Why so many people suddenly became AI consultants, and what is driving the market behind them.

AI usage disclosure: AI was used for research and references as well as rewriting the idea from notes to a structured article.

A few years ago, LinkedIn was full of digital transformation experts, blockchain consultants and Web3 strategists.

Today, many of those same conversations revolve around AI. New titles seem to appear every week: AI Strategist, Prompt Engineer, Agentic AI Architect, AI Transformation Advisor and Fractional Chief AI Officer.

It is easy to dismiss this as another technology hype cycle. But I think something more interesting is happening.

According to the 2024 Microsoft and LinkedIn Work Trend Index, there was a 142-fold increase in LinkedIn members adding AI skills such as ChatGPT and Copilot to their profiles during the period measured in the report.

That is a remarkable shift in how people present their professional skills.

Of course, adding AI skills to a LinkedIn profile does not make someone an AI consultant. Nor does it mean they lack genuine expertise.

But it raises a question.

Why has AI made it so easy to enter a market where people are selling professional technology advice?

I think the answer lies in several economic forces coming together: technology layoffs, cheaper AI tools, pressure on companies to adopt AI, and a growing gap between what people can demonstrate and what they can actually deliver.

1. The technology job market changed

The timing of the generative AI boom is important.

In November 2022, ChatGPT became publicly available. Around the same period, the technology industry was entering a major restructuring.

According to Challenger, Gray & Christmas, US technology companies announced approximately 168,000 job cuts in 2023 and another 134,000 in 2024.

That is more than 300,000 announced cuts over two years.

These layoffs were not all caused by AI. Companies were correcting earlier overhiring, responding to higher interest rates and reducing costs.

But the effect on the technology labour market was significant.

Imagine spending fifteen years working in technology consulting, project management, digital marketing or software delivery.

Suddenly, your company restructures. Traditional job opportunities become more competitive. Consulting clients are watching their budgets.

At the same time, AI is everywhere.

Companies are looking for AI expertise. Conferences are filled with AI presentations. Executives want to understand how to use it. New consulting opportunities are appearing.

For someone trying to remain relevant, learning AI and repositioning their services makes economic sense.

I do not think we can claim that most new AI consultants came from layoffs. There is no reliable evidence for that.

But we can reasonably say that a difficult technology job market and a fast-growing new field created strong incentives for people to change direction.

And AI made that change unusually easy.

2. It has never been cheaper to build something impressive

This is where generative AI differs from many earlier technology waves.

Think about what it took to become a database specialist, cloud architect or machine learning engineer.

You needed to learn programming, understand technical concepts, work with real systems and gain experience solving problems.

Those requirements have not disappeared.

But generative AI has made it much easier to produce something that looks technically sophisticated.

Today, someone with limited programming knowledge can use AI tools to build a chatbot, create a website, automate a business process or connect an LLM to company documents.

Sometimes this takes hours rather than weeks.

And the cost of the underlying technology has fallen dramatically.

Stanford's 2025 AI Index reported that the cost of querying a model performing at GPT-3.5's benchmark level fell by more than 280 times between November 2022 and October 2024.

That does not mean every AI application became 280 times cheaper. But it shows how quickly access to capable models became more affordable.

This is good news.

Small businesses can now experiment with technology that previously required substantial investment. Individuals can learn by building things rather than spending months studying theory before producing anything useful.

The problem is what happens when we confuse building a demonstration with understanding how to deliver a complete solution.

Someone can create a working AI agent without understanding how to manage failures, protect confidential data or control operating costs.

Someone can build a document chatbot without knowing how to enforce access permissions or test whether the answers are reliable.

The demonstration may look excellent.

But would you trust that person to deploy the same system across a company with hundreds of employees and sensitive customer information?

Maybe. Maybe not.

Knowing how to use a tool is not the same as knowing how to design a system around it.

That distinction is becoming harder for customers to see.

3. Companies are under pressure to buy AI expertise

Now consider what is happening on the customer side.

In Microsoft's 2024 Work Trend Index, 79% of surveyed leaders agreed that adopting AI was critical to remaining competitive.

Yet 60% said they were worried that their organisation lacked a clear vision and plan for implementing AI.

Think about those two numbers.

Most leaders believed AI was important, but many were unsure what to do about it.

That creates demand for external advice.

Hong Kong is experiencing its own AI rush

Hong Kong is not immune to this trend.

According to PwC's 2026 Global AI Jobs Barometer, AI-related job postings in Hong Kong increased by 50% in 2025. But AI-related positions still represented only 4.7% of total job advertisements.

Meanwhile, research published by HKUST in June 2026 estimated that Hong Kong's AI talent pool had grown to around 23,600 professionals in 2024, more than 280% higher than in 2015.

The interesting part is the mismatch.

HKUST found that 52.1% of new entrants to the AI workforce were relatively junior, while only 19% of advertised AI vacancies targeted junior candidates.

It also found that 85.3% of AI vacancies were for application roles rather than core AI research and development.

In other words, Hong Kong does not simply need more people who know about AI. Employers need people who can apply it to actual business problems, often with experience that goes beyond the technology itself.

The Hong Kong Productivity Council's 2025 survey of approximately 800 local companies adds another concern. While AI tools were widely used among surveyed businesses, 54% of companies with officially approved AI platforms said they lacked a complete or ongoing AI governance framework.

So we have a market where demand is growing, more people are entering the field, and many companies are still working out how to manage the technology properly.

That is a promising market for capable consultants. But it is also an environment where customers need to be careful about what kind of expertise they are buying.

A CEO might know that competitors are adopting AI but have little understanding of LLMs, data architecture, integration or AI governance.

They need someone to explain the options.

And when a market has many buyers who are uncertain about what they need, it creates opportunities for people who can confidently offer answers.

This is not unique to AI. It happens in other professional services too.

The difference is that AI is relatively new, changing quickly, and does not have a universally accepted qualification for someone calling themselves an AI consultant or architect.

A person might have twenty years of enterprise technology experience.

Another might have completed several online AI courses and built a few automations.

Both can describe themselves as AI consultants.

For a business owner without a technical background, it may be difficult to tell which person is qualified to solve a particular problem.

And that brings us to social media.

4. LinkedIn rewards visibility, not necessarily experience

Imagine two professionals.

The first spends most of the week designing systems, resolving integration problems, dealing with security requirements and supporting clients.

The second spends much of the week publishing AI content, sharing new tools, posting demonstrations and commenting on the latest model releases.

Who is more likely to become visible on LinkedIn?

Probably the second.

That does not mean the second person is less capable. They might be an excellent practitioner.

But social media visibility tells us very little about someone's ability to deliver a working enterprise system.

And AI has made content production easier too.

A person can now generate articles, diagrams, presentations, tutorials and videos much faster than before.

This creates an interesting situation.

The technology being promoted is also making it cheaper to promote oneself as an expert in that technology.

As more people enter the market, everyone competes for attention.

Simple messages travel well.

"Build five AI agents and replace your operations team."

"Automate your entire business in one weekend."

"Become an AI expert in 30 days."

These are attractive promises.

A detailed explanation of data permissions, system integration, failure handling and operational risk is considerably less exciting.

But those less exciting details are often what determine whether an AI implementation actually works.

There is a growing risk that businesses judge expertise by how confidently someone talks about AI rather than by what they have successfully delivered.

5. A working AI demo is not a working business system

Consider a simple example.

A company wants an AI assistant that can search its internal documents and answer employees' questions.

Consultant A builds a prototype.

They connect an LLM to a document repository, add a vector database and create a chat interface.

Within a few days, employees can ask questions and receive answers.

The CEO is impressed.

Consultant B approaches the same problem differently.

Before building anything, they ask several questions.

Who is allowed to access each document? What happens when permissions change? How will the system prevent confidential information from appearing in the wrong answers? How do we know the retrieved information is correct?

Then come other questions.

What happens when the model fails? How much will it cost when 500 employees start using it? Who will maintain it? How will it connect to existing systems?

And perhaps the most important question: does the company really need an LLM, or would better enterprise search solve the problem?

Both consultants might eventually produce something that looks similar.

But the work behind those solutions is very different.

The first has demonstrated that the technology can work.

The second is examining whether it can work reliably inside a real business.

This difference matters because companies are increasingly discovering that experimenting with AI is much easier than using it across an entire organisation.

McKinsey's 2025 State of AI survey found that 88% of respondents reported AI use in at least one business function.

However, nearly two-thirds said their organisations had not yet begun scaling AI across the enterprise.

That is an important gap.

AI adoption is widespread, but moving from experiments to established business operations remains difficult for many organisations.

And that is where experience in architecture, integration, security, operations and business processes becomes valuable.

Not because AI is particularly mysterious.

But because companies are complicated.

6. There is nothing wrong with being a new AI consultant

I think this point is important.

Not every AI project requires an experienced enterprise architect.

A small business that wants to automate appointment reminders, organize customer enquiries or summarize documents might be perfectly well served by someone who knows how to configure existing AI tools.

