AI Skeptics: Is AI Going to Take Your Job?

0:00 [Music]

0:06 hi everyone i'm dr sally eaves chair of

0:09 global cyber trust at cf cyber and ceo

0:12 of aspirational futures a huge warm

0:14 welcome to this juniper networks ai

0:16 skeptics event is ai going to take my

0:19 job we're here to bust some of the myths

0:22 and miss some misinformation around ai

0:24 particularly focusing on implementations

0:26 within ai and networking to do so i'm in

0:29 great company so first of all i'm going

0:31 to ask our stellar panel just give a

0:33 brief intro into their role darryl can i

0:35 start with you first yeah hi sally great

0:37 to be here um my name's daryl i work for

0:40 aston martin as the league network

0:42 architect and the team leader i'm in

0:45 charge of all the wired and wireless

0:47 connectivity right the way from your

0:49 desk all the way up to the main main

0:51 site so uh yeah that's me brilliant

0:54 stuff thank you baby sheamus over to you

0:56 next

0:56 hi my name is james mcgillicuddy i'm the

0:58 vice president of research and

1:00 enterprise management associates an

1:01 analyst firm and based in the united

1:03 states and

1:05 i've i do a lot of market research and

1:08 consulting

1:09 on

1:10 network infrastructure and operations

1:12 teams how they design build and operate

1:14 their networks brilliant thank you so

1:16 much for joining us today and next week

1:17 to you tom

1:19 my name is tom hollingsworth and i am an

1:21 independent analyst in the networking

1:23 community i work with a lot of other

1:25 independent analysts through an event

1:27 series called tech field day

1:29 but i do a lot of research into emerging

1:31 technologies such as artificial

1:33 intelligence you can see some of my

1:35 thoughts on that subject at my blog at

1:37 networkingnerd.net fantastic that's

1:39 brilliant i've attended one of those

1:40 days as well really really enjoyable so

1:42 thank you for that so perhaps we can get

1:44 straight in now to some of our key

1:46 questions and maybe let's start first

1:48 was the fundamentals so looking at what

1:50 ai is what jobs it can do within it and

1:53 networking and changing changing the

1:55 work that we do and has it changed you

1:57 and your roles and your organizations so

1:59 far now for me i'll come at this quickly

2:01 from two hats so one is a cto by

2:04 practice but also as a change manager as

2:06 well so particularly in telecoms i've

2:07 seen this in kind of three main pillars

2:09 so optimizing user experiences

2:12 simplifying operations and also this

2:13 move from reactive to more active

2:15 intelligence and end-to-end insights if

2:17 you will so i'd love to kind of go

2:18 around the table and kind of explain

2:20 those key fundamentals to your audience

2:22 today perhaps gerald can i go to you

2:24 first

2:25 yeah sure um so ai for us we're still

2:27 quite new into evaluating ai and we're

2:30 starting to deploy that into our

2:31 infrastructure

2:33 ai for us is not about kind of replacing

2:36 anyone it's about and this is a great

2:38 quote i heard the other day in a

2:39 training course ai is not about

2:42 replacing humans with robots it's about

2:45 taking the robot out of the human um and

2:48 it leaves you to get on with things like

2:50 strategy and high level things that the

2:52 business can can value i i use the

2:55 analogy of uh of a car then we we have

2:58 car sat navs um satnav's never replaced

3:02 the driver it just gave you a tool to be

3:05 a better uh more alert driver um so i

3:09 think that's uh that's how we should

3:11 view ai at the moment so yeah i love it

3:13 it's that complimentary partnership

3:14 coming to the forum what you were

3:16 discussing there i love that i love it a

3:17 good analogy that's great daryl thank

3:19 you and tom could i ask for your

3:20 perspective well i actually want to jump

3:22 on daryl's analogy there because one of

3:24 the things that i think that ai has been

3:26 very critical in doing is helping us

3:29 understand

3:30 how to modify systems and how to kind of

3:33 roll with the punches if you will part

3:35 of what we have spent so many years

3:37 doing in the networking space is kind of

3:40 dealing with the little fires that pop

3:42 up and if you think of it kind of using

3:45 daryl's analogy as a sat nav

3:47 i remember the days when satnav was

3:49 effectively going online

3:51 and downloading a list of directions

3:53 from a website like mapquest and lord

3:55 help you if you took the wrong turn at

3:57 step four because from that point

3:59 forward you were completely lost but a

4:02 sat nav not only gives you the overview

4:04 of what you're trying to accomplish but

4:05 it can course correct in the middle of

