AI Isn’t Delivering (And CEOs Know It)

We're well into the AI era now and the first reports are in. CEOs have been shouting from the rooftop about how AI is transforming the world. It's in their

words a gamecher and for some like the head of AI at Microsoft, hundreds of thousands of jobs, white collar jobs are going to be gone within 18 months. Are

companies making money hand overfist because of AI, growing revenues, slashing employee costs, and reaping the benefits? No, not really. The reality is

benefits? No, not really. The reality is that companies are spending heavily on AI, but most still can't clearly measure whether it's paying off. The reality is

that companies are spending tons of money and man-hour on finding ways to reap the benefits of AI, but no one really knows if it's working and whether or not they're saving anything. And

also, very few know how to scale early wins into companywide change. to get a good idea of what's going on right now.

The good news is we've got tons of consulting firms and tech analytics firms that have data on the actual effects of AI in the here and now. PWC,

Deoid, Gartner, McKenzie, they all show the same pattern. Most CEOs are not yet seeing meaningful growth from AI.

According to PWC, fewer than 12% of CEOs say AI has led to measurable revenue growth or cost reduction. Less than

onethird believe that they could see an increase in revenue down to the use of AI in some shape or form. This is across all industries surveyed in e-commerce digital first AIcentric companies. The

revenue generation is more apparent.

AIdriven advertising personalization customer-centric product suggestions and also companies offering software solutions with AI upgrade options are expanding their revenue generation opportunities. Outside of digital first

opportunities. Outside of digital first players like Amazon, Google and Meta, these remain relatively niche cases. For

most traditional companies, AIdriven revenue growth is still more aspiration than reality. I'm Mike and this is my

than reality. I'm Mike and this is my humble channel about everything business and careers. If you are finding my

and careers. If you are finding my content, this video and other videos engaging, then you can support the channel by clicking the subscribe button. It helps out massively. Thanks a

button. It helps out massively. Thanks a

ton. When it comes to the cost side, the picture is basically the same. AI

adoption is strongest in support functions like finance, HR, fulfillment.

The picture depends on the type of company. Long-standing incumbent type

company. Long-standing incumbent type companies like McDonald's, GE, and Sony are having a harder time realizing revenue and cost gains from AI.

According to PWC, 25% of companies reported cost savings due to AI, but 22% reported that they had increased their costs. McKenzie also found that AI

costs. McKenzie also found that AI hasn't infiltrated any single business function by more than 10% that applies to HR, finance, software engineering.

Less than 2% of companies stated that they had fully scaled a solution within one function like marketing or finance.

Less than 2%. Depending on the actual function, between 70 and 90% of respondents stated that they had not taken any steps at all to start up some

efforts with Agentic HR. The reality

seems to be that most companies are at a stage where they state that AI has some level of priority or importance to them.

That they're using AI in some way in as these reports call it at least one function. And a tiny fraction of

function. And a tiny fraction of companies have a fully scaled generative solution implemented in at least one business function. There was identified

business function. There was identified no incumbent company that had fully restructured itself around AI at this point in time. Whether it's Gartner, Deoid, PWC or McKenzie, the picture's

the same. It's trials, its pilots, it's

the same. It's trials, its pilots, it's localized attempts to find a way to get the benefit of AI in some way or another. AI isn't transforming revenue.

another. AI isn't transforming revenue.

It isn't cutting costs at scale. So, the

question is, why is every CEO acting like it is? As a largecale business transformation tool, AI is still in its infancy. We're just not at a point where

infancy. We're just not at a point where there are many broadly available, commercially viable and scale solutions that companies like GE, Boeing, and

Galaxos Smith Klein can latch onto. CEOs

and boards are demanding AIE strategies right now at a corporate level. And to

be clear, we're talking about companies that are not the ones trying to sell AI.

These companies and their leaderships have an idea that AI is going to be massively important for their business.

