The Missing Layer in Leadership Development: Why Individual Coaching Alone Won’t Move Your KPIs
If you have invested in strong individual leadership development and the organisational KPIs you were promised still have not moved, you are not imagining it, and the programme was not necessarily wrong. Four separate bodies of evidence now converge on the same explanation, and the newest of them has just been tested with hundreds of billions of pounds of real enterprise spending rather than a handful of academic studies.
The Evidence: Capability Is Necessary, Not Sufficient
A 2025 systematic review of 116 studies on leadership agility in VUCA environments (Syamsir et al.) found that individual capabilities - emotional intelligence, cognitive agility, behavioural adaptability - only translate into agile leadership behaviour when three organisational conditions are also present: psychological safety, timely information, and genuine decision latitude. Train the individual and leave the system untouched, and the training has a ceiling built into it before it starts.
A 2026 meta-analysis of 5,651 leaders across 14 studies (Ferede & Getahun) reinforces this from the change management angle. Strategic leadership does have a statistically significant positive effect on change success. The more revealing finding is how much that effect varied from one organisation to the next. Picture 14 companies trialling the same change leadership approach: if the results had all landed in a similar place, give or take normal variation, you could call the approach reliably effective. Instead, some organisations saw a strong positive shift, others saw almost nothing, and in a few cases the same approach appeared to make things worse. That spread is too wide to explain away as chance. It means the organisation you are operating in shapes the outcome more than the leadership model itself does, which is exactly why the paper’s conclusion is blunt: a one-size-fits-all leadership model is inadequate.
Separately, research into why senior leaders actually seek coaching - role transitions, team disengagement, the absence of honest feedback, strategic inflection points - shows that the moments capable leaders most need support are rarely capability gaps. They are context gaps: situations where what worked before does not automatically work now.
“Individual leadership development answers the question "is this leader capable?" It rarely answers the question your board is actually asking: "is this organisation capable of adapting fast enough?"”
Why This Matters for Business Impact and KPIs
This has direct implications for the L&D conversation about demonstrating business impact. If a leadership programme is measured purely on individual capability gains - confidence scores, 360 feedback improvements, self-reported skill growth - it can succeed on every metric you set and still fail to move the organisational KPIs it was commissioned to influence, because those KPIs depend on the organisational layer the programme never touched.
The research points towards a different measurement model: track psychological safety and decision latitude alongside individual capability. Track how quickly information reaches the leaders who need to act on it. Track behavioural and cultural change - willingness to embrace change, alignment with new goals - not just financial outcomes, because the change research specifically flags financial-only measurement as insufficient.
The Fourth Strand: What Gallup’s 2026 Data Just Proved at Scale
The most current, and by far the most expensive, confirmation of this thesis has nothing to do with an academic journal. Gallup’s newly published State of the Global Workplace 2026 report shows that 95% of organisations investing in artificial intelligence report zero measurable profit impact, and separate NBER research puts the proportion of executives seeing no productivity effect at all at 89%. Given the scale of global AI spend, this is the most expensive live demonstration yet of the missing layer this article is named for.
The report identifies the strongest predictor of whether an employee feels AI has genuinely changed how they work: whether their manager actively champions its use. Where managers do, employees are 8.7 times more likely to say AI has transformed their work, and 7.4 times more likely to say it lets them do what they do best. Fewer than a third of US employees in AI-adopting organisations say their manager actively supports the team’s use of it. In Germany, that figure is 21%.
This lands directly on the layer the first three research strands already identified. A capable individual leader, dropped into a team where engagement, psychological safety and genuine support have not been built, will not become an effective AI champion by attending a tools workshop. The barrier is organisational, not technical, and it is the same barrier that stalls change programmes and leadership agility alike.
The scale of the problem is structural too. Manager engagement globally has dropped 9 points since 2022, with the sharpest single-year fall - from 27% to 22% - landing between 2024 and 2025. Managers used to hold an "engagement premium" over the people they lead. That premium has gone. In best-practice organisations, 79% of managers are still engaged, which confirms this is fixable rather than inevitable.
“You cannot delegate resilience through a software licence. Managers are the layer where every technology investment either compounds or evaporates.”
There is a further nuance worth using directly with a sceptical board. Leaders as a group report higher life satisfaction than individual contributors but a worse day-to-day emotional profile: more stress, anger, sadness and loneliness. Engaged leaders do not show this pattern at all. They report lower negative emotion than the average individual contributor and are 14 points more likely to be thriving than the average leader. What looks like a leadership tax in the data is actually an engagement tax, and engagement is buildable, coachable and measurable.
For UK and European L&D functions specifically, the regional data adds a useful correction. Europe returns the lowest engagement of any world region, yet wellbeing sits well above the global average. This is not a burnout story. It is a quiet disengagement story, and it changes what kind of intervention will actually land with your leaders.
Leaders aren’t necessarily burnt out, instead they are disengaged. Photo Getty Images Unsplash
Multi-Level Intervention: The Practical Answer
The consistent thread across all four strands of evidence is that effective leadership development has to work at more than one level simultaneously. This is not a new idea in principle, but the evidence base for it has become considerably stronger and considerably more current, and it gives L&D leaders a more precise language for making the case internally.
Individual level: build cognitive agility, emotional intelligence and behavioural adaptability through sustained, reflective practice - coaching rather than one-off training
Team level: build genuine psychological safety and the trust that allows honest challenge, disagreement and experimentation
Organisational level: audit and address decision latitude, information flow and how mistakes are actually handled when they happen - not how the policy says they should be handled
Technology and AI rollouts: baseline manager engagement and active-championing behaviour before launch, not after adoption has already stalled
Diagnostic first: given how much outcomes vary by context, assess before you design, because the same programme genuinely will not produce the same result in every part of the business
Executive Presence and the Seat at the Table
There is also a direct connection here to the executive presence conversation many L&D leaders are having right now. Influencing at the senior table requires being able to make precisely this argument - that individual leadership capability and organisational readiness are two different investments, and that KPI movement depends on both. That is a business-impact argument, not a training argument, and it is the kind of framing that earns L&D a genuine seat in strategic conversations rather than a seat at the delivery table alone.
What This Means for Your Next Leadership Investment
Before commissioning the next leadership investment, four questions are worth asking that this evidence base now gives you the confidence to ask directly: What organisational conditions exist for the capability we are about to build to actually be used? What does our data tell us about where in the business the same intervention is likely to land differently? What are we measuring beyond individual capability that will tell us whether this investment moved the organisation, not just the individual? If the investment involves new technology, have we checked whether our managers are engaged enough to champion it, rather than assuming the training will take care of adoption on its own?
The multi-level advantage
Individual coaching, team facilitation and organisational diagnostic work are not three separate services. Used together, they address the full picture this evidence describes, which is precisely why leadership development that only operates at one level so often plateaus below the results it was capable of delivering.
For L&D professionals
If the AI enablement budget conversation in your organisation is happening in Technology or IT alone, it is happening in the wrong room. All four strands of evidence in this article point the same way: route leadership and technology investment through the organisational layer - psychological safety, decision latitude, manager engagement - not around it.
Ready to have a conversation?
This article is also published as a blog at theresiliencecoach.co.uk. For a structured, multi-level view of where your organisation currently stands, the Resilient Organisations Diagnostic and Resilient Leadership Diagnostic are both available below.
Leadership Resilience Diagnostic:
theresiliencecoach.co.uk/resilient-leadership-diagnostic
Book an introductory call: