With the rapid advancement of artificial intelligence (AI) technology, many businesses are turning to AI to increase their efficiency and productivity. McKinsey’s report on the state of AI for 2022 provides valuable insights into how AI will shape industries in the coming years.
What are the key lessons?
- AI is slowly reaching the peak of the maturity curve. That is, it’s no longer a special new sexy technology, but rather, an established reality.
- Talent shortages are still prevalent. However, given that this report reflects upon 2022, it remains to be seen whether these trends will persist in 2023. Also, who knows whether the new version of ChatGPT will reduce the need for junior developers!
- The majority of use cases involve straightforward applications like customer analytics, instead of fancier applications like generative AI.
- AI is primarily used to cut down costs and optimise efficiency, so it acts like a support function in most organisations.
Learn more by watching the video below!
Transcript: State of AI in 202
00:02 Hi, everyone, It’s this time of the year again when all the big consultancies, they released these reports about, you know, what happened in the past year.
00:12 And in today’s videos. In today’s video, I’m gonna be talking about the state of AI in 2022, A report by McKinsey.
00:20 You can find the link in the description below. So this is a very interesting report for a few different reasons.
00:27 First of all, McKinsey in this report, they reviewed the last five years and something which is clear, and I can also attest to that based on my experience, is that AI adoption is flattening out.
00:39 So while it’s higher it seems that adoption is slowing down and there isn’t a simple, okay, so AI has matured.
00:48 So as as you can see here, responses show an increasing number of AI capabilities embedded in organizations over the past five years.
00:57 So what this means is that the market for AI is different. So, whereas AI was the hot new thing, let’s say, five or 10 years ago, and it’s more of a mature technology like databases, maybe some things like chat d PT are gonna be revolutionary, but most industries, they don’t really care about this type of technology, even if it’s cool, okay, maybe they will care in 20 years.
01:18 But most of the use cases, they’re actually of AI in the industry. They’re much simpler. So when this means there means a few different things if you’re a graduate looking for a job I guess not the market is bigger, but at the same time, there’s more competition, right?
01:33 Simply because it’s a more mature market. However, ator demand is still more than supply in terms of service providers and and, and workers.
01:45 Secondly, if you’re a startup, I guess you know, you can still do many things in this area, but it’s not like, you know, it usually means if you are looking to create like new solutions, okay, it’s most likely that many companies have their own teams.
02:00 They’re creating their own products internally. However, the same time, if you’re a service provider or an entrepreneur, it’s probably gonna be easier to speak about AI because more companies now really understand what AI is about.
02:14 And when we look into functionalities, like how AI is being used, we see things like robotic process automation. So automation is obviously the biggest one.
02:25 Computer vision, natural language understanding virtual agents. And these are really things which have been commoditized to a large extent. So this is a space dominated by IBM and Google and, and the likes which is quite cool, right?
02:38 Which is fine. And it’s still you know, and this is like quite interesting because it means that, let’s say most of the low hanging crew is more, you know, is a very competitive space.
02:51 So if someone, for example, is looking to, you know, join this area to learn data science, or if someone is looking to come up with new ideas in this, this area, probably that have to look into like, let’s say new types of of applications like robotics maybe na like natural language generation, et cetera.
03:10 What’s interesting is that many companies like a good percentage of them, they’re using transfer learning or generative ai. Obviously this is at the bottom of this scale, but still it’s you know, 16 and 11%, which is not a small percentage for, for such and advanced type of technology which I find really, really exciting.
03:31 And when we look into the most popular use cases we can see like pretty vanilla stuff, okay? And I think this, one of the things that has come out of, you know, this of AI adoption in the last five to 10 years, and that is that most of the, let’s say, value in AI is in really simple things.
03:52 Okay? So it’s, it’s, it’s not usually most companies, they’re not AI first. So AI data science, they like support functions.
04:00 They plug themselves into existing functions, and they’re like, improve efficiency. So, so we see that most of the use cases they’re around optimizing service operations, creation of new AI products customer segmentation, AI based enhancement, like adding recommender system optimizing features, which could be things like data-driven analytics, things like that.
04:23 So really van vanilla stuff not anything too fancy. And I guess this is also a good lesson for many of us in this space who’ve been, you know, in this space for a long time.
04:33 And when we saw AI being adopted by organizations, we imagined that, you know, it could be transformational in, you know, all kinds of ways, but in the reality, the true transformation is not so much about doing let’s say new things, but more about taking things which already exist, like processes in a company, for example, and improving them.
04:55 Okay? I guess a good thing for Evan who’s in this area is that investment is, is rising. And I guess that AI is not unique in this.
05:05 I think it’s the same with web three. Okay? We’re in a bird market, but still with tech in general even if big tech has its own issues, I personally don’t see that, you know, that the trajectory is slowing down in the next 10 years.
05:19 I don’t think it’s going to happen. Because even if, let’s say adoption in developed countries, things like AI blocked and et cetera reaches a pink in the next 10 years, and there’s still many emerging markets, okay?
05:31 So we we’re going to see investment rising and rising, and that’s a graph, particularly like in this report because it, it shows how organizations are using AI in order to essentially make something out of it, right?
05:45 And there was this article like on the things it was five or 10 years ago in, in Harvard Business Review, saying that they’re defensive and offensive cases for data.
05:55 Okay? So offensive is when You, you try to make profit defensive is when you try to cut down costs. And you can see here a breakdown of the different functions and whether they’re using AI to decrease costs or increase profit.
06:12 And you can see that AI can actually be used in all of those cases, both in a defensive way and in an offensive way which I find really cool.
