Ask two people in the same company how digitally mature it is, and you will get two different answers. Ask the executive team and you will hear about the platforms they bought. Ask the people who use those platforms every day and you will hear about the spreadsheet that still runs the process.
Both answers are honest. Only one of them is true.
This matters more now than it did two years ago, because AI has turned digital maturity from a background concern into a board-level mandate. Companies are being pushed to “do AI” before anyone has confirmed they can. And the uncomfortable finding across the research is that most companies rate their own readiness higher than reality supports – often by a full stage.
Below: the evidence that this is systematic rather than personal, why it happens, what it costs, seven signs that it is happening to you, and a five-minute way to find out where you actually stand.
The Gap Between What You Own and What Actually Works
Here is the mechanism behind that number, and it is simpler than it looks.
Digital maturity is visible in exactly the places where it is easiest to fake. Licences, budgets, dashboards, a new platform rollout, a data team on the org chart. All of those are countable, and all of them get counted.
But maturity actually lives in the places that are hard to see: whether the systems genuinely talk to each other, whether the data is trustworthy, whether the people doing the work actually use what was bought, and whether any of it changes a business outcome.
Executives see the board deck. Employees live the friction. When those two views never meet, the gap between them is where transformation budgets go to die.
This is not a story about arrogant leadership. The research is consistent that the distortion is systemic, not personal. It shows up across industries, company sizes and continents. The same way a person cannot judge their own driving skill, an organisation cannot accurately judge its own maturity from the inside – because it is measuring itself against its own assumptions. The fix is not humility. It is external calibration.
The Number Nobody Says Out Loud
If overestimating were harmless, this would be an interesting psychology note. It is not harmless, because it changes what companies buy next.
The clearest evidence of the pressure comes from a 2026 survey by KUNGFU.AI and Wakefield Research, which found that 74% of C-suite leaders admit they have projected more confidence in their AI strategy to stakeholders than they actually felt. Nearly three quarters of executives. Not a few.
That is the environment in which investment decisions are being made. Confidence is being broadcast at a level the underlying readiness does not support, and everyone in the room can feel it.
It rarely ends well. And the reason is almost never the technology.
What Overestimation Actually Costs
The failure numbers in enterprise AI are remarkably consistent, and they all point at the same cause.
80%+
of AI projects fail – roughly twice the failure rate of non-AI IT projects.
RAND Corporation, 2024
95%
of organisations report zero measurable return on their generative AI investment.
MIT Project NANDA, GenAI Divide
86.9%
of companies delayed their AI deployments by around six months – driven by data gaps, not model quality.
AvePoint, 2026
Read those three together and a pattern appears. The projects are not failing because the models are not good enough. They are failing because the organisation was not ready for the model it bought.
When MIT researchers looked at why the returns were missing, the explanation was integration, workflow alignment and the readiness of the underlying data – not the AI itself. When AvePoint asked why deployments slipped, the answer was data governance and data ownership. Neither of those is a technology problem. Both of them are maturity problems.
Buying a Stage 4 solution on Stage 2 foundations is the single most reliable way to turn a budget into a pilot graveyard.
Why This Keeps Happening to Good Companies
This pattern has a name – and it is a shared human tendency, not a personal failing.
A 2026 study in the European Management Review examined how the Dunning-Kruger effect – the tendency to overestimate competence in areas where competence is low – plays out specifically in managerial self-assessment. The finding is not that managers are uniquely flawed. It is that self-assessment in complex domains is structurally unreliable, for everyone.
Digital maturity is a complex domain. It requires knowledge in data engineering, systems integration, organisational behaviour and process design all at once. Almost nobody holds the whole picture, which means almost everybody fills the gaps with assumption.
Here is what that looks like in practice. WalkMe’s fifth annual State of Digital Adoption report – a survey of 3,750 employees and executives across 14 countries – found that 54% of workers bypassed their company’s AI tools at least once in the previous 30 days and completed the work manually instead. A further 33% had not used AI at all. The licences were bought. The rollout happened. The adoption did not.
From the top, that looks like a mature organisation using AI. From the floor, it looks like a subscription nobody opens.
Seven Signs You May Be One Stage Ahead of Yourself
Diagnosis beats assumption. If three or more of these sound familiar, your perceived stage and your real stage are not the same number.
None of these signs mean you are behind. They mean you are ahead of your foundations, which is a very different problem – and a far more fixable one.
The Fix Isn’t Less Ambition. It’s Calibration.
The wrong conclusion from all of this is “we should slow down.” Ambition is not the problem. Uncalibrated ambition is.
The right move is the boring one: establish where you actually are, understand what the next stage requires, and invest in that instead of in the stage you wish you were in. Most organisations move through four stages, from manual and fragmented work, through integration, to optimisation. You can read the detail in the four digital maturity stages – but the short version is that each stage is a precondition for the next, and skipping one does not accelerate anything.
The uncomfortable part is that calibration cannot be done from the inside. Self-assessment in a domain this complex is reliably distorted, which is exactly what the research keeps showing.
That is why we built something quick and unglamorous. The 5-minute Digital Maturity Assessment measures six dimensions, returns your stage immediately, and gives you a personalised action plan rather than a generic score. It is based on European Commission recommendations adapted to the reality of growing businesses, and it measures the capability you actually have rather than the one you have purchased.
If you want the methodology behind those six dimensions before you take it, that is covered in how to measure digital maturity.
What to Do at Each Stage
The honest answer depends on where you land. There is no single next step, which is precisely why generic advice about AI is not useful.
Stage 1 or 2: foundations before anything else
Fix data quality, integration and the manual handoffs first. Every euro spent on AI at this stage buys a demo, not a result. Start with honest measurement, then close the weakest dimension.
Stage 3: automate and integrate before you add intelligence
You have the systems. The gap is the connections between them and the workflows on top. This is where automation pays for itself, and where your data becomes trustworthy enough to feed a model.
Stage 4: AI is genuinely available to you
At this point AI is not a gamble, it is an extension of capability you already have. This is the territory covered by intelligent transformation, and it is where the returns the rest of the market keeps missing actually live.
If the failing projects all share one trait, it is that they skipped this order. If the successful ones share one trait, it is that somebody insisted on knowing the real starting point first.
Take the 5-minute Digital Maturity Assessment and compare your result with what you expected. What you learn is your real stage across six dimensions, the specific gap holding you back, and a prioritised next step – not a transformation plan you will never start.
If you already sense the gap and want to talk through it, you can book a meeting with our team.
The Question Worth Asking Today
You are probably not behind on AI. You are more likely ahead of your foundations – and that is genuinely good news, because foundations can be fixed in months, while a failed three-year transformation programme cannot.
The only thing that makes it worse is not knowing. Confidence without measurement is the most expensive position a company can hold right now, because it feels exactly like readiness right up until it isn’t.
Measure. Then decide. In that order.
