Ep 20: AI Literacy and Responsible Innovation with Daan van Rossum

[00:00:00] [SPEAKER_00]: Hey, you, with the podcast in the eye.

[00:00:02] [SPEAKER_00]: But, of course.

[00:00:03] [SPEAKER_00]: Has your mobile HPA been activated?

[00:00:06] [SPEAKER_00]: You have a single long-term long-term limit to the state volume.

[00:00:10] [SPEAKER_00]: It's always great.

[00:00:12] [SPEAKER_00]: And for all these mobile functions for the internet to be available for the telecom.

[00:00:16] [SPEAKER_00]: Now, you can see one hour.

[00:00:18] [SPEAKER_00]: The mobile HPA in the Main Magenta App and already goes up.

[00:00:22] [SPEAKER_00]: In 5th genet, the Telecom.

[00:00:24] [SPEAKER_00]: Not yet.

[00:00:25] [SPEAKER_00]: Now, to activate the Main Magenta App.

[00:00:39] [SPEAKER_03]: Hey, it's Bob. In this episode I am joined by Don Van Rossum, who is the CEO and co-founder

[00:00:45] [SPEAKER_03]: of Flex OS, most of the lead with AI podcast, a writer, speaker, AI expert, and a LinkedIn

[00:00:52] [SPEAKER_03]: top voice for 2024. Don and I have an insightful discussion on the future of work and

[00:00:58] [SPEAKER_03]: the role of AI in shaping it. We explore the challenges and opportunities of AI adoption

[00:01:02] [SPEAKER_03]: in the workplace, the importance of employee autonomy, and the need for a responsible

[00:01:07] [SPEAKER_03]: AI integration. Don is up to going people on a regular basis with his lead with AI coaching

[00:01:12] [SPEAKER_03]: program, so stick around for his words of wisdom to elevate your AI cue and that of your

[00:01:16] [SPEAKER_03]: team. Hope you enjoyed this discussion and thanks for tuning in.

[00:01:22] [SPEAKER_04]: Hello everyone, welcome to another episode of Elevator. Your AI cue. I'm your host Bob

[00:01:27] [SPEAKER_04]: Oliver with me today is Don Van Rossum, is the CEO and co-founder of Flex OS. Don

[00:01:34] [SPEAKER_04]: I will let you give you proper introduction to what it is you do and what Flex OS

[00:01:41] [SPEAKER_04]: is all about and we'll get into it because I have many questions.

[00:01:44] [SPEAKER_02]: Okay, cool. Yes, happy to do that. Bob, thanks so much for having me on. So I am done

[00:01:50] [SPEAKER_02]: that that part you've got correct and I do have a company called Flex OS, so so far so

[00:01:53] [SPEAKER_02]: good without making it too long of a story dating you know 40 years back. I've been working

[00:02:01] [SPEAKER_02]: for quite a while, I think this is like year 25 in the workforce and I've had really

[00:02:07] [SPEAKER_02]: good work experiences and I've had not so good work experiences but I saw the potential

[00:02:12] [SPEAKER_02]: of having a place that you love working and that you love coming to and with good colleagues

[00:02:17] [SPEAKER_02]: and doing interesting work and meaningful work and then I heard from a lot of people that

[00:02:23] [SPEAKER_02]: don't have that and so the mission of the company was born, you know, long before the

[00:02:28] [SPEAKER_02]: company was born which is to create a happier future of work and to really look at like

[00:02:32] [SPEAKER_02]: how can we give people the opportunity to have a great work life in terms of like the kind

[00:02:38] [SPEAKER_02]: of work that they do, where they work, when they work, how they work and obviously the

[00:02:44] [SPEAKER_02]: pandemic has been an excellent for some of that like especially in the where we work

[00:02:49] [SPEAKER_02]: that has changed a lot in the last couple of years although companies are eager to go back

[00:02:54] [SPEAKER_02]: to prehistoric times or pre-pandemic times but but but some things have changed for

[00:03:00] [SPEAKER_02]: for the better and some things I think were still really early on especially obviously

[00:03:04] [SPEAKER_02]: in AI which is our mutual topic and so the mission carries on and we do it with a new

[00:03:11] [SPEAKER_02]: slatter, a podcast just like you. We have a team writing editorial content really focused

[00:03:17] [SPEAKER_02]: on the bigger topics and then we're doing some training and consulting as well.

[00:03:22] [SPEAKER_04]: Yeah, that's fantastic. One of the things that caught my attention in addition to your

[00:03:28] [SPEAKER_04]: courses and some of the great guests you've had on your show, you cover a lot in terms

[00:03:35] [SPEAKER_04]: of evaluating, you know, reviewing different solutions so I was curious what kind of

[00:03:41] [SPEAKER_04]: goes into that. I did spend some time as a market research analyst evaluating certain

[00:03:46] [SPEAKER_04]: solutions in the technology space, certainly I've spent time on G2 and other sites that

[00:03:55] [SPEAKER_04]: maybe take a more qualitative view as opposed to putting things into a grid or giving

[00:04:02] [SPEAKER_04]: solutions a new mayor score or things like that. So it seems like you cover a tremendous

[00:04:09] [SPEAKER_04]: amount of ground and I was just curious how big your team is or do you crowdsource some

[00:04:15] [SPEAKER_04]: of those reviews and how does that process work?

[00:04:20] [SPEAKER_02]: I would say all of the above, so we have a human team, we have freelancers and then we have

[00:04:26] [SPEAKER_02]: AI obviously doing quite a bit of the work. You know, I think you hit the nail on the

[00:04:31] [SPEAKER_02]: head with the G2 conundrum which is that if I'm actually looking for something and

[00:04:35] [SPEAKER_02]: I have to go through a list of 6,000 options with a lot of advertisements in between it's

[00:04:40] [SPEAKER_02]: kind of hard to find what I'm looking for. And certain categories are complex I mean if

[00:04:45] [SPEAKER_02]: you look at HR tech there are a lot of categories where it's very hard to say, you know

[00:04:50] [SPEAKER_02]: what is actually the best solution because it depends on so many factors right and so

[00:04:53] [SPEAKER_02]: that's where the quantitative approach obviously helps. But for certain other things like

[00:04:58] [SPEAKER_02]: just understanding what the AI in recruiting landscape looked like or what the most used

[00:05:04] [SPEAKER_02]: AI tools for HR there is much more like qualitative approach you can take and so with

[00:05:10] [SPEAKER_02]: all of those three teams I would call them we work together to try and just get sort of

[00:05:17] [SPEAKER_02]: a truth out there in terms of what are the solutions people should look into and I think you know

[00:05:23] [SPEAKER_02]: we do it in pretty traditional categories I would say like HR, I ask an HR software,

[00:05:28] [SPEAKER_02]: employee management software. I think obviously it gets a lot more interesting when you get into

[00:05:32] [SPEAKER_02]: the newer categories that maybe didn't even really exist two years ago like AI recruiting software

[00:05:38] [SPEAKER_02]: and so I think that's where we really shine in really keeping a finger on the pulse in terms of like

[00:05:44] [SPEAKER_02]: what kind of companies are out there you know what is worth looking into for people who are

[00:05:48] [SPEAKER_02]: looking for solutions and also giving them an opportunity to be spotlighted towards that audience

[00:05:54] [SPEAKER_02]: and obviously we are both part of the people the TPP community. So there's like often a lot of discussions

[00:05:59] [SPEAKER_02]: around hey I'm looking for days what would you recommend and typically those recommendations from

[00:06:04] [SPEAKER_02]: people you trust are a lot better than just some directory with like 6,000 list things and so we're

[00:06:09] [SPEAKER_02]: trying to be a little bit like the skilled version of that word still publicly accessible but it is

[00:06:13] [SPEAKER_02]: definitely a curated look into what is out there. I think it was Glenn Kathy who put us pull out this

