Leadership teams invest in the strategy. They rarely invest in the belief system needed to execute it. That's the gap.
There's a number that should be practically tattooed on every enterprise strategy deck: 97%.
That's the percentage of AI projects launched at scale that never reach full production, according to MIT Sloan's 2025 research. Across every sector, every region, every level of investment. Nearly all of them stall.
Now here's the part most organisations don't want to sit with: the technology isn't the problem.
BCG surveyed 1,000 C-suite executives across 59 countries and 10 major industries to figure out what separates AI projects that actually make money from the ones that become expensive corporate case studies in what not to do. The answer was uncomfortable for anyone who's been allocating budget like it's 2019.
Companies that turn pilots into profit follow what BCG calls the 10-20-70 rule: 10% of resources on algorithms, 20% on technology and data, and 70% on people and processes.
Seventy percent. On people.
I had a chat recently with former Westpac CTO David Walker, who ran his own meta-analysis of the credible research in this space. He got the same answer as BCG.
Which raises an obvious question: if the data is this clear, why are leadership teams still putting 70% of their attention on the technology?
I've been fortunate enough to MC quite a few enterprise tech events in the past 12 months, and more than a handful of C-suite executive roundtables. Anecdotally, my experience backs the data. We'll spend all this time talking about the technology and wrap up with a key takeaway: what should leaders do next on their AI journey?
The answer is always the same: invest in your people.
The gap nobody names
This scenario reflects a diagnostic tool I've taught for years now called the Vision Reality Gap. It shows up in every major change program I've observed, and it looks like this: leadership is genuinely excited, the strategy is solid, the roadmap is ambitious. Meanwhile, the people responsible for execution are quietly running a completely different story in the background: this won't work.
The AI version of this gap has a particularly sharp edge. Research shows 76% of leaders – three in four – are enthusiastic about AI. What they often assume is that their teams share that enthusiasm. They don't. Around 31% of staff across all age ranges are not on board. And it goes further than passive resistance.
A 2026 Writer/Workplace Intelligence survey of 2,400 knowledge workers found that 29% of employees had sabotaged their organisation's AI strategy in some form. Among Gen Z workers, that figure climbs to 44%. Roughly a third of those who admitted to sabotage said they did it because they were afraid of losing their jobs.
This is informed resistance. People aren't pushing back because they don't understand the technology. They're pushing back because nobody has given them a credible story about where they fit in the future the organisation is building.
That's not a technology problem. That's a narrative alignment problem.
What JPMorgan got right
In mid-2024, JPMorgan Chase launched LLM Suite – an internal generative AI platform, model-agnostic, connected to the bank's data and workflows. By March 2026, it was being used by more than 230,000 employees globally – about 75% of the workforce, with half of them using it every day. The platform hit 200,000 users in eight months.
Through opt-in, not a mandate.
The story isn't the US$19.8 billion budget. Plenty of organisations have spent at that scale and got nothing close to that result. The story is that JPMorgan closed the narrative gap before they deployed the tools.
Three moves made it stick.
First, Chief Analytics Officer Derek Waldron made AI the central narrative at executive retreats – not a side project, not a pilot to be quietly ignored, but the stated direction of the enterprise.
Second, Consumer Banking Chief Marianne Lake told investors directly that operations staffing would reduce by around 10% over five years – and CEO Jamie Dimon went on record: "We have displaced people with AI, and we offer them other jobs." Headcount held steady at 318,000 through redeployment. The honest story pre-empted the fear story.
Third, adoption was opt-in, with visible proof built in. When bankers could build a five-page deck in 30 seconds, the evidence did the work. Operations teams handled 6% more accounts. Fraud costs dropped 11%.
You don't need a $19.8 billion budget to replicate that logic. You need to close the gap before you deploy the tools.
What's actually in the gap
This is where most change frameworks leave money on the table. They diagnose the gap. They rarely address what lives inside it.
What lives inside it is belief.
A belief is an assumed truth – one that sits deep enough in a person's identity that it doesn't respond to a PowerPoint presentation or a change management comms plan. In my work across leadership teams, the beliefs that kill change programs aren't usually the obvious ones. They're not "AI is bad." They're more like: my value here is tied to what I know, and if the system can do what I know, I don't know who I am anymore.
That's an identity story. And identity stories don't move because someone in senior leadership had a good quarter and sent an all-staff email about the exciting road ahead.
The cryptology of belief systems – the hidden codes that govern how our beliefs are formed and protected against change – is genuinely complex.
Dr Carol Dweck's work on fixed versus growth mindset points to the mechanism: every person maintains a running internal account of what's happening, what it means, and whether to resist or move toward something new. Until that inner account shifts, no external strategy lands at the pace you need it to.
This is what I mean when I say: leadership teams invest in the strategy. They rarely invest in the belief system needed to execute it.
Closing the gap
The practical path forward has three moves.
Read first. Diagnose the real narrative – not the one in the board paper, but the one being told in Slack DMs and hallway conversations and the story behind the story of that resignation you didn't see coming. The gap between the leadership narrative and the ground-level reality is data. Most organisations are sitting on it and not reading it.
Write horizontally. The research is consistent: the biggest predictor of AI adoption isn't budget, it's whether the people doing the work helped write the strategy. Top-down CEO endorsement is necessary but no longer sufficient. People need to be co-authors of the story, not recipients of it. That requires a different kind of engagement – practical, hands-on, across every layer of the organisation.
Share value your customers can actually articulate. The final test of any change program isn't internal. It's whether the people you serve can tell you, in plain language, how it helped them. That requires a whole-of-company focus on outcomes that are human and legible – not just metrics that make sense inside the building.
Ultimately, the person in front of you is your AI strategy. That's not a soft idea dressed up in business language. It's the most precise description of where enterprise transformation succeeds or stalls. Profitable change is people-centred, driven by customer value, and grounded in a trusted story about how the future includes the people being asked to build it.
Close the narrative gap first and everything else follows.
Mark Jones CSP is the founder of The Story Code Co. and the author of The Story Code and Beliefonomics. He works with senior leadership teams to close the gap between strategy and execution through narrative alignment.
