0:04 The last big tech bubble burst at the
0:06 start of this century and we may be
0:08 heading that way again.
0:10 >> The kind of investment wave in AI we've
0:12 seen is like probably nothing ever
0:13 before in history.
0:15 >> Hundreds of billions of dollars are
0:18 being spent on automating workplaces.
0:20 >> We have this amazing technology.
0:22 However, we're not seeing adoption fully
0:25 yet in every pocket of the the economy.
0:29 Only 1% of CEOs have a fully formed AI
0:32 strategy. With such high stakes, will
0:36 businesses see a return on investment?
0:38 I'm Isabelle Barrett. I lead the FTS
0:41 working at Grand speaking, presenting,
0:43 and writing about management,
0:45 leadership, and workplaces. In this
0:47 series, I'll explore some of the most
0:49 pressing issues around the future of
0:52 work and talk to senior leaders about
0:54 how they are making work better.
0:56 >> 3 5 years from now, I think things will
0:57 look quite different
1:09 I'm here at the Charter Workplace Summit
1:12 in New York in rooms filled with senior
1:14 leaders from some of America's biggest
1:17 companies. These are the people tasked
1:20 with AI roll out and preparing the
1:22 workforce for the skills needed for the future.
1:22 future.
1:25 >> Last is David.
1:28 >> Every 6 months a new model is dropping.
1:30 Every 6 months something shifts within
1:32 the marketplace where you have to stay
1:34 up to date. With AI, we're still in like
1:36 the very very very early days of
1:38 everything happening. We have this
1:40 amazing technology with the promise of
1:43 productivity enhancing gains. Roughly
1:45 10% of companies are fully starting to
1:47 integrate AI into their processes. But
1:48 there's going to be years of this
1:50 happening. We have to figure out exactly
1:52 how we can use it and where it makes
1:55 sense to use it.
1:57 A staggering amount of investment has
1:59 been made in AI over the last few years,
2:02 and it now accounts for a 40% share of
2:07 US GDP growth this year. Over 75% of
2:09 businesses worldwide are using
2:12 generative AI in at least one function.
2:15 But despite this, a study by MIT Media
2:19 Lab found that 95% of Gen AI pilots in
2:21 the workplace failed. I spoke with
2:24 editor-inchief of charter Kevin Delaney
2:26 about the state of AI rollouts in industry.
2:27 industry.
2:30 >> Think about how AI is different from humans.
2:30 humans.
2:33 >> Companies are adopting AI at two
2:35 separate speeds. You have the tech
2:36 companies who are actually quite far
2:38 along to the point where they think of
2:42 AI agents as co-workers. On the other
2:44 hand, you have companies that are still
2:46 getting their heads around what AI
2:47 adoption means. And these are the
2:50 companies that are still trying to get
2:52 their employees to use chat GPT or
2:55 claude. A lot of them are not seeing
2:58 gains in productivity at this point. So
2:59 you have these two extremes.
3:02 >> So we hear a lot about um the need to
3:04 upskill the workforce for AI. What does
3:05 that actually mean? Are people actually
3:06 doing it or are they just letting people
3:07 get on with it?
3:09 >> People are trying to figure out what
3:11 exactly that means. And I think part of
3:13 the challenge is that we don't actually
3:18 know what the ideal worker skills will
3:22 be in 3 years or 5 years as AI is rolled
3:24 out more pervasively. There's a lot of
3:27 discussion about is the ideal worker in
3:30 a more AI deployed environment someone
3:34 who is a real specialist in a field or
3:36 is it someone who is a generalist who
3:38 kind of knows a little bit about the
3:40 business and how business operates and
3:43 who can communicate clearly and knows
3:45 enough to be able to check what the AI
3:46 is bringing back.
3:48 >> So we need a lot more experimentation
3:50 and possibly failure.
3:52 >> Yeah. And so that's uncomfortable for
3:55 leaders too. To be comfortable with
3:58 failure is something that you are not
4:00 generally taught in business school.
4:02 Failure generally is something that
4:04 executives are allergic to encouraging
4:07 in their workers.
4:10 >> After a day of off thereord discussions,
4:12 panels, and big picture sessions, what's
4:15 emerged is that there's no clear path
4:17 forward for Genai at work. It's still
4:20 all to be decided. the reimagination of work.
4:21 work.
4:24 >> Leaders have spent billions on preparing
4:27 for an augmented future, but for what gain?
4:32 So, at the FT, we wanted to look at how
4:34 is this roll out actually going and what
4:36 are companies saying about how they're
4:37 using AI. And so, we did this massive
4:40 analysis looking at um SP500 companies
4:43 in the US. Um we went through thousands
4:46 of earnings reports and um regulatory
4:48 filings and the s- the results were
4:51 quite surprising. Um in earnings reports
4:53 CEOs would often say you know AI is
4:55 amazing. It would bring incredible
4:57 productivity gains a Cambrian explosion
5:00 of innovation things like that. Um but
5:02 then in the filings which to be fair
5:04 tend to be more measured and um risk
5:07 averse no one really had anything
5:09 concrete to say of how they're actually
5:12 using it. And uh in those filings, the
5:14 risks outweighed the benefits very very
5:17 clearly. If you look at the SP500 index,
5:19 it's obviously going up, but a lot of
5:21 that growth is driven by seven big tech
5:23 companies. And the other companies on
5:26 the SP500 haven't necessarily grown that
5:29 much when they've said they use AI. AI
5:32 use is often phrased in their filings as
5:34 being something quite abstract. Um they
5:36 talk about productivity but don't really
5:38 offer any concrete examples of how
5:40 they're using it. Coca-Cola is one
5:41 example where in their earnings reports
5:43 they raved about how they're using
5:45 generative AI to transform their
5:48 business. Um, but in their filings, the
5:49 only example they could give was using
5:52 generative AI to create a Christmas ad.
5:54 It's definitely a mixed bag. The growth
5:56 of AI has led to a boom for
5:58 consultancies and learning platforms who
6:01 are keen to show business how to harness
6:04 the powers of AI at work. I visited the
6:07 HQ of AI upskilling platform Multiverse
6:10 and met with their CEO and founder Euan
6:13 Blair. So what are the ways in which
6:14 companies I guess your clients are
6:17 engaging with AI skills? Are they
6:19 hesitant? Are they all in? How is it
6:20 what does it look like?
6:22 >> So I I I think it's it's almost the kind
6:24 of polar opposite of hesitant. The kind
6:26 of investment wave in AI we've seen is
6:28 like probably nothing ever before in
6:31 history. So the big challenge a lot of
6:33 organizations are facing is how to turn
6:36 kind of potential AI gains into actual
6:38 realized AI gains. And that's where the
6:39 kind of training gap comes in because
6:41 what a lot of people are doing with AI
6:42 at the moment is the equivalent of
6:44 having an iPhone and just using it to
6:45 send text messages and make calls,
6:47 right? They're missing out on loads of
6:50 the capabilities that these tools
6:51 actually have. So we've seen a lot of
6:55 companies spend a lot of money on AI and
6:56 >> really a lot of money
6:59 >> and there haven't been right
7:01 >> particular productivity gains that I'm
7:02 aware of.
7:04 >> What's what's the where's this gap?
7:05 What's the gap?
7:08 >> We've seen accounts teams for example um
7:11 process invoices 50% more quickly and
7:13 with half the number of errors because
7:16 of introducing AI. We've seen um
7:18 software engineering teams increase
7:21 their speed of shipping code uh by 75%
7:23 in some cases. Those are big tangible
7:25 things that do actually have an impact.
7:26 One of the reasons we're not seeing
7:29 gains at the kind of big macro level yet
7:31 in terms of economic growth is this sort
7:34 of training and capability gap. Because
7:36 with previous versions of software, it
7:38 was often deemed enough to go and invest
7:41 in the technology and then over a period
7:42 of several years, people would figure
7:43 out how to use it and where to use it
7:45 and everything would be okay. The
7:47 difference this time is the inherent
7:49 capability of the systems is so much
7:50 greater. You need a lot of training to
7:52 be able to kind of fundamentally change
7:54 the way you work, but also the amounts
7:56 being spent are so much greater. So the
7:58 stakes are higher and that kind of
7:59 creates this this perfect set of
8:01 conditions where people realize the
8:02 people who spend the most on AI are not
8:04 the ones who are going to win. It's
8:05 going to be the people who have the most
8:08 AI enabled workforce and that's the kind
8:10 of space multiverse is playing in.
8:11 Everyone feels like they're behind the
8:13 curve when it comes to AI and they all
8:14 feel like they're not doing enough and
8:16 could be doing more. And that is kind of
8:18 creating this sort of um it's not even a
8:21 hype cycle but it's a just a desire to
8:24 kind of do more faster.