[00:06:21] [SPEAKER_04]: morning on LinkedIn around why aren't companies investing you know more in you know January

[00:06:27] [SPEAKER_04]: AI tools and my response was well I hope it's just because they have an actually figured out

[00:06:33] [SPEAKER_04]: what their strategy is and not because they don't think it's actually worth it. It's probably

[00:06:38] [SPEAKER_04]: about if you think about historically where you know people have been complaining forever that

[00:06:44] [SPEAKER_04]: they haven't invested enough in in L&D or it's not personalized enough for you know things like that

[00:06:50] [SPEAKER_04]: and so you're right that's definitely part of it but it just seems like unless you're a Microsoft

[00:06:56] [SPEAKER_04]: shop maybe like real pilot is on everyone's devices maybe you haven't fully enabled all the

[00:07:04] [SPEAKER_04]: all the copilates and I'm not even sure like how the architecture looks these days but

[00:07:09] [SPEAKER_04]: it just seems like and now with apples recent announcements you know that I always say

[00:07:15] [SPEAKER_04]: teen is going to have all these AI features and they're going to have you know onboard AI

[00:07:20] [SPEAKER_04]: custom AI chips and all the stuff and it's like in people's personal and professional lives

[00:07:25] [SPEAKER_04]: they are now inundated with AI so if you're going to now choose to not help guide them

[00:07:32] [SPEAKER_04]: in any way I mean that is like one of the most short-sighted who you're talking the foot kind of

[00:07:38] [SPEAKER_04]: you know situations I've ever heard of I mean you've got to figure it out and you've got to do so

[00:07:44] [SPEAKER_02]: you know quickly yeah absolutely yeah okay so like even take the example of Microsoft right so let's say

[00:07:50] [SPEAKER_02]: your company you're already on Microsoft and then and I've been in these discussions right so

[00:07:56] [SPEAKER_02]: you're already on Microsoft so therefore if everyone is already using their productivity software

[00:08:01] [SPEAKER_02]: when it comes to choosing your best option for AI it's probably going to be copilot even though

[00:08:07] [SPEAKER_02]: I don't think it compares to a lot of the tools out there but just for the sake of adding it onto

[00:08:12] [SPEAKER_02]: something you're already using and it being able to tap into all your documents and all your

[00:08:16] [SPEAKER_02]: presentations and again dating years or decades back that's a pretty powerful proposition right

[00:08:22] [SPEAKER_02]: then a company is still looking at it's $30 per employee per month on top of your normal

[00:08:26] [SPEAKER_02]: Microsoft fees right that's a 360 dollars per employee per year because you have to pay them annual basis

[00:08:32] [SPEAKER_02]: and so if you're a company of 1,000 people you're looking at 400k out the door and you're not

[00:08:37] [SPEAKER_02]: investing that 400k with the idea of all maybe a year later we'll pull the plug on it right because it's

[00:08:42] [SPEAKER_02]: a 400k for software fees then all the onboarding costs and time all the change management right

[00:08:51] [SPEAKER_02]: because again we've seen out from the research if you give people the tool that's not enough

[00:08:56] [SPEAKER_02]: if you just give people copilot mostly probably going to be like what is this nonsense there's

[00:09:02] [SPEAKER_02]: like a newer version of creepy on my desktop what do I do with it that also goes into you know

[00:09:06] [SPEAKER_02]: your Alan D team and they have to really be in it so they have to go into Microsoft training

[00:09:10] [SPEAKER_02]: and they have to understand it at that level and then they have to develop training and then you're

[00:09:14] [SPEAKER_02]: working with Microsoft and you know that's just on that side right you have to go through security

[00:09:18] [SPEAKER_02]: audits like company I talked to has to go through two weeks security audits to make sure that all

[00:09:23] [SPEAKER_02]: those files on one drive are not just going to everyone because suddenly this file that no one

[00:09:27] [SPEAKER_02]: would have seen that had a certain security permission now is available to everyone and that's very

[00:09:33] [SPEAKER_02]: scary right in a big global company so just looking at that practical case it sounds easy of like oh you

[00:09:40] [SPEAKER_02]: know AI is everywhere and you should kind of embrace it which I totally agree but you know it's

[00:09:45] [SPEAKER_02]: such a huge bet to take out let's say it's 400 plus 400 right like almost like a million dollar

[00:09:51] [SPEAKER_02]: bets for a thousand person company and then you don't really know what's going to come out of that

[00:09:57] [SPEAKER_02]: right and so I think you see really great case studies right so we've seen them on during

[00:10:00] [SPEAKER_02]: that case study yesterday interview it's probably CTO at BCGX you know they have a 3000

[00:10:06] [SPEAKER_02]: person team building AI applications they're working with a lot of these like really big companies

[00:10:10] [SPEAKER_02]: and helping them make that transition by the way they're all on chat to PT enterprise but

[00:10:16] [SPEAKER_02]: it is a huge project with you know getting it wrong is pretty bad too and so I think the

[00:10:23] [SPEAKER_02]: recent Asana AI state of work report showed it very clearly that 90% of the companies are still

[00:10:29] [SPEAKER_02]: at the very first stages of adopting AI and if you kind of overlay that with the Microsoft data

[00:10:35] [SPEAKER_02]: which showed that 75% of people are doing the BYO AI bringing your own AI to work then you're getting

[00:10:42] [SPEAKER_02]: a really bad situation where the company is not moving fast enough because there is so much to

[00:10:47] [SPEAKER_02]: be decided the employees are leapfrogging immediately and they're bringing all their own tools in

[00:10:54] [SPEAKER_02]: and that means use cases are fragmented the data is going to all kinds of providers like

[00:11:00] [SPEAKER_02]: who knows what other is doing with all my meeting transcript data or read AI like do I know

[00:11:05] [SPEAKER_02]: a lot about that company and so that's part of what we're trying to do now is actually starting to

[00:11:09] [SPEAKER_02]: sit down and look into what are the data privacy and like issues and what are there where

[00:11:15] [SPEAKER_02]: are they storing that data was the kind of like who's behind this company because there's just

[00:11:19] [SPEAKER_02]: like a lot of question marks out there in terms of the companies that are running

[00:11:22] [SPEAKER_02]: AI tools that we see right now. We keep saying you know HR leaders here's your opportunity

[00:11:28] [SPEAKER_04]: don't wander it and don't you know relinquish control to you know IT or to your you know

[00:11:35] [SPEAKER_04]: cheap data officer or somebody else like you guys are in and the driver's seat you're you know

[00:11:41] [SPEAKER_04]: proverbial seat at the table whatever but I just wonder you know at our general prizes

[00:11:46] [SPEAKER_01]: I feel like the technology leadership would really love to just before we move on I need to

[00:11:58] [SPEAKER_01]: if you're looking for the latest on product development marketing funding big deals happening

[00:12:04] [SPEAKER_01]: in talent acquisition HR HCM that's the show you need to listen to go to the work to find

[00:12:12] [SPEAKER_01]: network search out people tech markfever you can find them anywhere. Have fewer roads to choke as they

[00:12:21] [SPEAKER_04]: say when it comes to the vendor you know landscape right so you're right I mean co-pilot

[00:12:27] [SPEAKER_04]: the anecdotal you know feedback I get is it's just like all their other it's finally actually

[00:12:33] [SPEAKER_04]: they're not gonna prioritize user experience or you know some extra you know bells and whistles

[00:12:40] [SPEAKER_04]: at the risk of a driver privacy you know audits that are gonna keep coming and the whole slew of

[00:12:50] [SPEAKER_04]: risks that enterprise play yeah you'll have the shadow IT problem probably worse than ever

[00:12:56] [SPEAKER_04]: when it comes to these tools because if you don't like the answer you get from