4:07 your journey so if for example there is

4:10 a car accident that you must

4:12 route around or in modern sat nav with

4:16 capabilities to include more

4:18 intelligence if there is a congested

4:21 route that has become congested since

4:23 the start of your journey it can

4:25 automatically reroute you based on

4:27 certain factors like least amount of

4:29 time traveled or

4:30 whether or not you're going to have to

4:31 pay a toll on the road so it gives you

4:33 this capability to be smarter about what

4:36 you're doing instead of just relying on

4:38 the same old tried and true directions

4:40 that we've been following for a number

4:42 of years i love us great example also

4:44 what i love there is the ability to

4:46 personalize to what matters to you

4:48 whether that's that cost of the toll or

4:50 the shorter journey or environmental

4:52 impact even as well it's a great example

4:54 and james did you have any final

4:55 thoughts on that yourself

4:57 yeah sure um one thing

4:59 from my perspective one thing i see is a

5:01 lot of network operations tools and the

5:03 network monitoring network performance

5:05 management tools they excel at uh

5:07 presenting data to

5:10 admins engineers

5:12 they don't always

5:13 do a very good job of providing insights

5:15 they they point you to indicators of a

5:18 problem don't really tell you what the

5:19 problem might be

5:21 all the time

5:23 ai technologies are

5:26 sort of providing

5:27 those first

5:29 intelligent insights into the data that

5:31 those tools

5:33 typically provide

5:34 absolutely excellent thank you so much

5:36 for sharing that so let's go back to the

5:38 heart of our kind of question our title

5:40 for our event today is ai going to take

5:42 my job let's drill into that for the

5:44 audience because again it's something

5:46 that's grabbed a lot of headlines over

5:47 recent years although i think the

5:48 narrative is changing certainly for me

5:50 it's not about this replacement it's

5:52 that complementary strengths i mentioned

5:53 at top and also robots kind of taking

5:56 things out of humans for example and you

5:58 know making more time for meaningful

5:59 activity so shame if i could go back to

6:01 you and some of your research in this

6:02 area yes um i uh did some market

6:06 research on

6:07 network operations team interest in uh

6:10 ai uh last summer we surveyed a few

6:13 hundred

6:14 people who are actively engaged in using

6:17 this technology there wasn't a lot of

6:19 concern about

6:21 losing your job uh

6:23 there's a lot of

6:25 there was a sense that this can deliver

6:27 a lot of value to to make things run

6:30 better like one of the one of the early

6:32 use cases we were seeing uh

6:34 with a lot of traction was just the idea

6:37 of

6:38 more intelligent alerting and

6:39 escalations of events so that um

6:44 the alerts um connect the dots uh you

6:47 know there's a lot of contextualization

6:49 of data in those alerts um

6:52 more so than you may have seen in the

6:53 past and and also

6:55 uh the system gets smarter about routing

6:58 it to the right person that has the

6:59 expertise to take action on those alerts

7:02 although

7:03 when the alerts get uh smarter uh and

7:06 more contextualized uh people with less

7:08 sophisticated engineering skills can

7:09 actually take action more often so it

7:12 empowers people with with lower skills a

7:14 lot of the time that typically would

7:16 have just escalated something to a tier

7:18 three engineer rather than trying to

7:20 deal with it themselves so that's one

7:22 big thing that we saw in our research

7:25 fantastic really interesting i love that

7:26 one around that democratization of the

7:28 ability to take agency and to take

7:30 action um to more and more roles that's

7:32 really really interesting thank you and

7:34 daryl what are you seeing

7:36 yeah so um i can take that a step

7:38 further and and go on from what shane's

7:40 just saying there in some real world uh

7:42 uh examples there we've been deploying

7:44 and testing the the ai platform at the

7:47 moment um and it's um it's starting to

7:49 identify things that we might not have

7:51 known about and one particular one is uh

7:54 bad cable about network cable now

7:57 historically you wouldn't have

7:59 necessarily known that straight away um

8:01 it might manifest itself in a site is

8:05 you know poor performance or something

8:07 um but in order to identify that you

8:09 need to trawl through an awful dollar

8:11 logs and look at errors on interfaces

8:14 and and things like this so ai can pull

8:17 all that log files and alerting all into

8:20 one place and say hey you know what i i

8:23 think this is a bad cable on this