They don't quite know how. They don't

quite know where. There's a core belief that AI will eventually cut costs or unlock new revenue streams. This belief is driving a strong push to get going

with AI adoption across the business.

But aside from the vision that AI is integrated, integral to the business, the jump from strategy to tactics and operations is just weaker. Usually

strategy is about direction on the one hand and making deliberate choices. So

saying no to certain things on the other hand. Right now, it's the idea of AI

hand. Right now, it's the idea of AI that's driving strategy, not clearly defined business cases. That's why

priorities at the operational level feel scattered. Companies are struggling to

scattered. Companies are struggling to identify real savings, and when they do, it's hard to prove AI cost them. Deote

shows this in their report pretty clearly. And a lot of you probably know

clearly. And a lot of you probably know this from your daily work. You are

probably using some sort of LLM. You're

using it to help you write a company newsletter to make nicer PowerPoint slides or you've developed a little AI agentic model that helps you with some financial processes that were quite

timeconuming for you. You know that AI is helping you but how much how much more efficient are you and how can this efficiency be quantified? As Deoid

pointed out in customer service, AI is improving response times and user experience, but connecting that improvement to actual revenue growth is proving difficult. It's also proving

proving difficult. It's also proving increasingly difficult to identify scale benefits. For most companies, AI might

benefits. For most companies, AI might be a strategic pillar, but apart from a very small select number of businesses with an even smaller subset of tactical initiatives. Right now, most AI is being

initiatives. Right now, most AI is being driven by the logic of go use it and make us more efficient. These localized

unscaled solutions might be benefiting the business but it's proving hard to quantify their value generation and it's hard to scale these benefits across the

business and across an entire solution area. As one executive stated to deote

area. As one executive stated to deote everyone is asking their organization to adopt AI even if they don't know what the output is. There's so much hype that I think companies are expecting it to

just magically solve everything. The

picture that's being painted right now is clear in its lack of clarity.

Businesses are saying use AI and they're saying to staff, be AI fluent, but very few are generating revenue from it. The

amount of businesses saving money is offset almost equally by the amount spending more money while the majority have seen little to no change. The

strategy might be companywide, but the AI effort is localized, functional, and highly operational. There still doesn't

highly operational. There still doesn't exist a framework to measure the impact of AI consistently and it's not just a question of implementation. The biggest

problem is that no one really knows what the potential is. So almost all projects are a mix of identified target areas and experimentation. Amazon, Google, Meta,

experimentation. Amazon, Google, Meta, the AI giants, they are still mainly investing rather than at the stage of reaping the cost benefits of AI. They

are firing people but that's not purely because of AI. That's because they are generating so much money on ad revenue and their other businesses are so mature that they need less people to run the

business. Their cost savings are being

business. Their cost savings are being funneled into AI data centers. But

importantly, they're not indicative of companies outside of the sector in general. For everybody else, there's a

general. For everybody else, there's a German phrase which states which in English means we got to have a thing just like that thing. Which

basically says it all. It's a phrase that's meant to capture that situation where you see what the next door neighbor is doing and saying to yourself, "Oh, I I I want to do that. I

I want to copy that. I have to do that, too." There's no deeper strategic

too." There's no deeper strategic understanding of how it will benefit you. You just see that it's there and

you. You just see that it's there and you want it too. Everybody hears that AI is going to lead to all sorts of benefits, but nobody right now really

knows quantifiably what they are.

problem with AI adoption right now is down to four core problems that all companies seem to be struggling with.

Problem one is that it's a solution without an identified link to a problem.

So if you have rats in your shed, that's your problem. You know that there's a

your problem. You know that there's a product called rat poison that can solve that problem. Rat poison is explicitly

that problem. Rat poison is explicitly designed to deal with rats. The problem

is for companies right now, AI is not rat poison. In this analogy, it's a

rat poison. In this analogy, it's a hammer. You're standing there with a rat

hammer. You're standing there with a rat problem and somebody hands you a hammer and says, "Use this hammer to solve your rat problem." Now, is there a way to

rat problem." Now, is there a way to solve the rat problem with the hammer?