[00:13:00] [SPEAKER_04]: oh pilot what are you gonna do you're gonna hop on your phone and you're gonna ask

[00:13:04] [SPEAKER_04]: uh Apple intelligence or you're gonna ask Gemini or bread who you know whoever you're gonna go

[00:13:10] [SPEAKER_04]: to an alternative because we already know that the same prompt in different general AI tools

[00:13:17] [SPEAKER_04]: that actually going to give you at least a slightly different answer by the way it's you know trying

[00:13:23] [SPEAKER_02]: to understand the question or yeah I would say vastly different right there's gonna be huge

[00:13:28] [SPEAKER_02]: differences in how something like co-pilot would answer versus you know using cloud and again

[00:13:33] [SPEAKER_02]: that's just for asking questions but when you go deeper into the flow of work and you're actually

[00:13:37] [SPEAKER_02]: looking at like what am I mostly doing at work they're gonna be tools that are gonna outshine

[00:13:42] [SPEAKER_02]: something like co-pilot pretty quickly so I don't think that you can really avoid people bringing in

[00:13:47] [SPEAKER_02]: their own AI tools but it does mean that there has to be some vetting and like you said like who's

[00:13:52] [SPEAKER_02]: carrying that agenda because obviously IT has a big say in that at the other site like 90% of

[00:13:58] [SPEAKER_02]: this isn't really about technology it's really about people because I think Ethan Mallick said

[00:14:02] [SPEAKER_02]: it the best that this is not software right this is not software in the sense that you know

[00:14:07] [SPEAKER_02]: software was always submitted you could install or using the clouds and it had a button and when

[00:14:11] [SPEAKER_02]: you push a certain button and a certain action would take place and software is very regular it's

[00:14:16] [SPEAKER_02]: very strict it's very methodical you can implement it you can train people on it and then they use

[00:14:20] [SPEAKER_02]: the software this is like a coworker this is bringing new people into the workforce that just

[00:14:25] [SPEAKER_02]: happens to be AI's right and that's exactly how we use AI in our team I've said many times

[00:14:31] [SPEAKER_02]: like whether I go if I if I remember something I have to do whether I go on Slack

[00:14:36] [SPEAKER_02]: and at mention one of my colleagues or I go to chat to PT and I at mention one of my GPT's

[00:14:41] [SPEAKER_02]: it's almost the same to me only difference is the GPT's worked 24-7 they never ask for a raise

[00:14:47] [SPEAKER_02]: they never go to HR because I say something wrong right so I think like this is really a human

[00:14:53] [SPEAKER_02]: challenge and if we really look at AI as another colleague and as part of our new workforce

[00:14:59] [SPEAKER_02]: it has to be an in-first of all it has to be a company issue CEO issue but then very

[00:15:04] [SPEAKER_02]: very quickly in line has to be an HR issue because you're really talking about

[00:15:08] [SPEAKER_02]: a new generation of employees coming in and how do our current employees work well with them

[00:15:13] [SPEAKER_02]: and I would say that like whenever I'm working very hands-on like in the lead with AI course when

[00:15:18] [SPEAKER_02]: you're working one-on-one with people or with teams they really see people suddenly like it's

[00:15:24] [SPEAKER_02]: just like flipping a switch they see it completely differently because they've been hearing about it

[00:15:29] [SPEAKER_02]: as a piece of software not as a new coworker but the moment that they see it as a coworker

[00:15:33] [SPEAKER_02]: they also approach it completely differently and one person who was in the course actually said

[00:15:38] [SPEAKER_02]: and pretty senior person obviously because it's for executives he said I didn't realize how

[00:15:44] [SPEAKER_02]: bad I was a delegating until I started working really with AI but that's it right it's another

[00:15:50] [SPEAKER_02]: person that you delegate work to and the better your instructions are the better the outputs are

[00:15:55] [SPEAKER_02]: and so therefore to me it's totally an HR issue and not so much an IT issue maybe IT is involved in

[00:16:01] [SPEAKER_02]: which platform just to like but again for most companies that is already pretty obvious because if

[00:16:05] [SPEAKER_02]: you're already on Microsoft you're probably going to choose co-pilot doesn't really make sense

[00:16:09] [SPEAKER_02]: to separate all the data that AI needs to be really good and the AI right if you're already

[00:16:15] [SPEAKER_02]: on Google you're going to go on Gemini right when that rolls out fully in enterprise right

[00:16:19] [SPEAKER_02]: so it doesn't really make sense for companies so I would say IT definitely has to be involved

[00:16:23] [SPEAKER_02]: in that otherwise to me is purely an HR issue yeah no absolutely great on that this is really about

[00:16:29] [SPEAKER_04]: people in behavior change and when I was in NBC Universal there was a consultancy that

[00:16:35] [SPEAKER_04]: laid out this automation strategy and automation like mature in scale where you had

[00:16:41] [SPEAKER_04]: you had RPA, a robotic process automation and you had intelligent automation where it was

[00:16:45] [SPEAKER_04]: a little bit more than just you know if then kind of statements and had a little bit of

[00:16:50] [SPEAKER_04]: smarts to it and then and then there was cognitive automation which really gets into where we are now

[00:16:56] [SPEAKER_04]: with actual AI and having it as part of your decision support system which is incredibly

[00:17:05] [SPEAKER_04]: important in the talent in HR spaces as you know well and so you can't just look at historical

[00:17:11] [SPEAKER_04]: patterns of decisions and say well you know I see I know what you're going to do so I'll just take

[00:17:16] [SPEAKER_04]: care of it like whoa whoa whoa whoa now that's not exactly how work not to mention the fact

[00:17:22] [SPEAKER_04]: that most organizations don't have the data quality and data maturity, data analytics maturity

[00:17:28] [SPEAKER_04]: to but even do that even if you you wanted to which it's in a whole other topic around I

[00:17:35] [SPEAKER_04]: talked to another guest about like just because you can automate something or use AI for something

[00:17:40] [SPEAKER_04]: doesn't mean you necessarily good and I think these are responsible and ethical kinds of questions.

[00:17:47] [SPEAKER_02]: Absolutely but I also think that like the best case that is out there right now right if you

[00:17:51] [SPEAKER_02]: look at the Moderna Open AI case study if you look at the Microsoft HR VP of HR like they

[00:17:58] [SPEAKER_02]: they wrote this really great case study and in all of those examples it's really not so much

[00:18:02] [SPEAKER_02]: the company deciding what to do I mean the company may have chosen which platform which in Microsoft

[00:18:06] [SPEAKER_02]: case is pretty easy to choose they may have chosen a platform but in all those instances they're

[00:18:11] [SPEAKER_02]: really leading it to the employees to say here's where I'm going to apply it because just for

[00:18:15] [SPEAKER_02]: since at a few months ago I think in a different context but your frontline workers they know

[00:18:20] [SPEAKER_02]: what work they do much better than you so it doesn't really make sense to try and solve this at a

[00:18:26] [SPEAKER_02]: leadership level or at an organizational level again as long as you introduce this as a new

[00:18:31] [SPEAKER_02]: co-worker that can alleviate a lot of the work that you probably don't really like right what BCG

[00:18:35] [SPEAKER_02]: calls the toyol part of your job and it can free you up to do the more joyful part of your job as

[00:18:41] [SPEAKER_02]: long as it's introducing that way it makes much more sense to leave it to employees and say

[00:18:46] [SPEAKER_02]: where is the use case going to be for you and in both those case studies it was very much

[00:18:51] [SPEAKER_02]: you know that use cases came up that they would never have thought about because again like at

[00:18:55] [SPEAKER_02]: a central level you don't really know the work as deeply as the people on the grounds and so I think

[00:19:00] [SPEAKER_02]: then you do need training I think I totally agree with that part but on you do need training on the