8:24 particular interface here um you might

8:27 want to look at this and so it goes back

8:29 to that is it going to replace me and

8:31 the answer to that is is absolutely no

8:34 you you find yourself between two

8:36 extremes at the moment and this is where

8:38 i see things going on the one extreme

8:40 you've got oh we're getting an alert for

8:42 bad cable we'll ignore it because people

8:45 turn off their pcs all the time on the

8:48 other extreme we'll get oh it's a bad

8:50 cable i'm going to go and replace it

8:52 because the ai told me to that cable

8:54 might be the uplink for the site and so

8:57 it's kind of like an extra pair of hands

8:58 just guiding you in that direction to

9:00 say hey you know what you might want to

9:02 go and have a look at this and that

9:04 saves you an awful lot of

9:05 troubleshooting because it's already

9:07 correlated that uh that data so at the

9:10 moment we're seeing ourselves in in that

9:12 balance and um it's definitely going the

9:13 right way so it's really promising at

9:15 the moment

9:16 i love that almost takes us back to that

9:18 analogy and sat nav earlier doesn't it

9:20 you've got a myriad of paths and it's

9:21 filtering out that noise into what's the

9:23 optimum direction and path to take so i

9:25 really like that excellent and tom what

9:26 are you seeing there particularly around

9:28 that kind of single source of truth

9:29 aspect i think it's really interesting

9:31 well it's interesting that a lot of

9:33 people have have looked at this ideal of

9:35 having a single source of truth for your

9:37 network but if you asked most engineers

9:39 or operations teams today what is the

9:42 state of your network they would go to

9:44 some kind of a design document and say

9:46 oh well this is what it looks like and

9:48 my first question when someone does that

9:50 is okay now what does the network

9:52 actually look like

9:53 and that usually throws them for a loop

9:54 because well no no it's supposed to be

9:56 like this and i'm like

9:57 von mulkey said that no plan survives

9:59 contact with the enemy well no network

10:01 configuration survives implementation

10:03 and so there's even if there's a little

10:05 bit of a disconnect between what you

10:07 intended to put out and what was

10:09 actually configured those two

10:11 disconnects create two sources of truth

10:14 desired state and actual state

10:17 and nobody wants to do the investigation

10:20 work to find out what actual state is

10:22 there's a quote that can be attributed

10:23 to kurt vonnegut it's a flawed human

10:25 nature that everybody wants to build but

10:26 nobody wants to do maintenance and

10:28 that's effectively what this is is going

10:30 out to figure out what's going on so

10:32 what you can use an ai platform to do is

10:35 to find out what current state is

10:37 implementation

10:38 and compare it to desired state and to

10:41 try to find out where the disconnect

10:43 happened in those two was it

10:45 unintentional i copied a configuration

10:47 to a switch or i had some kind of a

10:49 staging server that had an error in the

10:50 middle of deployment and now suddenly

10:53 the switch thinks it's the spanning tree

10:54 root for the whole network or was it

10:56 intentional where you have someone who

10:58 has a lot more skill than they think

11:02 they do

11:03 and says oh well i know that every time

11:04 i try to put this on that particular

11:06 model of switch it fails so i'm going to

11:08 modify this command just a little bit so

11:10 that what actually goes under the switch

11:11 works

11:12 even if the intention was pure what

11:15 you've created is a disconnect in state

11:17 where a command later on or some kind of

11:19 a desired uh deployment could fail

11:22 because of something you had no clue

11:24 about what happened because of the

11:26 disconnect between those two and

11:28 part of the reason why humans don't like

11:30 this is because maintenance is boring

11:32 it's mind-numbing

11:33 and an ai doesn't sleep it doesn't get

11:37 tired it will not stop until it does

11:39 exactly what it's been told to do

11:41 and what it returns is the truth as it

11:44 sees it it is not trying to save its job

11:47 it is not trying to get a promotion it's

11:49 not trying to make somebody else look

11:50 good it is telling you what it found so

11:53 you can trust it you may not inherently

11:56 believe it but you can trust that the

11:58 answers that it's returned are accurate

12:01 i love that it's a really really good

12:02 example there and sheamus can i bring

12:04 you back in as well because i know we

12:06 spoke previously you were drilling in a

12:08 bit more as about how it helps find out