Probably. But is it to stand there really still and bash the individual rats on the head with it? Or is it to build some sort of hammer rat trap smasher thing? Or is it to board up the

smasher thing? Or is it to board up the shed so the rats can't get inside? When

companies buy enterprise solutions like Salesforce or invest in a factory full of CNC machines, it's because those investments solve a defined problem.

Right now isn't a solution in itself.

It's a tool towards a solution. Problem

number two, at a business level, you are expecting staff to come up with solutions, which is admirable. People

are smart, they're innovative, they're creative. But the reality is in most

creative. But the reality is in most companies, most people are just good at whatever they do. They're not software developers. They're not project

developers. They're not project managers. They are not process

managers. They are not process improvement specialists. Asking your

improvement specialists. Asking your bookkeeper to massively transform your bookkeeping process through AI is not something they're skilled at. They might

be able to automate one or two local processes, but they're likely not going to be fundamentally changing processes for greater efficiency gains. Right now,

in a lot of companies, a lot of employees are treating AI more like Microsoft Excel than Microsoft Cloud.

The point being, cloud is an enterprisewide scale solution that solves a host of business challenges while also opening up a mountain of possibilities. Excel is a day-to-day

possibilities. Excel is a day-to-day tool which is indispensable but usually doesn't deliver scale solutions. It

delivers localized solutions within specific use cases. In every company I've ever been in, the ubiquitousness of Excel or G sheets and the way they used to support all sorts of business tasks

are both a blessing and a curse. Their

blessing is the ease of use and customizability. But the curse is when

customizability. But the curse is when you want corporatewide insights, scalability security and replicability.

Excel sheets just don't do that. And

right now, neither do a lot of those localized AI solutions. Problem number

three, you are in some sectors seeing AI enterprise solutions being developed, but they are few and far between. LLMs

with a specialized focus towards law firms and legal work exists. LLMs and

customized enterprise agents specialized in creative work are also showing up. If

I'm Mercedes and I want to use AI to revolutionize the way I work with R&D, I might have to invent the Agentic AI solution myself. Potentially, no one yet

solution myself. Potentially, no one yet has a bespoke solution that's thoroughly tested and commercially proven with other buyers as well. In many sectors,

no one has a bespoke, thoroughly tested, and commercially proven solution for me as a company. There's nothing wrong with AI being a hammer as such, but to really

get the scale benefits, AI needs to move from being a tool to a solution to specific problems. Problem number four, although companies to some degree are stating that some AI has led to

increased revenue, what you're still mostly seeing is a cost focus. Save

money, cut employees. That's where the real CEO dream scenario is. And from a business perspective, that's fine.

Cutting costs is something investors like. But what we're seeing quite

like. But what we're seeing quite consistently is that the options for revenue generation are more limited. If

you want societal transformative change, then you want a tool in AI that is additive, that creates value, not just reduces cost. Cutting cost in the short

reduces cost. Cutting cost in the short term is fine, but down the road, you're just going to be paying more for electricity to power those data centers.

The cost side will always eventually equalize. The focus has to be on the

equalize. The focus has to be on the innovation and growth side. And that's

where AI is currently proving to be lackluster in a lot of firms. Even the companies building the models are still figuring out how to turn massive infrastructure spend into durable

profit. Not that it won't happen, but

profit. Not that it won't happen, but we're a far way off. At this point in time, AI to me feels like it's everywhere in everything all at once,

but it feels invisible. It feels

intangible. It's still for a lot of companies aspirational in nature and not transformational. We all standing there

transformational. We all standing there with the hammer and thinking of 13,000 different ways to kill a rat. Those were

my thoughts. Share yours in the comments. Be sure to subscribe. See you

comments. Be sure to subscribe. See you

soon.