[00:19:07] [SPEAKER_02]: you know how do you use it ethically and how do we you know what data do you input and what they

[00:19:11] [SPEAKER_02]: do not input because look open a i may say that if you switch this toggle the model is in train

[00:19:17] [SPEAKER_02]: on it but the data flow somewhere and you know you never know what's going to happen right so

[00:19:21] [SPEAKER_02]: there still needs to be some training and I think Jared Speterrow the AI chief at Microsoft

[00:19:26] [SPEAKER_02]: said in a webinar a while ago that you know one of the key things you need to teach people is

[00:19:32] [SPEAKER_02]: better judgment right what outputs of AI do you accept and what do you add it and what do you throw away

[00:19:39] [SPEAKER_02]: right that kind of judging but again in very much the the same fashion as we were talking about how

[00:19:44] [SPEAKER_02]: you work with it in terms of instructing receiving the output back it's just like working with a

[00:19:49] [SPEAKER_02]: colleague if you breathe a junior colleague to do a project for you you know going to be hands

[00:19:54] [SPEAKER_02]: often just forward to your boss right you're probably going to check that work and you're going

[00:19:58] [SPEAKER_02]: to spot the things that are not good about it and you're going to re-brieve that part until you

[00:20:02] [SPEAKER_02]: have to write the right work so that training on judgment and the training on instruction and

[00:20:07] [SPEAKER_02]: what is ethical and training what falls within the guidelines of the company is all going to be a huge

[00:20:12] [SPEAKER_02]: part of how companies work with AI and letting the employees work with AI yeah I mean I feel like

[00:20:18] [SPEAKER_04]: I never like hearing stories about like employee monitoring because I think it just shows a

[00:20:26] [SPEAKER_04]: complete lack of trust and you're treating your employees like students or children and the

[00:20:34] [SPEAKER_04]: hour they are adults let's not forget yeah I mean I saw this when when social media the

[00:20:40] [SPEAKER_04]: admin social media I was at IBM and they were quick to you know develop some social what they call

[00:20:45] [SPEAKER_04]: the social computing you know guidelines like we expect you to behave like an adult and that means

[00:20:50] [SPEAKER_04]: not bending all day on Facebook and Twitter and whatever and by and by a way if you have worked

[00:20:55] [SPEAKER_02]: to do because I think this also goes into a lot of the debate around like remote work and I would work

[00:21:00] [SPEAKER_02]: if you have work to do and you're clear on your objectives and you actually like what you're

[00:21:04] [SPEAKER_02]: working towards like your goals in the company's goals are aligned you're happy to be a work

[00:21:08] [SPEAKER_02]: therefore you know you cannot be on Facebook and thinking a similar way if you know again there

[00:21:12] [SPEAKER_02]: is a bit of education required maybe but the moment that you understand broadly how these models

[00:21:17] [SPEAKER_02]: work and what happens with the data and what are good use days use good use cases and bad use cases

[00:21:23] [SPEAKER_02]: your interest in the companies interest are very aligned I think there's somehow this

[00:21:28] [SPEAKER_02]: perception and I spend a lot of time on the anti work rate it because you get the real

[00:21:32] [SPEAKER_02]: sentiment of people and on certain tick-tockers comments and you really see like when employees see

[00:21:38] [SPEAKER_02]: you know some video about some situation to workplace and they all go on like that's exactly

[00:21:42] [SPEAKER_02]: like with me there's such a distrust between employees and companies that really doesn't need to be

[00:21:48] [SPEAKER_02]: there because the incentive should be pretty aligned the company wants to achieve something

[00:21:52] [SPEAKER_02]: if the company achieves that it makes money and therefore it can pay its employees

[00:21:56] [SPEAKER_02]: and if it's also something that you like working on and therefore doing good is doing well etc

[00:22:02] [SPEAKER_02]: the incentive should be pretty aligned and therefore these things should fall pretty naturally

[00:22:06] [SPEAKER_02]: but I also see obviously in that's where our whole mission comes from I also see plenty of examples

[00:22:11] [SPEAKER_02]: where that's not the case and again in the case of remote work where companies are so distressing

[00:22:16] [SPEAKER_02]: of people and telling them to come into the office for quote unquote camaraderie and community

[00:22:21] [SPEAKER_02]: where people just travel for an hour or a sit behind their laptop and do the same zoom cause they would have

[00:22:26] [SPEAKER_02]: and actually are are handing in a lot of the productive time they've been outspend

[00:22:31] [SPEAKER_02]: committing to the office so I recently interviewed Amy Lashikahal from ADP and she said time to

[00:22:37] [SPEAKER_02]: grown up if I work like we cannot treat work as something that we have all the smarts and like you said

[00:22:44] [SPEAKER_02]: about we deal with a bunch of kindergartners and we need to control them and we need to we know

[00:22:49] [SPEAKER_02]: what's better for them we know when they should show up for work and we know what tools to use

[00:22:53] [SPEAKER_02]: I just don't think that works in this day and age like maybe even five years ago that's still

[00:22:57] [SPEAKER_02]: would have worked I just think like the genius out of the bottle we're working in very different ways

[00:23:03] [SPEAKER_02]: and again that's what you see in AI where the employees are leapfrogging their companies because

[00:23:07] [SPEAKER_02]: they're they're getting smarter about these tools than most of the companies again

[00:23:12] [SPEAKER_02]: I train executives and nothing against the people who are taking the course but I just see it all the time

[00:23:17] [SPEAKER_02]: that they are now and at least they're awake to it at least the people that I'm coaching

[00:23:22] [SPEAKER_02]: they're awake to it that their employees are running faster than they are they're walking across

[00:23:27] [SPEAKER_02]: the floor or they're seeing things on slack where they're like I don't even know how that works

[00:23:30] [SPEAKER_02]: like I don't know how you got that I don't know how you did that right so they're doing this sort of

[00:23:34] [SPEAKER_02]: like we do it two week almost like bootcamp where it's really you know a bit of hard work but

[00:23:39] [SPEAKER_02]: you get to the point where you actually you understand AI and you're using a couple of tools

[00:23:43] [SPEAKER_02]: and you're starting to develop your use cases but you know the people are working like they're

[00:23:47] [SPEAKER_02]: going faster so I think if there was ever an age for autonomy and agency with employees it's

[00:23:53] [SPEAKER_02]: now you know and yeah let them make wise decisions but again why would they not like what are

[00:23:58] [SPEAKER_04]: you doing in your company that the people are against you? Well I think it's a bit ironic that

[00:24:05] [SPEAKER_04]: we are still like laser focused on productivity where generally they I use that's actually

[00:24:14] [SPEAKER_04]: like a that's like the new vanity metric in my opinion and yet we still have the return to office

[00:24:20] [SPEAKER_04]: and mandate you're literally hindering my ability to produce a metric that you like from

[00:24:27] [SPEAKER_04]: from a productivity standpoint in a way you're still measuring productivity largely at an individual

[00:24:33] [SPEAKER_04]: level and we don't work in isolation we work as part of teams that's isn't that why you

[00:24:40] [SPEAKER_04]: you claim you want to back in the office in the first place because you want your point

[00:24:45] [SPEAKER_04]: the camaraderie and all that so I just feel like a lot of organizations that there's a lot of

[00:24:51] [SPEAKER_04]: lip service there's a lot of like we don't know is that we're just gonna we're just gonna batten

[00:24:55] [SPEAKER_04]: down the hatches and I just think that's a huge mistake I think shadow IT is worse than ever

[00:25:01] [SPEAKER_04]: I think that to your point the the consumerization of IT right where you know what's happening

[00:25:07] [SPEAKER_04]: in the consumer space is happening at a faster clip than what's happening inside of organizations

[00:25:12] [SPEAKER_04]: you know that that disparity is pretty bad I think to your point and I get it I get it