12:10 the how you know how things are broken

12:12 for example not just how to fix

12:13 something that's a really interesting

12:14 take as well

12:15 i remember having a conversation with um

12:18 a network operations manager a couple

12:21 years ago who he told me like he might

12:23 have 7 000

12:25 interfaces down on his network he only

12:28 cares about the ones that matter um

12:30 if there's applications that are trying

12:32 to run over there's interfaces he wants

12:33 to fix somebody it points to the fact

12:35 that um a lot of

12:37 tools that network operations teams use

12:39 do not provide actionable insights uh

12:41 typical enterprise

12:43 tells in our research a typical net ops

12:46 team less than half of alerts being

12:49 pumped out by their tools are actionable

12:51 like indicators of actual problems and

12:53 so there's a lot of noise that needs to

12:55 be sorted through first before you even

12:57 try to fix something to figure out if

12:59 it's actually

13:00 broken if there's actually something

13:02 broken that needs to be fixed

13:04 and then yeah once once you figure out

13:06 something's broken you need to know

13:08 how it broke uh not just how

13:11 how to fix it um

13:13 because you don't want it to break again

13:14 um

13:15 so

13:17 you know you have to be thinking about

13:18 proactive problem prevention and

13:20 optimization as you're troubleshooting

13:22 ai can sort of help you do that because

13:24 ai kind of connects the dots um look at

13:26 how the network is looking at how the

13:28 network is serving the business and

13:30 how to optimize that it will open up the

13:32 opportunity for more problem solving

13:35 that that improves how it improves the

13:37 network's ability to serve the business

13:39 not just fighting fires but but um

13:41 preventing those fires from ever

13:44 happening over the last like 10 years

13:46 our research has shown that

13:49 between 30

13:51 30 and 40 of all it service degradations

13:55 are detected by end users before it

13:57 operations is aware of them

13:59 that means they're reporting them to the

14:01 help desk and they're not being not

14:03 being productive because of those

14:05 problems

14:06 to some extent

14:07 before net ops is even responding uh you

14:11 need to you need to turn that around you

14:12 need to get that closer to as close to

14:14 zero as possible i think

14:16 ai can can help you get there to some

14:18 extent

14:20 absolutely there's a key point there i

14:21 think two pillars of what you were

14:23 saying really struck me a that move to

14:25 be more proactive as you were just

14:26 talking about there that active

14:28 intelligence i said at the top but also

14:30 the ability to filter out the noise and

14:32 one thing i read some research i've been

14:34 involved in it was looking at for

14:35 example burnout in a lot of operations

14:37 team and that was one of the biggest

14:38 contributors to that and also to churn

14:40 as well so again making a huge

14:42 difference in that respect as well so

14:44 really interesting points all around

14:45 there fantastic

14:46 and perhaps we can now drill into that

14:48 impact point so around your

14:50 organizations what have you seen ai

14:52 delivery in particular how have you been

14:54 able to demonstrate that value across

14:55 the organization and what other

14:57 implications is this brought about

14:58 particularly around skills and education

15:00 which implicit first acknowledged that's

15:02 a real strong point area for me i really

15:04 love looking at that area too so perhaps

15:06 we can bring all those points together

15:07 look at the impact and the implications

15:09 and perhaps shameless if i could go back

15:10 to you first well as you're bringing ai

15:12 in the organization uh one thing that

15:16 you're going to see is um

15:19 many

15:20 people in your your it ops organization

15:23 are

15:24 not confident in their ability to

15:25 evaluate ai they have multiple vendors

15:28 telling them that they've developed ai

15:30 technology machine learning technology

15:32 that can transform operations they don't

15:34 know how to

15:36 determine whether

15:38 what they're being sold is is just a

15:40 bill of goods or not

15:42 um

15:43 they also want to know that there's a

15:46 variety

15:47 of um of network data that's being used

15:50 to train those algorithms like are the

15:52 is the is the training that that this

15:55 vendor

15:56 does with its algorithms going to you

15:59 know be representative of what i see in

16:00 my network like are they are they are

16:03 they training this in a realistic way so

16:04 it's actually going to understand

16:05 networking um so that's uh something