06:21 And probably the most interesting finding is that supply chains see the greatest benefit. And maybe one of the reasons is that supply chains can be very complicated and inefficient and brittle, as you saw during the disruption with Covid and Lockdowns and the war in, in Ukraine.
06:37 Okay? And in terms of the important mitigation of AI related risks it seems that most companies, I mean, there’s not a big change over the last few years, it seems cyber securities, you know, one of the biggest concerns, and I think that the current situation with the war, unfortunately and also global politic tensions are only going to increase to increase that.
07:02 Okay? And let’s go down below and see about the AI leaders where really this, this is like an interesting graph, which pretty much shows things that us who work in this space already know.
07:17 Some of you who’ve been following my work, you know, my books, the Decision Makers Handbook to Data Science also the new book about emerging technologies where I talk extensively about data strategy.
07:29 And, you know, the work I’ve been doing with Tester Academy and pretty much the things that I’ve been teaching there, I see them here as well, right?
07:37 That usually the AI high performers, they have a clearly defined AI vision strategy. They have a roadmap. Senior management is aligned with the goals of the organization thinks, which are true not only for ai, but pretty much for any technology.
07:53 Okay? If you wanna know more about this, if you wanna, this is pretty much true, I guess, for any kind of digital transformation initiative.
08:01 Okay? So if you wanna know more about this, just, you know, shoot me an email. I have lots of content to share with you.
08:11 So I mean, this simply is, I mean necessarily this, this data corroborates with what US practitioners in this area already know.
08:20 And now let’s move on. And in the last part of the report they talk about talent and the number of positions they hired for this report reiterate what many people know that demand for tech workers is larger than the supply.
08:38 However, I think a large part of this report of this survey was conducted before the layoffs, which happened towards the end of last year.
08:46 So it remains to be seen what’s gonna be the effect in the supply We have, we have now some very highly qualified people who’ve been laid off from various big tech companies looking for jobs.
08:57 We also see maybe slow down of investment this year. So maybe things are gonna normalize a little bit. But again, we see that software engineers they’re obviously the first hire and then data engineers, they’re still surpassing data scientists.
09:12 Which again, comes at no surprise because I guess data scientists can work only when a data engineer has built the pipelines.
09:20 And that’s like an interesting finding. I guess in the next few years, th this is going to normalize. So I would expect data engineers to be on the same level in terms of salary and hiring patterns as data scientists.
09:35 And it seems to be, and also in terms of hiring most most companies say that hiring treat much, any of those for any of those roles is quite challenging.
09:48 Okay? And I guess it’s quite challenging. Again, that’s an effect of, of demand and supply. So I guess that’s good news for anyone who’s looking for a job in this area because it means that you can still get a premium in, in your salary, and it’s still a good time to really join any of these professions, right?
10:05 Whether it’s data scientist, data engineer, et cetera, et cetera. Something that’s also stands out is that the AI high performers they’ve hired from a variety of like roles and it’s usually easier for them to hire than other companies, I guess because they’ve actually done the foundational work so that it’s easier for someone to join.
10:33 It’s more, it’s more attractive environment for data scientists to see more opportunity there. Okay? And this really goes back to, you know, that really as an organization, you need to have the right strategy and you need to have a clear vision.
10:46 And really it’s usually top organizations that also attract top talent, okay? So if you want to be a top organization, then you need to start thinking like one an alternative seems to be to hiring, seems to be upskilling with, I mean, you know, me, my, I’m also training data scientist, and I really believe in upskilling.
11:06 I think it can definitely help cure part of the problem. But I still think that many companies might be a bit defensive about upskilling and it remains to be seen whether it can be like the solution to, to the current demand supply issues.
11:24 And finally ai like high, high, high, high AI performing organizations, they seem to be hiring a lot from top tier universities with, you know, doesn’t come at a surprise as well as global technology companies.
11:40 It’s the same with all big organizations. If they can hire people from Harvard, Cambridge, et cetera, they’ll just go and do it.
11:46 But I don’t think all organizations can hire this kind of graduates or candidates, and not everyone should, right? And I don’t believe that you need to always pay a huge salary or a premium to get these people, because unless you do the foundational work in your business, in your organization, you may not trip the benefits.
12:05 And that’s why I really believe that an organization shouldn’t focus just on hiring graduates or professionals with flashy names, but actually see what they can get, how they’re going to use this talent, and what’s the expected ROI for ai.
12:22 And then there’s some working in terms of diversity, which I guess it’s more of a problem in, in, in tech in general, not just in in ai.
12:30 And that’s pretty much the end of the report. Okay. So that was a bit of a short video. I don’t want to expand upon it forever.
12:39 But I guess if you were to summarize it, it would be that a couple of key findings. First of all, AI is maturing.
12:45 So in the next few years will be facing a different market with lots of opportunities for those who are seeking a job.
12:53 But maybe in terms of innovation, startups, new entrants, it starts to get maybe a little bit crowded. The good thing, the good news is that there’s still lots of investment in this area, which is great.
13:03 And also conversation with companies can be easier because now they’re more aware of ai, what it is, what it can do.
13:10 And then but if you’re, if you looking to get into the serve, you’re a graduate, it’s a great time to be in.
13:16 And my, and also something else, which again, I’d like to reiterate is that it, if you are a business owner, entrepreneur, a senior professional manager, what’s going to make the difference is not, let’s say a new cool platform is not a new cool tool, but it’s really the foundational work in terms of understanding how AI fits with your overall business strategy.
13:39 So if you’re interested in becoming a data scientist, or if you are a manager or entrepreneur and you’re interested in using AI in your business, then please make sure to get in touch, and I’ll be more than happy to help you out.
13:51 Thank you. Let’s have a great 2023 and see you soon.