[00:25:19] [SPEAKER_02]: it's terrifying I mean I get it like I'm not running a big company I'm not in an IT function

[00:25:24] [SPEAKER_02]: in a big company I will be terrified too if you're seeing people you know from whatever monitoring

[00:25:29] [SPEAKER_02]: you're doing or just anecdotally you're seeing people bring in tools like fireflies other

[00:25:35] [SPEAKER_02]: giving them access to every meeting that you attend and you're using three four models and

[00:25:40] [SPEAKER_02]: oh you just heard that Gemini is really good for this and now you're gonna go on your lockdowns

[00:25:45] [SPEAKER_02]: I will be terrified too but hey that's the reality and again like the company is not going to

[00:25:50] [SPEAKER_02]: go fast enough to deliver a credible alternative and like the same I totally get why companies would

[00:25:56] [SPEAKER_02]: choose for a Gemini install or a copilot install because it links so well with the current infrastructure

[00:26:04] [SPEAKER_02]: and it does you know again I need data so that the best source of data is all your messaging

[00:26:10] [SPEAKER_02]: and all your meetings and all your files I totally get it but I think it's hard to avoid that someone

[00:26:15] [SPEAKER_02]: brings in their own AI tool and says look this is what I need for X right and how agile you are

[00:26:21] [SPEAKER_02]: and how responsive you are to people making those requests and bringing those things in I think that's

[00:26:27] [SPEAKER_02]: going to determine in a big part how competitive you're going to be in the marketplace right because

[00:26:32] [SPEAKER_02]: a company where like that's why I thought that modern I case that even of course it's marketing

[00:26:36] [SPEAKER_02]: but why I thought it was so impressive that if a highly one of the most highly regulated industries

[00:26:42] [SPEAKER_02]: one of the most highly regulated companies in the world like moderna can give people kind of like

[00:26:47] [SPEAKER_02]: card blanche and just develop whatever you want create your own GPT share them within the company

[00:26:53] [SPEAKER_02]: of course with again with some training and with some guidance and again they had Slack channels

[00:26:57] [SPEAKER_02]: where open AI people were in that channel you could ask question they had office hours they celebrated

[00:27:02] [SPEAKER_02]: best practices the CEO would tell everyone what GPT he developed and what work etc etc I mean if

[00:27:09] [SPEAKER_02]: they can do it it's it's clearly possible but I don't like the reality is different I get people are

[00:27:15] [SPEAKER_04]: going to be building custom things and then pushing them to like an internal library like an internal

[00:27:22] [SPEAKER_04]: repo or something like that and yeah you got to have some standards around that and then

[00:27:28] [SPEAKER_04]: you've got to have guard realism in just a certain way back to what we're talking about before

[00:27:32] [SPEAKER_04]: about like this continuous monitoring I mean it could just be around you know usage and making

[00:27:37] [SPEAKER_04]: sure that people aren't uploading certain things I mean just like you would have hopefully you've

[00:27:43] [SPEAKER_04]: good you know cyber you know policies and monitoring that make sure people aren't uploading

[00:27:49] [SPEAKER_04]: you know proprietary narrow data where it shouldn't go so even just extending the

[00:27:55] [SPEAKER_04]: some of the things that we've done already in the data privacy and and cybersecurity space

[00:28:01] [SPEAKER_04]: it seems like like anybody could be the weakland right that's why I always say

[00:28:06] [SPEAKER_04]: when it comes to response to AI we're all responsible on a come-to-level and so

[00:28:11] [SPEAKER_04]: you know without putting too much you know constraints on it just it just seems like

[00:28:16] [SPEAKER_04]: you should at least standardize on where there's a use kit like you mentioned

[00:28:19] [SPEAKER_04]: you know order and read.ai which the letter which I use it is significantly better

[00:28:28] [SPEAKER_04]: than Google's transcription whenever recorded Google meets like I don't know what

[00:28:33] [SPEAKER_04]: what happened because Google has been like the you know poster child for you know language and

[00:28:40] [SPEAKER_04]: translation and whatever they invented transform or so yes. Yeah and yet they can't do a transcript.

[00:28:48] [SPEAKER_02]: But like that's that's the yeah again like the incentives are completely different right like someone

[00:28:53] [SPEAKER_02]: you know a founder somewhere because I mostly were a start-ups like a founder somewhere can say

[00:28:58] [SPEAKER_02]: yeah that's the status quo I mean that's how I interviewed with the guy from intercom yesterday and he said

[00:29:03] [SPEAKER_02]: you know I saw Zendesk declining and I saw that was my opportunity right it's like someone just can see

[00:29:08] [SPEAKER_02]: something happening in the space put one-on-one together and then suddenly you know you have a new idea

[00:29:13] [SPEAKER_02]: and you can probably execute that much better and much more targeted than the big platform scan so

[00:29:19] [SPEAKER_02]: the big platforms was still offer some of those solutions just like you know we work a lot of

[00:29:24] [SPEAKER_02]: like the hybrid workplace booking software right you know Microsoft now is also integrating that

[00:29:29] [SPEAKER_02]: and Google's integrating that and if you put a calendar invite you can already select like where

[00:29:34] [SPEAKER_02]: meeting will take place and it will be some stuff around that's booking but it will never go as

[00:29:38] [SPEAKER_02]: deep in the UI will never be as targeted and polished as a singular provider focusing on that's

[00:29:44] [SPEAKER_02]: on that issue and if you just look at like all the options out there right now you know those

[00:29:49] [SPEAKER_02]: guys can run so much faster they can develop so much faster I think that's always been the case obviously

[00:29:54] [SPEAKER_02]: in software but I think the model is going to be very different now where you know you could build

[00:29:59] [SPEAKER_02]: platforms where again even the UI could be changed you saw a little bit of that in the in the

[00:30:04] [SPEAKER_02]: Gemini demo by the way Gemini excels in demos that never seem to come true but you saw a little bit

[00:30:09] [SPEAKER_02]: of that in the Gemini demo where UI was being produced in front of your eyes to display certain

[00:30:15] [SPEAKER_02]: content in a certain way right so you know you could see that you build platforms that could

[00:30:20] [SPEAKER_02]: tailor UI even more specifically than what that startup can do but I think there's no there's

[00:30:26] [SPEAKER_02]: two ways about it that there will always be the bigger platforms and then they will be startup

[00:30:31] [SPEAKER_02]: to develop something is just that now everyone can be their own CTO like everyone can create their own

[00:30:38] [SPEAKER_02]: AI startup because you know have you know a devan you know you now have these co pilot coding agents

[00:30:47] [SPEAKER_02]: so the accessibility of creating something better while it's being much higher and therefore you

[00:30:53] [SPEAKER_02]: can probably see a lot more competition which it only be really interesting for us not CIOs

[00:30:59] [SPEAKER_02]: you're not really good for us to see what's out there exactly where are your thoughts on like

[00:31:05] [SPEAKER_04]: the interoperability of some of these you know agents co pilot so whatever I mean let's say

[00:31:10] [SPEAKER_04]: you've got an AI power talent solution and that needs to talk back to the HRIS or

[00:31:19] [SPEAKER_04]: performance management systems or whatever that's you know in a whole another you know AI

[00:31:25] [SPEAKER_04]: solar system are these all going to be sort of interoperable so that they speak to each other

[00:31:31] [SPEAKER_04]: so that you don't have to rely on rely less on like the big LLM you know vendors I mean what's

[00:31:37] [SPEAKER_04]: the evolution of all these agents you know running around I've heard terms like it agentic

[00:31:45] [SPEAKER_04]: computing yeah I'm very ex of agents and interoperable because I think it's one thing to just

[00:31:50] [SPEAKER_04]: create something that can do a particular task almost like an appliance and it's another to