16:08 that we've seen uh early on

16:11 those are the types of things that they

16:12 want to see i've also heard people say

16:14 you know i want to just get past the

16:15 like people specifically told me like

16:17 architects and

16:18 fortune 500 companies

16:20 fortune 100 companies

16:22 who say there's a lot of marketing

16:25 to speak a lot of buzzwords going around

16:28 they just want to know how you're

16:30 solving problems like how you solving

16:31 the problems that i have can you show me

16:33 examples of that so that i know like

16:36 what this can do that's something that

16:38 they want to be able to take back to

16:39 their organization be like hey look what

16:40 i saw you know this is this i think this

16:42 is something that we could use to really

16:44 uh change the way we we manage our

16:46 network here

16:48 at those tangibles those examples so so

16:50 important make it relatable to that

16:52 particular sector and that clarification

16:54 you're spot-on around language i think

16:56 zero trust would be a classic example of

16:58 that uh doing something on that earlier

17:00 on today and again that's somewhere

17:01 where there's quite a lot of confusion

17:02 so getting that right and making sure

17:04 there's not misinformation absolutely

17:06 critical as well so i love that tangible

17:08 focus and tom what do you see in your

17:10 area

17:11 so a lot of the unease about deploying

17:13 ai within an organization comes from two

17:15 different areas the first is that it's

17:17 going to take my job away and and we've

17:19 kind of talked a little bit about that

17:20 throughout this whole thing but more

17:22 importantly there are a lot of people

17:24 that don't feel comfortable with the

17:25 idea that an ai is going to be able to

17:28 determine what the source of your

17:29 problem is instantaneously and give you

17:31 a recommended fix it kind of eliminates

17:34 that that kind of feeling that we get

17:36 when we we figure out a really hard

17:37 problem like working in sudoku or trying

17:40 to figure out a puzzle and we get that

17:41 rush of endorphins when hey this is the

17:43 the secret and i figured something out

17:46 to combat that second problem um you

17:49 kind of have to get into a mode where

17:51 you're you're communicating more

17:52 effectively and you're using ai as a

17:55 tool to do that how many times has your

17:58 average engineer ops person kind of been

18:00 under the crunch because there's a huge

18:02 problem that we need to fix and the

18:03 executives are frustrated and something

18:05 needs to happen and you're sitting here

18:07 literally racking your brain trying to

18:09 find out what the source of the problem

18:10 can be eliminating choices one at a time

18:14 to them you look like an ai because you

18:16 are doing the same kind of learning and

18:18 analysis that an ai would do to spit out

18:21 an answer they trust you because you're

18:23 a living breathing person we should

18:25 trust ai in the same way we should still

18:27 communicate we should still tell people

18:29 what happened and and more importantly

18:31 even if ai is the solution that

18:34 solve that problem we do need to tell

18:36 people that ai came up with that um we

18:38 in my old job as a managed service

18:40 provider we used to have a saying on the

18:42 the wall if you didn't tell your

18:43 customers that you fixed something did

18:45 you actually fix anything in a way users

18:47 need the same kind of reassurance inside

18:49 of your organization

18:51 but the other thing that is very

18:52 important is that people need to look at

18:53 ai not as a way for them to have their

18:56 jobs taken away

18:57 but it will remove tasks from your jobs

19:00 kind of like daryl said we're taking the

19:01 robot out of the person think about all

19:03 the things that you do on an average day

19:05 that are just repetitive and boring and

19:07 mind-numbing but necessary that having

19:11 that taken off of your plate having that

19:12 given to something that will not stop

19:15 will always accomplish the results and

19:16 send you a report telling you what it

19:18 did frees you up to think about other

19:21 things to spend more time building and

19:23 imagining and creating instead of doing

19:25 maintenance and i think that that

19:27 framing helps people understand you're

19:30 not going to lose your job in fact

19:31 you're probably going to get effective a

19:33 promotion to something more exciting and

19:36 new

19:37 and that's a good thing overall

19:39 absolutely it's that changing of the

19:41 narrative isn't it and the word

19:42 enablement is kind of ringing through my

19:43 eyes when we were describing these

