AI Cosplay

How Performative AI Is Inflating a Dangerous Market Bubble

Most organisations are not using AI like they say they are, it's cosplay. There are announcements, trials, chatbots, automations but no tangible AI impact.

There is a gap between the story companies tell about AI and the reality of how work is actually done. That gap is being filled with smoke and mirrors, not capability.

We are living through the era of what I call performative AI, and it carries real risk not just to organisations but the global economy.

AI is talked about as if it is transforming the day to day reality of teams, yet the experience of most employees is largely unchanged.

This is not a failure of the technology but its a capability failure. Which, if unchecked, will inflate a bubble (The Big Short anyone?) that will eventually burst with more far reaching consequences than most can imagine.

AI Cosplay, Explained

AI cosplay is what happens when organisations pretend they are innovating rather than doing the hard work required to actually change. It is innovation by performance, not practice.

I see this pattern in many places:

  • Chatbots no-one uses
  • “AI analytics” that are mere dashboards that are hard to reframe.
  • Email or calendar tools that produce more confusion than clarity.
  • AI centres of excellence producing slide decks but not capability.
  • Declarations that an organisation is “AI first” while all the real work still happens manually.
  • Annual reports listing AI initiatives that have delivered no measurable productivity uplift.

It looks impressive from the outside, reads well in quarterly statements and gives the illusion of progress.

This is where the cosplay comes in, signalling competence instead of building it.

The deeper issue is simple and one I have seen many times, most organisations do not know how to redesign work, they know how to buy tools but not how to rework processes.

So they cosplay.

Why is this Happening?

Executives are (largely) not stupid or reckless but they are incentivised wrong, they are under internal/external pressure to do 'something' yet the capability gap remains huge.

1. Boards demand visible AI activity not outcomes

They want to hear we are “doing something with AI” without understanding what meaningful adoption looks like.

2. Vendors oversell and underdeliver

Everything is “AI powered” now. The label gets retrofitted to pre-existing or simple features so that organisations then believe they have adopted AI, when they are really just rebranding.

3. No one is redesigning how work actually happens

AI requires new workflows, new trust structures, and new decision models. Most organisations put this in the too hard pile, and AI becomes the cover image for a jigsaw with missing pieces.

4. Innovation teams are rewarded for activity, not impact

Launching a pilot in 3 days gets you recognition but transformation takes months of organisational negotiation. The former is rewarded, the latter is hard.

5. The capability gap is real

AI is failing organisationally, not technically. The models are extraordinary but most workplaces are not ready to use them.

Why AI Cosplay Is Dangerous: The Data Problem

A deeper operational truth is that most organisations could not adopt AI meaningfully even if they wanted to.

Why? Because their data is often a fragmented, vendor-controlled mess.

SaaS has left companies with data scattered across dozens of systems, governed by multiple external providers, and structured in incompatible ways. No organisation can train a truly useful model on data it does not control or cannot access easily.

The organisational data foundation is fundamentally unfit for the intelligence layer sitting over it.

AI Cosplay Is Inflating a Market Bubble

Here is the part no one is really talking about yet.

AI cosplay inside organisations creates an illusion of adoption, and that is feeding a global economic assumption that AI productivity gains are already happening, or will happen very soon.

Investors are (incorrectly) factoring in:

  • massive enterprise AI uptake,
  • significant operational cost reductions,
  • deep workflow automation,
  • exponential demand for AI infrastructure,
  • and major AI-driven revenue expansion.

But if the reality inside organisations is cosplay, then the fundamentals supporting the AI boom are overstated.

This is exactly how bubbles form: markets are driven by storytelling rather than outcomes.

You can map the risk clearly, this pattern has been seen time and time again, think .com boom, subprime loans, Black Monday etc...

  1. Organisations report AI activity, not results
    Boards hear that AI is “rolling out”. Investors assume adoption is real.
  2. AI vendors forecast exponential growth
    Demand appears explosive, even if it is mostly pilot projects and messaging.
  3. Infrastructure providers scale aggressively
    NVIDIA, chip manufacturers, data centre operators all expand for a future that may not materialise at the pace expected.
  4. Productivity numbers fail to lift in line with expectations
    AI is everywhere you look, but nowhere in outcomes.
  5. Confidence shock
    The market realises enterprise AI adoption is far slower and more superficial than the story suggested.

If that happens, the correction will not just hit AI vendors. It will hit index funds, sovereign wealth portfolios, cloud providers, and the entire AI-adjacent supply chain.

This is not a prediction of doom. It is an observation of misalignment between narrative and reality.

What Real AI Adoption Looks Like

Real AI adoption does not look like a chatbot. It does not look like a pilot. It does not look like an “innovation lab”.

Real AI adoption looks like a redesign of work itself.

  • Workflows change.
  • Roles change.
  • Cultural norms shift.
  • Decision making structures adapt.
  • Teams offload low-value digital noise.
  • Humans begin working on the human parts of work.

And at the centre of this shift (I believe) is something far more powerful than a generic AI tool.

The AI Twin

A model trained on your own data, that learns how you work and your context. A twin that mirrors some of your capability and sits beside you, not above you.

This is not cosplay. This is capability.

It moves organisations from tool-centric to human-centric, from vendor-led to self-sovereign, and from cosplay to meaningful transformation.

In short it is real, grounded, operational intelligence.

Closing Thoughts: Bubble or Breakthrough

The technology is not the problem, but cosplay is. It inflates expectations, distorts valuations, and puts a sticking plaster over the hard work required for genuine transformation.

If we keep pretending, this bubble is going to burst.

If organisations start rebuilding work, reclaiming their data, and designing capability around how people actually operate, AI will deliver the productivity lift the markets are already pricing in.

The gap between theatre and capability will decide who controls the next decade.

AI does not need more hype. It needs more honesty. It needs less cosplay and more design. It needs to be real.