[00:31:55] [SPEAKER_04]: interconnect you know in a sort of IoT kind of contact you have a smart home or you just have a smart

[00:32:02] [SPEAKER_04]: fridge that doesn't talk to any other person or machine yeah and maybe maybe we're fine with

[00:32:09] [SPEAKER_02]: our fridge is not talking behind our back to other appliances you know it's like well he ate that

[00:32:13] [SPEAKER_02]: morning like come on but so not gonna think bye bye I mean I think that's where we're you see Apple

[00:32:18] [SPEAKER_02]: Intelligence kind of getting into right where just to say look we know certain intends within applications

[00:32:25] [SPEAKER_02]: we're gonna unlock that for developers so you can sort of name certain intends within your

[00:32:30] [SPEAKER_02]: application now to become something that you can instruct Siri to do so you're getting a lot deeper into

[00:32:36] [SPEAKER_02]: one kind of central point and then access into everything right and again like I just I just move back

[00:32:42] [SPEAKER_02]: over ten years in Asia and you look at the super apps there like a wee chat right

[00:32:47] [SPEAKER_02]: everyone wants to be the home page everyone wants to be the starting points right so obviously Apple

[00:32:51] [SPEAKER_02]: is making a really strong play for it right now but it's the same thing of jetty beauty they would

[00:32:56] [SPEAKER_02]: like to be the starting point for everything right that's why you can control complexity and

[00:33:00] [SPEAKER_02]: consensus and all trails all from your jetty beauty main interface that's only going to be more so

[00:33:06] [SPEAKER_02]: when it's all done through voice right and then obviously co-pied once to be your main interface

[00:33:11] [SPEAKER_02]: to the world and then Gemini of course wants to do that too so everyone wants to be that central

[00:33:15] [SPEAKER_02]: point and I think we even see this in something like an eightfold saying we're not gonna wait for

[00:33:21] [SPEAKER_02]: service now or someone else to build an all encompassing AI I think SANA I did it too we're going

[00:33:27] [SPEAKER_02]: to give you the universal assistant that yeah it works with us but it also works with a bunch

[00:33:31] [SPEAKER_02]: of other applications because again we don't want you to have to wait for another company to do that

[00:33:36] [SPEAKER_02]: obviously you know they also have some benefits benefit from that I think everyone wants to be

[00:33:41] [SPEAKER_02]: that first points because you know you know if your Google you control all of search that's

[00:33:46] [SPEAKER_02]: pretty lucrative right like if you're the main interface the main choke points through which

[00:33:50] [SPEAKER_02]: all of the demands go you can decide which application to send that to so I think that's going

[00:33:56] [SPEAKER_02]: to be something that has to be solved and then everything on the back end right like who is going

[00:34:00] [SPEAKER_02]: to unlock what for whom that's probably going to be a big puzzle and you know that's why

[00:34:06] [SPEAKER_02]: I love following Josh Perseent because he talks a lot of people at you know service now

[00:34:11] [SPEAKER_02]: and work day and all of these other big platforms and it seems like everyone at the same time

[00:34:15] [SPEAKER_02]: is working on something very similar which is the universal assistant who we're gonna end up

[00:34:20] [SPEAKER_02]: using I mean that's a good question right are we all gonna spend all of our time in work day AI

[00:34:24] [SPEAKER_02]: and just do everything from there I don't know maybe what do you think the average person

[00:34:30] [SPEAKER_04]: might just be like wait which which one do it what's my starting point to do this and can I just

[00:34:36] [SPEAKER_04]: have on my homepage to your point one interface call it whatever you want call it you know Jarvis

[00:34:44] [SPEAKER_04]: whatever and then it just knows what other agents right I need time off oh I need to handle that

[00:34:50] [SPEAKER_02]: with that software I need to put an expense claim and I do it with that software yeah that makes sense

[00:34:56] [SPEAKER_04]: that sounds good let's do it yeah yeah we even in the consumer space it was like it became an

[00:35:04] [SPEAKER_04]: unsolved you know dilemma because we weren't sure and there was no one at the time that was

[00:35:10] [SPEAKER_04]: building something that was you know to your point like this sort of universal you know translator or

[00:35:16] [SPEAKER_04]: you know universal you know virtual assistant or whatever but I mean even in the consumer space that

[00:35:21] [SPEAKER_04]: hasn't been solved right I mean apples trying to wedge themselves in and you know progress along

[00:35:27] [SPEAKER_04]: the wall garden kind of you know mentality which is totally understandable but you know if I just

[00:35:33] [SPEAKER_04]: have a general question I have an Alexa echoed off right there I've got my my iPhone and my apple watch

[00:35:44] [SPEAKER_04]: I've got Google home I've got and then I have whatever is on my Lenovo laptop right so I have five

[00:35:54] [SPEAKER_04]: right and I don't know who's gonna give me the best answer or the answer that I like or need or whatever

[00:35:59] [SPEAKER_04]: so if you think forward to true adoption not just implementation but actual user adoption going back

[00:36:07] [SPEAKER_04]: to the people's centric nature of all of this that's a huge friction point people are like

[00:36:13] [SPEAKER_02]: not sure which direction to go and for some companies that may mean right Bob that and this is

[00:36:20] [SPEAKER_02]: the discussion I had with with Matt from Bcg yesterday at Bcg x right they have 3000 AI engineers

[00:36:27] [SPEAKER_02]: right Bcg is 30,000 people they have 3000 people purely doing AI like building engineering right

[00:36:35] [SPEAKER_02]: because for a lot of companies the answer may be build your own so almost like doesn't matter what

[00:36:42] [SPEAKER_02]: model is behind that again there may be some implications in terms of if you're already doing most

[00:36:47] [SPEAKER_02]: most of your computing on Azure Cloud you know maybe you're going to tap into cloud or like whatever

[00:36:52] [SPEAKER_02]: you know if you're on AWS you may use you know model X so they may have some linkage there but for

[00:36:58] [SPEAKER_02]: a lot of companies that maybe and I've spoken to people in the last cohort of the course whose

[00:37:02] [SPEAKER_02]: companies have developed their own and there you can you can be a little bit more directive because

[00:37:07] [SPEAKER_02]: that's basically you can say well at least we all know to your points out of the 56 touch point this

[00:37:12] [SPEAKER_02]: is the only sanction channel to go through when it comes to work related you know using using an AI

[00:37:18] [SPEAKER_02]: for meaning transcription using an AI for answering questions using an AI for analyzing data

[00:37:24] [SPEAKER_02]: is very simple we have our own here's how you use it right and no McKinsey had developed something

[00:37:30] [SPEAKER_02]: already last year this company I spoke to as part of the cohort they have done it and then at least

[00:37:35] [SPEAKER_02]: first point is very clear I think then the problem that you still may face is that people again they

[00:37:41] [SPEAKER_02]: use that and they can just feel that it's not on par with what they're using in their private life

[00:37:48] [SPEAKER_02]: and it's like wait yesterday I generated with a little bit of back and forth like the perfect

[00:37:53] [SPEAKER_02]: recipe for something I want to cook based on a photo that I took of my fridge and what I had there

[00:37:56] [SPEAKER_02]: which is now a normal at soundstreet you're listening is now a normal use case and now I'm coming

[00:38:01] [SPEAKER_02]: to work and the thing doesn't even do XYZ right so you may still run into that issue but that's

[00:38:06] [SPEAKER_02]: like an whole other layer that I think all these kind of integrators and companies are jumping on

[00:38:11] [SPEAKER_02]: which is letting companies develop their own first points which then corresponds with both the

[00:38:16] [SPEAKER_02]: proprietary company data like I think it's called Lily right at McKinsey where it's digging from

[00:38:20] [SPEAKER_02]: 60 years up McKinsey data that is your unique advantage right that's your district differentiator

