Marvis + ChatGPT + Zoom Integration Demo

0:00 um

0:05 so let's actually look at how much of

0:07 this is real uh so so this is chat GPT

0:09 connected to the Marvis conversational

0:12 interface uh and essentially you know

0:15 being able to ask uh chat CPT hey you

0:19 know

0:19 um can we actually leverage large

0:22 language models uh in in Marvis right in

0:26 the in the conversational system so

0:27 here's a question we've actually piped

0:30 this question through the uh the open uh

0:34 um chat GPT model out there and so we're

0:36 actually getting

0:38 um you know natural language generation

0:41 Marvis always had natural language

0:45 understanding and we obviously had all

0:48 this data to arrive at the right answer

0:50 but for the first time we're actually

0:53 introducing now natural language

0:55 generation previously if you asked us a

0:58 question how do you configure Dynamic

0:59 Port profiles we'd throw up a bunch of

1:01 links like this like Google used to do

1:03 now we're leveraging GPT natively in

1:07 Marvis and so this is coming to you uh

1:10 uh to your dashboard uh fairly soon uh

1:13 in the Q3 time frame I'm super excited

1:16 about the zoom integration this one is

1:19 is is incredibly powerful in what we can

1:22 do with zoom uh basically

1:25 you know native integration from Zoom so

1:29 and teams will follow the second half of

1:31 this year but Zoom now if you are a Zoom

1:33 customer we can natively bring in Zoom

1:36 data into your missed Cloud you can

1:39 actually compare and contrast it with

1:41 the Lost latency Jitter you see out

1:43 there that Zoom sees labeled data and

1:47 Bob said labeled data is gold when it

1:50 comes to Ai and this is labeled data for

1:53 the first time we have a cloud

1:54 architecture in this industry that can

1:57 handle the zoom data every user every

2:00 minute every Zoom call all the time

2:03 we're consuming this from organizations

2:05 right now and actually be able to even

2:07 predict Bob didn't go into the details

2:10 of explainability and predictability the

2:12 predictability is US building a model

2:14 where even if you don't have zoom calls

2:16 at a site can I tell you if you know the

2:19 wireless clients or the wireless

2:21 RSSI maybe the van latency the round

2:24 trip time what is causing this kind of

2:27 implications on Zoom one of the best

2:30 parts of the zoom integration is it's

2:32 natively integrated into the marvelous

2:34 conversational system so The Marvelous

2:36 conversational assistant you could ask

2:37 tell me what Zoom calls do you know

2:40 happen today at the scale of your

2:42 organization you know I have one

2:44 customer uh we're doing they're doing

2:46 about 30 000 Zoom calls a day just one

2:49 customer small customer 30 000 Zoom

2:52 calls a day you're able to just ask the

2:54 question who's who's doing Zoom calls

2:56 and how is that experience that Green

2:58 Dot says those were great Zoom calls the

3:00 user said that was good no problems but

3:03 the best query my absolute favorite

3:07 query on this whole feature is basically

3:10 saying you know what we have all these

3:13 clients but how do I know how do I know

3:16 if there was one bad Zoom call today

3:19 right just ask that question to Marvis

3:21 and Marvin says aha there were sites

3:24 that had bad Zoom calls Kumar's MacBook

3:26 Pro had a bad Zoom call today and you

3:28 click on it and voila for the first time

3:31 in the networking industry we're taking

3:33 data from truly real-time data applying

3:37 AI on it we're applying that sharply

3:39 model that Bob talked about and being

3:41 able to come up with an answer saying

3:42 hey that Zoom call was bad because that

3:45 user was roaming at the exact same time

3:47 so we'd say okay let's triple check this

3:49 and let's ask you know Marvis hey Marvis

3:52 you know was that really used having a

3:54 bad experience of course that user was

3:56 having a bad experience and it's it's

3:58 our favorite it's it's RF Engineers

4:00 worst nightmare is this interband

4:02 roaming all the time that happens it's

4:05 now native in uh in the dashboard so

4:08 this is available today if you're a Zoom

4:10 customer and a missed customer connect

4:12 with your account teams and we can make

4:14 this happen we're taking not just zoom's

4:17 perspective of it but also the user's

4:21 feedback when you have that Zoom call

4:23 hey how was that Zoom call nobody ever

