The latest TNTN Live set out to explore AI skills, but what emerged was less a discussion about technology and more an observation about how work itself begins to change once those capabilities are taken seriously. It became clear very quickly that the problem organisations are facing is not a lack of access to AI, nor even a lack of understanding of what the tools can do. The problem is that most are still approaching AI as an accessory to existing workflows, rather than something that fundamentally reshapes them.
There is now a well-established pattern playing out across industries. Organisations invested early, rolled out licences, encouraged teams to experiment, and expected to see meaningful gains. What followed was a burst of activity, a spike in generated content, and then a quiet plateau. The promised transformation failed to materialise. Instead, AI became something that sat on the edges of work, useful in moments, but not central to how outcomes were delivered. This has led to a growing sense of scepticism, not because the technology is weak, but because it has been applied in ways that preserve the very structures it is capable of replacing.
The discussion moved beyond this familiar narrative by reframing the role of AI entirely. Rather than seeing it as a general purpose assistant, it is far more useful to think of it as a collection of discrete, repeatable capabilities. In other words, as skills. This distinction matters because it changes how the technology is used. A prompt is a transaction. It is a request followed by a response. A skill, by contrast, is something that can be relied upon, invoked repeatedly, and embedded into a broader system of work. Once that mental shift occurs, the impact becomes far more tangible.
When AI is treated as a set of skills rather than a tool, the structure of work begins to compress. Tasks that previously required multiple steps, often distributed across different roles, start to collapse into a single flow. Activities such as gathering information, structuring it, producing an output and refining it are no longer discrete phases. They become part of a continuous interaction. The gain here is not simply speed, although speed is an obvious byproduct. The more important change is the removal of friction between steps. Work becomes more direct, less dependent on coordination, and less constrained by the need to pass information between people and systems.
This compression has a second order effect that is arguably more important. Capability becomes decoupled from individuals. In traditional organisational models, capability is tied to roles. If a particular type of work needs to be done, it is assigned to the person or team that owns that expertise. This creates dependency, delay and, often, bottlenecks. Once AI skills are embedded into the flow of work, those capabilities become available on demand. They are not perfect, and they do not eliminate the need for human oversight, but they are sufficiently reliable to change behaviour. The question shifts from who can do this to what is required to move this forward. That shift is subtle, but it has profound implications for how teams operate.
The conversation made it clear that the real gains do not come from any single capability in isolation. They come from the combination of capabilities. Writing, analysing, structuring, translating and evaluating are all useful individually, but their true value emerges when they are orchestrated together. This is where most organisations are currently falling short. They are using AI in isolated pockets, applying it to individual tasks without rethinking how those tasks connect. The result is incremental improvement rather than systemic change. When those capabilities are combined into a coherent system, the effect is very different. Entire segments of workflow can be replaced, not by removing work, but by changing how it is executed.
This inevitably shifts the role of the individual. As execution becomes easier, the value of judgement increases. The work does not disappear, but it moves up the stack. Less time is spent performing tasks step by step, and more time is spent defining outcomes, shaping inputs and evaluating results. This is not a reduction in responsibility. If anything, it is an increase. The individual becomes accountable for direction rather than execution. This is one of the reasons the current moment feels uncomfortable. AI promises efficiency, yet many people feel busier than ever. The tools accelerate output, but they also raise expectations. Once it becomes possible to do more, more is expected. The burden shifts from producing work to ensuring that the work being produced is the right work.
There is also a clear lesson in why many early implementations have failed to deliver value. When AI is used as a shortcut, the output is predictably shallow. Content is generated quickly, but without depth or originality. Decisions are accelerated, but without sufficient context. This is what has led to the growing frustration with what is often described as low quality, generic output. The issue is not the capability itself, but the absence of structure around it. Skills, when properly defined and embedded, operate within a system. They contribute to a larger outcome. Without that system, they produce noise rather than value.
What emerged from the discussion is that the organisations seeing meaningful gains are those that have moved beyond experimentation and started to rethink the architecture of work itself. They are identifying the capabilities that matter, embedding them into their processes, and orchestrating how those capabilities interact. They are not asking how to use AI in a general sense. They are asking what their organisation can now do that it could not do before, and how that should change the way work is designed.
This leads to a more fundamental realisation. As capability becomes increasingly abundant, it ceases to be the primary constraint. The constraint shifts to intent. It becomes less important whether something can be done, and more important whether it should be done, how it should be done, and what outcome is actually being pursued. This places a premium on clarity, judgement and leadership. The organisations that succeed in this environment will not be those with the most advanced tools, but those that are most effective at directing and orchestrating the capabilities available to them.
Mad Skillz was not a discussion about features or product updates. It was a reflection on what happens when capability becomes embedded within the system of work rather than layered on top of it. The lesson is not that AI will make people marginally more productive. It is that it has the potential to change the structure of work entirely. That change does not happen automatically. It requires a deliberate shift in how organisations think about capability, process and value.
The work that is emerging from this shift is different in nature. It is less about execution and more about direction. Less about producing output and more about shaping outcomes. The tools are already capable. The question is whether organisations are willing to redesign themselves around what those tools make possible. That is where the real gains are realised, and that is the challenge that now sits in front of every leader paying attention.