That person may deliver an excellent service at a reasonable price.

There is no reason to make simple problems unnecessarily complicated.

And someone who entered the AI industry two years ago might be more capable with modern tools than someone who has spent twenty years working with older technologies.

Experience alone does not guarantee competence.

The issue is not whether newcomers should be allowed to offer AI services.

Of course they should.

The issue is whether people are selling services that match their actual capabilities.

Building automations is one service.

Providing AI training is another.

Designing and governing an enterprise AI platform is something else entirely.

They require different skills, carry different risks and should not be presented as interchangeable.

A consultant does not need to know everything.

But they should understand the limits of what they know.

7. The economics may not be as attractive as they appear

There is another irony in this AI consulting boom.

The same tools that make it easy to start an AI consulting business also make it easy for competitors to do the same thing.

Suppose you start a business offering AI workflow automation.

You use commercially available LLMs, automation platforms and cloud services.

Your competitors can access almost exactly the same tools.

A potential customer might even build a simpler version themselves.

What makes your service different?

Perhaps you understand a particular industry. Perhaps you have strong relationships with customers. Perhaps you can integrate AI with complicated existing systems or provide reliable ongoing support.

Those are real advantages.

But simply knowing how to use popular AI tools is becoming a weaker competitive advantage as more people learn them.

This creates pressure on prices.

And while consultants compete to sell services built on top of AI platforms, the model providers, cloud companies and software vendors continue charging for the underlying infrastructure.

That does not mean AI consulting cannot be profitable.

It means the business model needs something more durable than access to tools everyone else can use.

In fact, I suspect this is a broader problem affecting many small professional service businesses.

AI reduces the cost of producing work, but it can also reduce what customers are willing to pay for that work.

Customers may assume that a report, presentation or software application should be cheaper because AI helped produce it.

At the same time, more providers can compete for the same projects.

Lower production costs do not automatically translate into higher profits.

Sometimes they simply lead to more competition.

8. What happens when the excitement settles?

I do not think AI consultants are going away.

AI will continue to create genuine opportunities for people who can help businesses use technology effectively.

But the market will probably become more demanding.

Today, knowing how to use a new AI tool can be enough to attract attention.

Tomorrow, that knowledge may be as ordinary as knowing how to use Excel, Google Workspace or a CRM system.

The tools themselves will continue improving.

Things that require specialist knowledge today may become standard features in business software.

When that happens, customers will have less reason to pay someone simply to configure them.

The more valuable work will involve understanding business problems, connecting systems, managing risk and demonstrating measurable results.

That does not necessarily mean only traditional technology professionals will succeed.

Some of today's newcomers will become excellent engineers, consultants and business advisers.

Others will find profitable niches delivering simpler services.

And some will discover that building a sustainable consulting business is considerably harder than building an impressive AI demonstration.

Ultimately, the market should become better at asking a few straightforward questions.

What have you actually delivered?

Does it work outside a demonstration?

What did it cost?

What business problem did it solve?

And who takes responsibility when it stops working?

Those questions matter more than any professional title.

Conclusion

I do not think the sudden growth of AI experts is simply a story about people chasing hype.

It is also a story about economics.

A difficult technology labour market encouraged people to develop new skills. Cheaper AI tools made it easier to enter the field. Companies facing pressure to adopt AI created demand for advice. And social media gave people a relatively inexpensive way to market themselves.

All these forces came together at roughly the same time.

The result is a market where it has become easier to demonstrate technical capability, easier to claim expertise and harder for some customers to distinguish between the two.

There is nothing wrong with making technology more accessible.

In fact, that may be one of AI's greatest benefits.

But access to powerful tools should not be confused with the experience needed to apply them properly.

AI has made it much cheaper to look like an expert. Whether it has made someone a better expert is a different question.

And that is something the market will eventually have to work out.

References

1. Challenger, Gray & Christmas (2024–2025). Annual Job Cut Reports. US technology-sector announced job cuts for 2023 and 2024. Challenger, Gray & Christmas
2. Microsoft & LinkedIn (2024). 2024 Work Trend Index Annual Report: AI at Work Is Here. Now Comes the Hard Part. Read report
3. Stanford Institute for Human-Centered Artificial Intelligence (2025). AI Index Report 2025. Research and Development chapter. Read report
4. McKinsey & Company (2025). The State of AI in 2025: Agents, Innovation, and Transformation. Read report
5. Stack Overflow (2025). 2025 Developer Survey. Findings on developer adoption and trust in AI tools. Read survey
6. Sources: PwC (2026), HKUST (2026) and HKPC (2025)