19:45 different aspects around around this

19:46 today so couldn't agree more strongly

19:48 tom absolutely daryl did you have any

19:50 final thoughts on this around that value

19:53 also the skills implications of this

19:54 piece too sure yeah i mean we're

19:57 starting off quite small with ai at the

20:00 moment we are a small team and so we

20:02 can't bring in something in big bang

20:04 because you know there's a there's only

20:06 there's only three of us now um so we're

20:08 starting off very small where we see a

20:10 little bit of not so much resistance but

20:12 um our operations teams uh our service

20:15 desk network operations and so on um

20:18 they don't um they don't dislike ai they

20:21 think you know everyone kind of accepts

20:23 okay this is a great tool where we do

20:26 find um maybe a little bit of pushback

20:28 is with management and they say oh hang

20:30 on a minute is this is this just another

20:32 one of the toys that it's buying um so

20:35 we kind of say well actually no look

20:37 what we can do

20:39 before i had to count you know how many

20:42 mac addresses were on the network or how

20:44 many ip addresses were in use now i can

20:47 just ask

20:48 a high-level like language question like

20:51 how many users are on this site and it

20:54 gives me an answer and so management is

20:56 sort of seeing this and i'm demoing this

20:58 very informally it's almost like a

21:01 oh hey while you're here just have a

21:03 look at this you know it's just kind of

21:04 just as somebody passes by and uh and

21:07 we're seeing a lot of uh value from that

21:09 so um just having that very small um

21:11 focused uh uh try not so much trial but

21:15 um small focused deployment and growing

21:17 from there is uh it's hugely beneficial

21:20 yeah i really like that i like the fact

21:21 you're bringing to the floor as well

21:22 this could be an incremental change as

21:24 well so i love that that's really

21:26 important for a lot of businesses so

21:27 that's a great point thank you for

21:28 sharing that and with my telco hat on as

21:30 well certainly things that i've seen up

21:32 close and personal is around you know

21:33 faster problem solution that resolution

21:36 um fewer on-site visits for example as

21:38 well so you've got that big operator

21:39 benefit and equally from the consumer

21:41 you've got a better reliability

21:42 measurability huge point as well

21:44 unpredictability too so again that kind

21:46 of shared value proposition through ai

21:49 so perhaps as we start to bring all

21:50 these different elements to a close we

21:52 do a bit of a round table again and kind

21:54 of just bring each of you like a closing

21:56 thought to take away to share with the

21:58 audience it could be something that's

21:59 come out from this conversation or kind

22:00 of a top tip that might be relevant for

22:02 them wherever they are in their journey

22:04 towards implementing ai at the moment

22:05 and sheamus got to go for you first

22:08 one thing is um

22:09 most network operations teams have too

22:12 many tools um

22:14 they're using 10 15 20 tools monitor and

22:17 troubleshoot their infrastructure

22:20 ai

22:21 can

22:22 consolidate that i think quite a quite a

22:25 bit they can make some tools less

22:26 relevant they may not go away but that

22:28 you know the insight you're getting out

22:29 of the ai will

22:31 will lead to your people spending less

22:33 time going from tool to tool to tool to

22:35 get the answers they need because the

22:36 answers are in the ai

22:38 solution um depending on where it lives

22:40 i've heard from people saying i'd love a

22:42 ai solution just pulls all the data from

22:44 all my tools and presents it as a single

22:46 view with insight um

22:48 i think and once

22:50 once you see that um you see the value

22:53 um i also find you know people spend a

22:56 lot of time

22:58 just generating reports you know like

23:00 hey we got to audit this we need to know

23:02 what you know what what's what

23:03 is everything in compliance you know

23:05 they're golden configs

23:06 like i talked to people recently who who

23:09 devote 10 15 20 hours a week

23:11 generating reports

23:13 uh

23:14 i think

23:15 an intelligent solution can can optimize

23:17 that so

23:19 if you feel when you're doing stuff like

23:21 that like swiveling from tool to tool

23:23 the tool or generating reports over and

23:25 over again

23:26 not really delivering a lot of value to

23:27 the business like you're spending most

23:29 of your time just

23:31 doing repetitive tasks as everyone's

23:33 been talking about here you're not

23:34 demonstrating your value and when i'm

23:36 trying to understand the return on