4:25 answers that question except when the

4:27 call sucks everybody's like oh yeah that

4:29 call was bad right we need that labeled

4:32 data right to say all of all is well in

4:35 all these other instances but this one

4:37 call wasn't good if you get that labeled

4:39 data it's awesome and then again this is

4:43 another really really good advancement

4:46 from Marvis perspective in terms of sort

4:49 of Marvis new actions I won't do a full

4:52 Deep dive on the new actions here

4:54 tomorrow the entire boot camp there is a

4:57 switching section there there's a

4:59 wireless session brand new Marvis

5:01 actions are coming to the dashboard we

5:03 want you to please participate in

5:05 tomorrow's boot camp it's going to be

5:07 phenomenal it's it's customers

5:08 presenting with juniper product product

5:11 teams and you will see this is one of my

5:14 new favorite marvelous actions basically

5:17 think of it this way imagine all the iot

5:20 sensors video cameras all of that in our

5:22 Network today

5:23 what happens suddenly if one of those

5:25 video cameras stops sending traffic

5:27 who's watching it if a thermostat stops

5:30 sending data to its Cloud who's watching

5:32 it Marvis is watching it and this is

5:35 Marvis catching it right and so this is

5:37 game changing and then last but not

5:39 least uh probably uh my favorite of all

5:43 of these are is this Marvis as a member

5:45 of your team Marvis is coming to A

5:48 team's Channel near you and so Marvis

5:51 actually we've submitted to Microsoft

5:53 for the app integration here and um as

5:56 soon as um uh it gets approved this will

5:58 be released but you can invite Marvis to

6:01 a conversation with your fellow team

6:03 member and say Hey you know Ryan uh

6:06 what's up with this particular infusion

6:08 prop in this Hospital ask Marvis right

6:10 so Marvis can participate in a

6:14 conversation within your own team and

6:16 this is super cool uh and you know

6:19 whatever you could do on the VNA for the

6:21 most part uh you know you can actually

6:23 do natively in this so and all of us how

6:27 many of you are teams customers here

6:28 everybody right literally everybody who

6:31 is not and so this now don't have to log

6:34 into the Miss dashboard Marvis is

6:36 sitting on your team's Channels with you

6:39 and it could stay there it literally can

6:42 stay on that team's conversation and

6:44 then just keep answering questions as

6:46 you go through this you can troubleshoot

6:47 switching issues troubleshoot you know

6:50 wireless issues whatever summary you see

6:52 today on the conversational assistant in

6:54 the dashboard is coming to uh this

6:57 team's channel so super stoked about

6:59 this

7:00 you know the the little button that I

7:01 love is adding it to your team Marvis as

7:04 a member of your team so this is what

7:07 Bob and team have been working on so

7:09 thank you very much for all of that Bob

7:11 and the Innovations on AI

7:18 last year last year oh by the way we

7:21 have to give away some stuff so who is

7:23 celebrating a birthday today are closest

7:26 to today

7:28 who's got it stand up please just tell

7:30 me a name

7:32 literally are you celebrating a birthday

7:34 today you got to stand up you have five

7:36 people pointing at you

7:39 yesterday all right happy birthday and

7:42 on behalf of juniper I have a Wi-Fi

7:44 access point and a T-shirt and whatever

7:47 else you want happy birthday to you so

7:50 uh so

7:55 I think legally I can't say whatever

7:57 else you want but uh but you know we'll

8:00 erase that from the record yesterday

8:02 last time last year when we had this

8:05 event uh the best highest voted session

8:08 was the customer panel because this is

8:11 uh truly where you hear the stories

8:14 right the what the why the how our

8:17 customers have transformed their

8:19 networks and uh this one we actually

8:21 have mic set up there as well I'm going

8:23 to have a few questions but really it's

8:25 yours if you're new to Juniper uh if

8:27 you're new to mist if you are an

8:29 existing customer but like something

8:31 that one of the customers has said walk

8:33 up to a mic you know interrupt us this

8:36 is a conversation for all of us to have

8:38 that's why this is an intimate session

8:39 so without further Ado please uh bring

8:42 up my customer panel come on up

8:46 thank you

8:49 [Music]

8:51 foreign

8:52 [Music]