8:25 So when you think about the financial
8:27 gain of AI, a lot of that money is
8:29 flowing into tech companies. AI
8:31 companies, management consultants, and
8:34 companies adopting AI aren't necessarily
8:37 seeing those magical financial gains
8:38 that they were promised. But it's worth
8:40 bearing in mind that it's still really
8:42 early on. Um it's really early in the
8:44 deployment stage of these technologies.
8:45 Just a few years ago, they were still in
8:48 the lab. And so we have to be patient.
8:50 But obviously the question is how long
8:51 do we have to wait? Obviously,
8:53 businesses are hoping that these use
8:55 cases and gains will come sooner rather
9:00 >> The number of people turning to
9:03 commercial AI platforms on a daily basis
9:06 has been astronomical. The rate of
9:09 adoption for chat GPT alone outpaces the
9:11 rise in use of the internet when it was
9:14 first launched. But the gulf between
9:17 work related and personal usage is growing.
9:24 So what you often see are these shadow
9:28 use cases where official uh corporate AI
9:30 initiatives are often untouched or
9:32 unused and people just use the AI tools
9:34 they like and this is often because
9:36 there there hasn't been necessary
9:38 communication between leadership and
9:39 staff about what they need and what kind
9:41 of tools they actually want but
9:43 different rules apply at workplaces
9:45 right workplaces often have sensitive
9:47 information or accuracy really matters
9:49 and so you have to pay attention to the
9:51 fact that these models often do make
9:53 factual mistakes and that could be
9:55 really embarrassing or even catastrophic
9:57 for an organization. So, every
9:58 organization needs to be thinking about
10:00 this and thinking about how these tools
10:03 apply to them and what they want their
10:05 employees to know about how to use them.
10:07 Some of the biggest challenges that
10:09 businesses face are that they just
10:10 aren't ready for this digital transformation.
10:12 transformation.
10:15 To use AI well, you need good structured
10:18 data, good cyber defenses, and perhaps
10:22 most importantly, AI literate staff.
10:24 I went to Google's newest campus in New
10:27 York to meet Amanda Broofphy, director
10:30 of Grow with Google. It's Google's
10:31 professional training arm and offers
10:34 courses to businesses and individuals on
10:37 how to use AI. What's your advice for
10:40 leaders who have maybe a cohort of staff
10:42 who are still very skeptical of AI or
10:45 slow to adopt? I think you need to find
10:46 how to make the AI work for that
10:48 specific person in their role and what
10:51 they're doing. What makes AI so powerful
10:53 is when you can translate it into what
10:56 you are doing today and now that's
10:59 specific to you. So if a marketer is
11:01 trying to use AI and we're helping them
11:03 figure out how to use this to write
11:05 social captions for their social media
11:07 posts for customer service to think
11:09 about how they use this to write
11:11 responses back in a way that's polite
11:13 when someone's getting upset and it's
11:15 escalating. Making it custom to that
11:17 person and role is when you actually see
11:19 the real benefits. And so being able to
11:22 test that for you is what allows that
11:23 skepticism to go away and see the real
11:25 benefit from it. One of the big problems
11:28 with AI rollout is that people aren't
11:30 really getting trained. So what do you
11:32 say to employers? You need both the
11:35 technology and the training, right? You
11:36 need the tools in the training. It's an
11:38 and not an or. And so what we're finding
11:40 is just rolling out the technology isn't
11:43 enough. We have a course, the Google AI
11:45 essentials course. And what we've seen
11:47 is that being able to teach people how
11:49 to use the technology, how to prompt and
11:51 make sure that they're using it in an
11:53 effective and reliable way helps them to
11:55 get to use it every day to upskill and
11:57 reskill. What I think makes AI different
11:59 is it's not learning about it. It's you
12:02 have to use it and do it. You have to
12:04 have the daily practice to make it a
12:06 regular habit in the work that you do.
12:07 It's one of those ones that you need to
12:10 have the intrinsic interest to be able
12:12 to see the value of AI in the day-to-day
12:14 of your professional and personal
12:17 benefits and the employer needs to be
12:19 able to deliver and have this available
12:22 for employees so that people are
12:24 consuming this information for their for
12:27 the company. And what's your best tip
12:30 for anyone watching this who wants to
12:32 get better with AI in their job? You
12:34 need to be able to prompt the AI
12:35 effectively to make sure you get the
12:37 desired output that you want.
