Worldwide Locations
Consultancies will tell you what to do. Platforms will give your teams something to work through. Between the two sits the part nobody has claimed: the people who actually have to build the thing. That middle is where strategy becomes capability, and it is where most transformation programs quietly stall.
Three markets, and the gap between them
If you look at how technology transformation gets bought, you will usually find three suppliers in the room.
The first is strategy. A consultancy arrives, studies the business, and produces a target state that is often genuinely good. It tells you what the organization should become.
The second is content. A platform arrives with a library covering everything, available to everyone, priced per seat. It tells your people what exists.
The third is the vendor itself, whose interest is in its own product working inside your estate.
What none of the three owns is the part in between: the named engineers, architects and analysts who have to take the target state and build it, on real infrastructure, under real load, with a real audit date. That is the unclaimed middle. It is also where the value in the whole program is either realized or lost.
Why the middle went unclaimed
It went unclaimed because it is hard to productize. You cannot ship it as a license, and you cannot bill it as a deck. It requires people who have built the thing before, enough of them, in enough places, authorized on the platforms the work actually runs on.
For a while it also looked unnecessary. The automation wave made it seem as though content could stand in for expertise, and for a stretch of years the market behaved as if that were true.
The numbers have not cooperated. In the middle of the largest automation wave most of us have worked through, spend on learning from a real expert is rising rather than falling. IDC's forecast for IT education and training services, cited in Darren Bance's POV this month, has the Americas market climbing from roughly 10.3 billion dollars in 2024 to 12.8 billion, with instruction delivered virtually by an expert gaining ground fastest in exactly the areas everyone is trying to staff: AI, cloud and security.
When the work gets harder and the consequences get sharper, organizations reach for people. They always have.
What the middle looks like when somebody occupies it
The clearest illustration this month came from a client of ours in a heavily regulated industry. Board-level AI push, strong engineering team, the full platform library bought and worked through, badges and all. Then the models had to go into production on infrastructure that genuinely mattered, and the program stalled.
Nothing was wrong with the strategy. Nothing was wrong with the content. Nothing was wrong with the engineers. What was missing was anybody who had done this specific thing in a business this exacting, sitting alongside them while they did it for the first time.
We paired those engineers with instructors who had built and shipped the same class of system under the same kind of scrutiny. A few weeks later the work that had stalled was moving and the board's program was back on track. Same people. Same ability. They had finally learned it from someone who had been there.
The four things the middle actually requires
If you are assessing who can hold that middle for you, these are the four properties that matter. They are also, not coincidentally, the four proof points we hold ourselves to.
| What the middle requires | What that means in practice |
| Real world experts | A global community of 900+ practitioners who have built, secured and operated systems like the ones your business depends on, authorized by the vendors those systems run on, including Microsoft, AWS, Cisco and Palo Alto. |
| Technical and business breadth | The technology deep enough to build it, and enough commercial fluency to explain to a board why it matters and what it costs to get wrong. |
| Programs built around the work | Not a catalogue mapped to a job title. A program shaped around the specific thing your organization has to deliver next, taught in real lab environments. |
| Outcomes that prove it worked | Evidence at the level of the work moving, not a completion percentage. Someone stays with it long enough to see whether it held. |
Four questions for your next partner conversation
Whoever you are talking to, these four questions separate the middle from the edges of it.
- Who specifically will be in the room with my team, and what have they personally shipped?
- Is this program built around your catalogue or around the work on my roadmap?
- What environment will my people practice in, and what is allowed to go wrong in it?
- How will we both know in ninety days whether this worked, in terms my board recognizes?
The position, plainly
We are not reinventing ourselves for the AI moment. For thirty years the belief has been the same: people get good at hard technical work by learning it from people who have already done it. That did not change when the cloud arrived, or when security became everyone's problem, or when AI started rewriting what the work is.
What changed is that the moment finally makes it obvious why it matters. If your AI and infrastructure plans depend on capability your teams do not have yet, another license will not close that distance. People will.
The work is still human. Build the people. Stand by the work.
Learn more about our cybersecurity solutions at: https://www.globalknowledge.com/us-en/cybersecurity/