smart enough to hand it

0:08:01.920 --> 0:08:04.960
<v Speaker 3>off to a human agent with all the context, so

0:08:05.000 --> 0:08:07.000
<v Speaker 3>you don't have to repeat your account number again and

0:08:07.040 --> 0:08:09.920
<v Speaker 3>you don't have to go out and you know, kind

0:08:09.920 --> 0:08:12.320
<v Speaker 3>of restate your problem just the way that we actually

0:08:12.360 --> 0:08:15.240
<v Speaker 3>feel right now. And that's what gets the frustration levels high.

0:08:15.720 --> 0:08:18.680
<v Speaker 3>And I'm I'm really hopeful about that that entire category.

0:08:18.720 --> 0:08:20.600
<v Speaker 3>So that's the second category where you're seeing a fair

0:08:20.640 --> 0:08:24.680
<v Speaker 3>amount of momentum, and then there's there's many more use

0:08:24.720 --> 0:08:27.520
<v Speaker 3>cases and automated workflows that you'll start to just see

0:08:27.560 --> 0:08:33.120
<v Speaker 3>get more and more proliferated throughout the entire corporate landscape.

0:08:33.200 --> 0:08:35.560
<v Speaker 2>Well don't judge me, G two, but I just started

0:08:35.559 --> 0:08:38.400
<v Speaker 2>really playing with chat Gipt and I came in and

0:08:38.440 --> 0:08:40.440
<v Speaker 2>I was like kind of blown away. I came in

0:08:40.440 --> 0:08:42.440
<v Speaker 2>one morning like, Tim, you've got to start playing with You're.

0:08:42.320 --> 0:08:43.880
<v Speaker 1>Saying you soon you're going to be paying two hundred

0:08:43.880 --> 0:08:45.760
<v Speaker 1>bucks a month for the agent future.

0:08:46.000 --> 0:08:46.440
<v Speaker 4>I see it.

0:08:46.600 --> 0:08:47.760
<v Speaker 2>Everything's a monthly faith.

0:08:48.520 --> 0:08:50.640
<v Speaker 3>I don't know if you folks have tried out this operator.

0:08:50.800 --> 0:08:53.839
<v Speaker 3>It is unbelievable what it can do because.

0:08:54.800 --> 0:08:57.079
<v Speaker 1>A month version, right, yeah.

0:08:56.920 --> 0:08:58.960
<v Speaker 3>It's it's the two hundred dollars a month version. And

0:08:59.000 --> 0:09:00.920
<v Speaker 3>by the way, the interesting part is they're losing money

0:09:00.920 --> 0:09:02.719
<v Speaker 3>on the two hundred dollars a month, which tells you

0:09:03.200 --> 0:09:06.320
<v Speaker 3>the amount of impact and the amount of adoption rate.

0:09:06.360 --> 0:09:08.600
<v Speaker 4>That is most people think it's bad. I think it's

0:09:08.600 --> 0:09:09.200
<v Speaker 4>a really.

0:09:08.960 --> 0:09:12.040
<v Speaker 3>Good sign for EI because there's no shortage of people

0:09:12.120 --> 0:09:14.600
<v Speaker 3>wanting to use this. They're solving a real problem. So

0:09:14.760 --> 0:09:18.480
<v Speaker 3>you can literally tell Carol, tell this agent, Hey, I

0:09:18.760 --> 0:09:20.640
<v Speaker 3>want to go out for a movie with my spouse

0:09:20.679 --> 0:09:22.480
<v Speaker 3>tonight and I want to make sure that I can

0:09:22.559 --> 0:09:25.760
<v Speaker 3>have dinner. Can you go find me a dinner reservation.

0:09:27.200 --> 0:09:28.319
<v Speaker 4>And I'll just do it for you.

0:09:29.000 --> 0:09:32.240
<v Speaker 2>Yeah, listen, come back soon. I really would like to

0:09:32.280 --> 0:09:35.000
<v Speaker 2>continue this conversation with you. Really enjoyed it. G Two Patel,

0:09:35.320 --> 0:09:38.240
<v Speaker 2>executive vice president and chief product officer over at Cisco,

0:09:38.640 --> 0:09:40.800
<v Speaker 2>joining us from Silicon Valley