12:39 Highlighting pieces like who's the
12:40 audience you're trying to reach? What's
12:42 the goals in the context? What's the
12:44 reference materials? And so being able
12:47 to prompt AI effectively is critical to
12:50 get the output that you will then see to
12:51 make this a regular habit and the
12:52 efficiencies that you want.
12:54 >> So do you think journalists make good
12:55 prompters? I bet we do.
12:57 >> I think you make excellent prompters
12:58 because you're good at the questions.
12:59 It's exactly what it is. You understand
13:01 who the audience is, what the questions
13:03 are. I think journalists are excellent prompters.
13:04 prompters.
13:06 >> Perhaps not surprisingly, the tech
13:09 sector has been an enthusiastic AI adopter.
13:11 adopter.
13:14 I met with Cisco's UK and Ireland CEO
13:17 Sarah Walker to see how it's working for
13:21 them. So internally at Cisco, it's a
13:23 tech company ahead of the curve.
13:26 >> What does AI usage look like generally
13:27 internally here?
13:29 >> Really, really broad spectrum. So if I
13:30 think of it in terms of our product
13:32 development, um things like our WebEx
13:35 platform have AI agents built in and
13:37 they do some fabulous things which have
13:39 made my life a lot easier and more
13:41 efficient. We've also then got some
13:43 really great platforms that we use as as
13:46 employees. There's different levels of
13:48 adoption of that. As you can imagine,
13:51 some are super proficient, some still
13:52 are trying to get to grips with with
13:54 what that means. But that's where
13:56 adoption becomes key because for for us
13:58 to really capitalize on the efficiencies
14:01 that those investments um can and should
14:04 deliver um our next task is, you know,
14:07 how do people adopt that and and make
14:09 that a part of their kind of DNA and how
14:10 they operate on a daily basis.
14:11 >> From talking to people, there's a kind
14:13 of you know, people bring in AI systems
14:15 and then they don't really monitor
14:17 adoption. How can leaders get over that?
14:19 >> Well, first of all, you have to lead by
14:21 example because my team will never adopt
14:23 those sorts of platforms if I'm not
14:25 talking about it and using it myself.
14:27 So, we did a master class actually with
14:28 our senior leadership team across the
14:30 UK. And I speak really really positively
14:33 about pro workforce and pro AI. It's not
14:36 an eitheror and using AI doesn't mean
14:38 that at some point in the future your
14:40 role will be replaced by it. This is
14:43 about using these applications to say
14:45 how do you become more efficient in the
14:47 things that you can and should automate.
14:49 And candidly, it's human nature to want
14:51 to find a quicker and more efficient way
14:53 to do things. We've always been like
14:55 that just because it's now called AI or
14:57 that's more kind of broadly known. We
15:00 shouldn't be we shouldn't be fearful um
15:02 of that. But it is a a common mistake
15:04 that businesses make that thinking just
15:05 because you've got the applications or
15:07 the opportunity that adoption will
15:10 follow. Everyone should definitely try
15:11 these tools. They're a lot of fun to
15:13 play around with and that's the quickest
15:15 way to learn how these might work for
15:16 you or how they might not work for you.
15:18 You have to use them for use cases where
15:20 the tools are actually beneficial
15:22 instead of expecting it to be some sort
15:25 of magic wand that can fix all problems.
15:26 And so currently we're operating in the
15:28 fact that this all will work and it'll
15:30 lead to amazing things in the future.
15:32 But if that were to change, if this were
15:35 a ma massive bubble that were to burst,
15:37 um the reality is that a lot of these AI
15:39 experiments, only the use cases that
15:42 actually work and that bring benefits to
15:44 employees will stay. Everything else I
15:46 can't really see surviving.
15:49 The challenge of AI rollout in
15:50 workplaces doesn't have a
15:53 one-sizefits-all solution. Businesses
15:56 need input from staff, but equally those
15:59 staff need support and training from
16:01 their leaders if any of us are to
16:04 realize the financial and productivity
16:07 gains that AI promises. I'm old enough
16:09 to remember when the internet rolled out
16:12 in the mid 1990s, and it seems to me
16:14 we're at a very similar early stage of
16:17 the cycle with Gen AI. There's a lot of
16:20 boom and bust to come and with it
16:23 disruption and I hope excitement at work.