23:38 investment in the tool like an ai

23:40 solution it's about

23:42 freeing up labor and allowing that labor

23:44 to do something that's more important to

23:46 the business

23:47 and that's what i think people should be

23:49 thinking about as they're looking at ai

23:51 is like how do i take my people who are

23:54 really knowledgeable and skilled and

23:56 empower them to support transformation

23:59 of our business not not keeping the

24:00 lights on but like you know spreading

24:02 the light elsewhere

24:04 i really like that bring up that

24:05 capacity for higher order work against

24:07 that changing of the narrative isn't it

24:09 so so important thank you so much and

24:11 tom over to you for your closing

24:12 thoughts

24:13 i really like a lot of the things that

24:14 sheamus said and i think that one of the

24:16 reasons why we kind of have this tool

24:20 aversion is because the tool is the

24:22 solution to whatever problem we have

24:24 so every time we come into a new issue

24:26 that we need to solve we we just go buy

24:29 something and make the problem go away

24:31 for a little while longer

24:32 and i think that people are looking at

24:34 ai like that oh great this is just

24:36 another tool that goes into the box that

24:37 i never use

24:39 change the thinking about what you're

24:40 looking to use ai for you're not buying

24:43 a tool you're hiring someone no it may

24:46 not be a flesh and blood person but you

24:48 are going to invest training in that

24:49 person you are going to give them tasks

24:52 and job roles that will multiply the

24:55 ability of your other

24:57 assets to get things done if you look at

25:00 it that way if you look at it as you are

25:03 hiring someone then you'll appropriately

25:05 invest in that person you'll give them

25:08 the capabilities they need to accomplish

25:10 the tasks and roles and responsibilities

25:12 and kind of like sheamus said then that

25:14 elevates the rest of the team because

25:16 now they're not spending hours of their

25:18 day tracing down log messages or

25:21 generating reports for the executives i

25:24 mean with a properly configured ai the

25:26 executives can request the report

25:27 whenever they like to see it and they

25:29 can verify that the information that's

25:31 in the report is good information that

25:34 they wanted and those are the kinds of

25:36 things that give you the capability to

25:39 spend more time doing more impactful

25:41 things generating revenue being more of

25:44 an asset to the company in the bottom

25:46 line as opposed to just doing the

25:48 regular maintenance task that nobody

25:50 wants to do

25:52 absolutely very well seth thank you so

25:54 much and daryl over to you for final

25:56 thoughts

25:57 great thanks a lot i can carry on from

25:58 what tom is saying there really i mean

26:00 i'm always telling the uh the operations

26:02 uh staff that uh progress is progress

26:05 whether you're fixing a vault or

26:07 deploying something progress is progress

26:10 sometimes it's small progress sometimes

26:11 it's big ai is not going to fix

26:14 everything all at once

26:16 and so i say

26:18 on tom's note there about think of it as

26:21 hiring another person it's higher in

26:22 that pair of hands to just progress your

26:24 fault along the right lines you don't

26:27 you're not throwing it all against the

26:28 wall and seeing what sticks ai really

26:31 focuses you and it just progresses you

26:33 forward in that incremental step that we

26:34 talked about earlier i don't mind if a

26:36 fault takes a long time to fix as long

26:38 as there's progress in that in that way

26:41 so then i can report to management and

26:43 say

26:44 it is taking a while but we're seeing

26:45 this progress and i think that's

26:46 definitely what ai is bringing us at the

26:48 moment

26:49 absolutely fantastic i think we brought

26:50 in you know a lot of themes for the four

26:52 here you don't need to know everything

26:54 you know all the plumbing so to speak

26:56 it's how you can work with ai it's

26:58 changing this narrative is this

26:59 complementary strengths aspect and going

27:02 from menial tasks to the meaningful how

27:04 we can get that higher order works i

27:06 think so many takeaways here a lot of

27:08 practical examples as well how we're

27:10 seeing organizations at different stages

27:12 using this within itn network so

27:14 fantastic thank you all three of you so

27:16 shameless tom darrell thank you so much

27:18 for joining us today and thank you all

27:20 for joining our webinar as well it's

27:21 been fantastic and a lovely discussion i

27:23 think one we can continue

27:25 great thank you very much

27:27 [Music]