[00:38:25] [SPEAKER_02]: and then on top of that it can tap into one of the other models right and then the company can

[00:38:30] [SPEAKER_02]: make a cost effective by saying for certain things we don't need you know like TBT40 we can just use

[00:38:35] [SPEAKER_02]: like a really cheap model that does you know really you know it takes a bit longer but you know

[00:38:39] [SPEAKER_02]: like it can do this and that so that's like an whole other layer where a bigger company

[00:38:43] [SPEAKER_02]: says to think about it's not even choice of which tool do I buy or which tool I sanction

[00:38:49] [SPEAKER_02]: we're actually going to have to build our own because again that's going to be our differentiator

[00:38:55] [SPEAKER_02]: and you saw this at JP Morgan right J. and more like 2,000 people now dedicated on AI

[00:38:59] [SPEAKER_02]: and they're building their own AI interface again their own AI model because again they have

[00:39:05] [SPEAKER_02]: such unique data sources and they're seeing that as their competitive advantage so they were really

[00:39:10] [SPEAKER_02]: early in the AI news because they were forbidding people to use JTBT but then they got up pretty

[00:39:14] [SPEAKER_02]: nicely by building their own model and now giving people like a totally unique and competitive advantage

[00:39:20] [SPEAKER_02]: in being able to dig into again decades of data that only day actually have access to and then

[00:39:29] [SPEAKER_04]: just just to summarize that there what you're highlighting is that once you can build your own

[00:39:36] [SPEAKER_04]: stuff then everyone is both a builder and consumer of this technology which just you know sort of

[00:39:44] [SPEAKER_04]: reemphasizes the need for responsible AI by design through you know usage and and all your decision

[00:39:51] [SPEAKER_04]: making processes so the talent space that is full talent life cycle kind of awareness and

[00:39:58] [SPEAKER_04]: responsibility and then to your point earlier you know you've got to think about this when you do

[00:40:04] [SPEAKER_04]: upscale people around AI it's not just here's the output of the tool let's think critically

[00:40:11] [SPEAKER_04]: about where it may have gone you know or I or you know just don't take it blindly like a

[00:40:17] [SPEAKER_04]: calculators you know output right but now people are going to take this initiative and go and start

[00:40:23] [SPEAKER_04]: thinkering themselves as they know you know through the community we've seen a lot of people

[00:40:28] [SPEAKER_04]: yep you know do that already and I know you're yep you've got a course coming up to teach people

[00:40:32] [SPEAKER_04]: how to you know build their first AI employee right so so everyone is going to be

[00:40:39] [SPEAKER_04]: getting into this game and anyone frankly anyone with 20 bucks can go and build a custom

[00:40:44] [SPEAKER_04]: Dbt and and TITPT plus so this is something we all have to you know really think

[00:40:51] [SPEAKER_04]: you know deeply about and going back to the productivity saying it's like if you're incentivizing

[00:40:57] [SPEAKER_04]: people based on metrics that are fairly on you know throughput and a very good see and productivity

[00:41:04] [SPEAKER_04]: you've got to really rethink that like at what cost yeah I mean that's not treating people as people

[00:41:10] [SPEAKER_02]: right which is ironic in the age of AI and so that's what that's why I totally agree with like

[00:41:16] [SPEAKER_02]: Debbie Lovic and the BCG teams kind of perspective on you know AI is a way to reduce people's

[00:41:23] [SPEAKER_02]: toil at work it's a way to increase people's joy at work that drives more engaged employees

[00:41:29] [SPEAKER_02]: and more engaged employees as we all know from the get-up data you know they take less sick days

[00:41:35] [SPEAKER_02]: stay with the company longer their better colleagues their better leaders you know so from from

[00:41:40] [SPEAKER_02]: every perspective like again use AI look at how people are doing the work today and then let them use

[00:41:45] [SPEAKER_02]: AI to take out that part of the job that they don't they don't find any joy in anyway right and I think

[00:41:50] [SPEAKER_02]: you look at that and to do it from like with with the ethics in mind again you look at that

[00:41:55] [SPEAKER_02]: Microsoft HR case study they said we did very light education upfront we gave people the tools and we

[00:42:00] [SPEAKER_02]: had go and code away go and create you know little programs and by the way what is the first thing

[00:42:07] [SPEAKER_02]: that the HR team built an FAQ bots of course but but they said look here are five or values

[00:42:14] [SPEAKER_02]: five or principles for how we build AI at Microsoft and everyone who builds because you're now

[00:42:20] [SPEAKER_02]: a builder to your points everyone who builds has to adhere to those policies or to those

[00:42:25] [SPEAKER_02]: to those poor values and so again you're making it a very shared discussion you're making it

[00:42:29] [SPEAKER_02]: very much like a co-creation and you're telling people look we uphold these rules and now

[00:42:35] [SPEAKER_02]: they're going to build you're going to uphold these rules as well and those principles are

[00:42:39] [SPEAKER_02]: very straightforward and you know you're using it for good and the data that I write so I think

[00:42:44] [SPEAKER_02]: that's all very possible and in that case study Chris Fernandez mentions you know you don't have

[00:42:51] [SPEAKER_02]: to be technologists to roll out chat to PT or or or co pilot you don't have to be technologists

[00:42:55] [SPEAKER_02]: or all of AI you just need an adoption plan you just need some kind of idea of how it's going to

[00:43:00] [SPEAKER_02]: play a role in your company and as I'm now working not only with individual executives but also

[00:43:05] [SPEAKER_02]: with entire like leadership teams sometimes that's the part that's the hardest part is to kind of

[00:43:10] [SPEAKER_02]: you know they all have employees using AI they feel a little bit left behind sometimes the older

[00:43:16] [SPEAKER_02]: people are the more that's the case and then they need to be aligned on okay now I kind of get what

[00:43:23] [SPEAKER_02]: they are AI I do and again like mentioning things like AI is not software it's your new colleague

[00:43:28] [SPEAKER_02]: and letting them build their first AI employee and understanding that and again this is like

[00:43:33] [SPEAKER_02]: bursting bubbles a lot of the time understanding that a lot of things that we think is uniquely

[00:43:38] [SPEAKER_02]: human work is extremely easy to automate with AI because a lot of what we do is if this then

[00:43:45] [SPEAKER_02]: that and if AI cannot do it outright it can at least give you the first draft and so an example

[00:43:51] [SPEAKER_02]: is something very simple like answering emails right we think that that's like no one could ever take

[00:43:55] [SPEAKER_02]: that over and everyone has these like nightmares of AI sending all kinds of rubbish as reply to

[00:44:01] [SPEAKER_02]: emails and then I let people break down how do you answer emails right I said well I open my

[00:44:06] [SPEAKER_02]: inbox I click on the title I read the email I understand what that person is saying based on everything

[00:44:12] [SPEAKER_02]: I know I think what could I reply I choose an option I reply it's like oh interesting that's

[00:44:17] [SPEAKER_02]: exactly how grandma AI works and that's exactly how now Apple intelligence will work because I saw

[00:44:22] [SPEAKER_02]: that they copied that from from grandma and then go by it will work in the same way I mean there's

[00:44:26] [SPEAKER_02]: so much work that we do as humans that is so easy to automate and so just doing that and just

[00:44:34] [SPEAKER_02]: allowing people to do that and getting people to take those like 20 30% productivity gains and then

[00:44:41] [SPEAKER_02]: not applying that to just doing more work right my first lesson is you got to meditate on

[00:44:47] [SPEAKER_02]: Parkinson's law the amount of work will always expends it the amount of time that you allow for

[00:44:51] [SPEAKER_02]: it right if we're going to have a 30% productivity gain but then we're going to fill up with other

[00:44:56] [SPEAKER_02]: nonsense work that we don't really like doing and what's the point right so that's like really

[00:45:00] [SPEAKER_02]: my culture action as you got to use this for something that makes people feel better about

[00:45:06] [SPEAKER_02]: the way that they work and the work that they do and so use that extra time maybe we split

[00:45:11] [SPEAKER_02]: the difference right like maybe you can work a little bit last time I put everyone on a four

[00:45:14] [SPEAKER_02]: day work work week schedule because I don't need them for five days anymore and the other part

[00:45:19] [SPEAKER_02]: use that for more strategic work use it for more collaboration use it for spearheading new

[00:45:24] [SPEAKER_02]: initiatives thinking about the millions of things you could do and that were impossible maybe

[00:45:29] [SPEAKER_02]: two years ago because we didn't have AI right so there's so much possible there and I don't

[00:45:33] [SPEAKER_02]: think it takes the technology as I think it just takes someone Amy Leskeh Kalkal is that like a

[00:45:38] [SPEAKER_02]: courageous leader who steps up and says the reality is here here's the opportunity here are the

[00:45:44] [SPEAKER_02]: guidelines here maybe those core values or those core ethical values and now go and experiment

[00:45:50] [SPEAKER_02]: and let's all share in real time because we've never had something in a company where from the CEO

[00:45:55] [SPEAKER_02]: to the front desk work or everyone was learning at the same time it's never happened right there

[00:45:59] [SPEAKER_02]: is no time for the editing team to learn something to then go and teach it because we're all

[00:46:03] [SPEAKER_02]: here in the press release yesterday right and that's all learned at the same time let's all share

[00:46:07] [SPEAKER_02]: what we what knowledge we gain and what are the best use cases and then collectively you know

[00:46:12] [SPEAKER_02]: use this tool to our advantage I think that's the only way to do it if you can establish

[00:46:17] [SPEAKER_04]: you know a sort of let's learn from from our you know failures let's you know do this

[00:46:22] [SPEAKER_04]: gather let's experiment in a safe environment you know you put up a sand box or whatever I mean

[00:46:28] [SPEAKER_04]: IBM was created that very different they had program called technology adoption program

[00:46:32] [SPEAKER_04]: unfortunately there were no local no code to go back then so only developers were actually

[00:46:38] [SPEAKER_04]: bewildered you know there's definitely gonna be some potentially cultural sort of changes

[00:46:43] [SPEAKER_04]: in the way they think about allowing people the time to do that but you're right I mean

[00:46:49] [SPEAKER_04]: you've got to show them that you're going to be okay with this kind of experimentation

[00:46:54] [SPEAKER_04]: and that the time that you do save some of that is going to be reinvested in the workforce

[00:47:01] [SPEAKER_04]: in some capacity it's either like you said you know give them a little bit more you know flex time

[00:47:07] [SPEAKER_04]: or give them the means to actually take on some of those entire value everyone's talking about oh

[00:47:14] [SPEAKER_04]: look think of all the things you could do well let's we're that yeah or we save 20% so we're going

[00:47:20] [SPEAKER_04]: to cut 20% of the people and now you're right back where you started because now you got to pick up

[00:47:25] [SPEAKER_02]: the slack with their nonsense or whatever and by the way and good luck because again like

[00:47:31] [SPEAKER_02]: the competitor that doesn't do that and that basically supercharges that 20% of people

[00:47:35] [SPEAKER_02]: is going to be way better off than the company that cuts 20% of the workforce and so I think

[00:47:40] [SPEAKER_02]: that's because I know we're at time I think that's like the one thing that companies need to understand

[00:47:44] [SPEAKER_02]: is that it's still so early days for AI so I know everyone has stressed about it and the whole

[00:47:49] [SPEAKER_02]: BIO AI thing is a little bit intimidating and there are a lot of things that are very stressful but

[00:47:54] [SPEAKER_02]: if you pick it up today you can still be ahead of 80% of companies you can still be ahead of

[00:48:00] [SPEAKER_02]: basically everyone in your industry most likely unless I think you're working insurance where

[00:48:05] [SPEAKER_02]: they cut on a bit earlier like if you pick it up today and you sit with you again like with your

[00:48:09] [SPEAKER_02]: leadership team for two three weeks and you get everyone to understand it and use it and then you're

[00:48:14] [SPEAKER_02]: all allowed into your company and you allocate time for experimentation and you do town halls where

[00:48:20] [SPEAKER_02]: share use cases and you celebrate openly that people use AI so that we don't become these like

[00:48:26] [SPEAKER_02]: hidden sideboards where we're secretly using AI which is happening a lot which doesn't contribute

[00:48:30] [SPEAKER_02]: to anything positive at all this is still a huge opportunity but I do think the longer you

[00:48:37] [SPEAKER_02]: weigh the harder it's going to get because if you then step in six months from now and everyone has

[00:48:42] [SPEAKER_02]: already a year head start on you it's not as you're starting on that day you're really starting

[00:48:46] [SPEAKER_02]: behind because then you have to contextualize and grasp overall idea of what is AI and how

[00:48:52] [SPEAKER_02]: does it work in an organization all that then they've made become so they but right now is still

[00:48:56] [SPEAKER_02]: great great time to start I do have one final question for you before that you go when you

[00:49:01] [SPEAKER_04]: hear the title of this podcast elevate your AI cute what what do you think of when you hear

[00:49:05] [SPEAKER_02]: AI cute there's a big miss understanding about AI and that is that AI is technology which

[00:49:13] [SPEAKER_02]: technically it is but it's the most human technology that's ever been invented and so elevating

[00:49:19] [SPEAKER_02]: the AI in a way has to be about seeing this as a new coworker seeing this as a collaborator I think

[00:49:27] [SPEAKER_02]: as Ethan Molly calls it a co-intelligence seeing it as that co- and the other part of that is that

[00:49:32] [SPEAKER_02]: it doesn't mean that we only have to learn about or we only have to you know let AI

[00:49:39] [SPEAKER_02]: do what AI does best right I interviewed Dr. Alex and Rassamual was was regional technologist who

[00:49:47] [SPEAKER_02]: has been you know doing technology for a lot longer than I have and she said it I think really

[00:49:52] [SPEAKER_02]: well which is that yes AI is evolving really quickly but we have to evolve too and sort of the

[00:49:59] [SPEAKER_02]: people that you know you're seeing this very clearly again in the Sana data you saw it in a Microsoft

[00:50:04] [SPEAKER_02]: in both those cases you're seeing the power users report much more that AI is making them

[00:50:10] [SPEAKER_02]: more productive that AI is making their work more joyful that they get a lot of benefit out of AI

[00:50:15] [SPEAKER_02]: the people who use it like once a month or once a week and do something random with it and

[00:50:21] [SPEAKER_02]: give a really bad prompt and then get a really bad output and then toss it aside again and say

[00:50:25] [SPEAKER_02]: you see it doesn't work I think leveling up the AI IQ is really about understanding like what this is

[00:50:32] [SPEAKER_02]: and really starting to build that relationship and having to evolve on both sides and also evolving

[00:50:37] [SPEAKER_02]: yourself to say how do I work like that person in my class who said I realize how bad I am

[00:50:43] [SPEAKER_02]: at delegating not that I'm working with AI I think that's a really big part of it is that

[00:50:48] [SPEAKER_02]: you gotta evolve yourself you're gonna work very differently with AI then you work with people

[00:50:52] [SPEAKER_02]: before then you work with technology before which you've got to learn this dynamic you got to learn

[00:50:57] [SPEAKER_02]: or you're uniquely good at what you bring to the table and what AI can take over and what

[00:51:02] [SPEAKER_02]: relationship is between you two I think that's a really big part of it yeah absolutely fantastic

[00:51:07] [SPEAKER_04]: done thank you so much for taking time and this is a fantastic conversation as usual

[00:51:14] [SPEAKER_04]: thank you my pleasure