AI Won’t Design Your Learning Strategy — But It Will Expose Whether You Have One
LEADERSHIP FROM THE INSIDE — PHILIP KNOX | PGKCONSULTANCY.COM
AI Won’t Design Your Learning Strategy — But It Will Expose Whether You Have One
A practitioner’s perspective on AI and L&D from someone who has been in the room
The noise vs. the reality
There is currently no shortage of commentary about how artificial intelligence is going to transform learning and development. Most of it is written by people who have never conducted a training needs analysis, sat in front of a sceptical finance director trying to justify a development budget, or tried to explain to a Chief Constable why last year’s leadership programme hasn’t yet changed the culture on the ground.
I have done all of those things. So this is written from the inside — from the perspective of someone who has designed and delivered learning across government, public safety, aviation, and energy sectors, and who currently teaches at CIPD Level 7. Not as an AI commentator. As a practitioner.
And the honest assessment is this: AI is a significant development for L&D. But the organisations that will benefit from it are those that already have clarity about what they’re trying to do. For everyone else, it will mostly help them produce irrelevant content faster.
The fundamentals haven’t changed
The shift from training to learning — from sage on the stage to guide on the side, from centralised delivery to learner-led approaches — has been underway for decades. None of that is altered by AI. The Systematic Training Cycle still applies. ADDIE still applies. The questions are still the same ones they have always been: what problem are we solving, for whom, how will we know if the learning worked, and is this actually a learning problem at all?
That last question is the one that separates the practitioner from the order-taker. Xerox moved away from training needs analysis towards a performance consulting model for a reason: too often, training is reached for first, and the TNA becomes a mechanism to design a delivery that was already decided. The right starting point is the performance gap, not the course catalogue.
AI does not change that logic. It adds capability within it.
Where AI genuinely helps: analysis and diagnosis
The first question an OD practitioner asks is not ‘what should we train people on?’ It is ‘is this a training need, or is it something else?’
This is where AI begins to make a practical difference — not because it answers that question, but because it helps surface the data you need to answer it yourself.
AI can analyse patterns across large datasets — exit interview themes, engagement survey trends, performance metrics, incident data, complaint patterns, 360-degree feedback at scale — and surface connections that manual analysis would miss or take months to find. It can flag where performance gaps cluster around particular roles, levels, or contexts. It can help distinguish between a skills gap, a knowledge gap, a motivation issue, or a structural problem that no amount of training will fix.
The MAGIC model (Mandatory, Adapting, Growing, Improving, Corrective) remains a sound starting framework. But applying it to organisational data at any meaningful scale used to require significant time and resource. AI changes that calculation.
What it does not change is the OD judgement about what the data means. Understanding whether a pattern in performance data reflects poor individual capability, ineffective line management, a toxic team culture, or an organisational design problem that sits upstream of anything L&D can address — that still requires human expertise, contextual understanding, and professional courage. AI is a diagnostic aid. It is not a diagnostician.
The leadership question
Mike Walsh, in The Algorithmic Leader, argues that the future leader is not the smartest person in the room, but the one asking the smartest questions. That framing is useful for L&D practitioners thinking about AI, but it is also directly relevant to what we should be developing in leaders themselves.
AI is reshaping the environment in which leadership operates. Algorithmic literacy — understanding how AI systems work, where they apply, and crucially where they fail — is becoming a core leadership competency. So is the capacity for systems thinking, ethical judgement under data-driven pressure, and what Walsh calls ‘narrative intelligence’: the ability to make sense of complexity and communicate it with clarity.
None of this replaces the fundamentals of transformational leadership. Clear vision, role modelling values, genuine care for individuals, and the willingness to stretch and challenge people — these do not become less important when algorithms handle the routine. If anything, they become more important. When process is automated, what is left is the human quality of leadership: culture, values, relationships, and the tone that leaders set simply by how they behave when things are difficult.
The challenge for leadership development programmes — and this is one I engage with directly in CIPD Level 7 teaching — is that linking LMD interventions to measurable organisational outcomes has always been hard. The CIPD’s own data shows that investment in leadership development dropped by 7% between 2021 and 2023, and that less than half of L&D teams are currently delivering it. That is a profession-wide problem. And it will not be solved by AI alone.
Where AI genuinely helps: evaluation and learning transfer
Evaluation has always been the weakest link in L&D, and nowhere is that more painfully true than in leadership development. The CIPD’s research is direct on this: only 13% of organisations evaluate beyond learner satisfaction. Only 24% of L&D professionals know how long it takes to achieve competence in a role they design development for.
That is not a data problem. It is a thinking problem. Organisations that treat evaluation as an afterthought — bolted on after delivery rather than designed in from the beginning — will not be rescued by AI analytics. The question is not how to measure what happened. It is what you were trying to change, how you would know if it changed, and who in the organisation needs to see the evidence. That thinking has to come first.
The deeper challenge is learning transfer. The CIPD’s 2023 Leadership Development Scientific Summary is clear: leadership programmes frequently fail to embed learning into the workplace. Lack of follow-up, unsupportive line management cultures, low accountability after the programme ends, and the time lag before behavioural change becomes visible — these are consistent barriers. Kirkpatrick’s four-level model remains the most referenced framework, but moving beyond Level 1 (the happy sheet) to Level 3 (behaviour change) and Level 4 (organisational results) requires longitudinal tracking that most organisations simply do not do.
This is where AI makes a genuine practical difference. It can analyse engagement survey trends over time, correlate them with cohorts who completed development programmes, surface patterns in 360-degree feedback data at scale, and flag where team performance metrics shifted in the periods following leadership interventions. The Theory of Change approach — mapping from development inputs through intermediate indicators to strategic outcomes — becomes significantly more evidenced with AI-enabled analytics behind it.
AI can surface the data. It cannot tell you what questions to ask before you start.
The practitioner’s honest assessment is this: AI expands what is possible in evaluation. But it does not replace the professional judgement about what matters, what to measure, and what good looks like. Those are human decisions. And they are exactly the decisions that CIPD Level 7 is designed to develop.
The harder truth
If your L&D strategy is not aligned to business performance, AI will help you produce irrelevant content more efficiently. If your organisation treats training as a compliance exercise rather than a development opportunity, AI will not change that. If your leaders do not model the behaviours that your learning interventions are trying to develop, no amount of technological sophistication in the design will make the learning transfer.
Culture is still the determining variable. And as Edgar Schein long established, culture follows leadership — not policy, not structure, and not technology. The organisations that will genuinely benefit from AI in their learning and development are those where senior leaders believe in development, where line managers create the conditions for learning to transfer, and where L&D has a seat at the table when organisational performance is discussed.
AI is a powerful tool in the hands of a capable practitioner operating in a learning culture. In any other context, it is faster noise.
The practitioner’s challenge
Before worrying about AI, make sure you can answer the questions that have always mattered.
Is this actually a training need, or is it something else? What does success look like, and how will you measure it beyond the end-of-programme questionnaire? Do your leaders model the behaviours the learning is designed to develop? Is your organisation genuinely committed to development, or to compliance?
If you can answer those questions clearly, AI becomes a genuinely useful tool. If you cannot, it is a distraction from the work that actually needs doing.
Philip Knox | FCIPD | Leadership Consultant & Educator | pgkconsultancy.com
Former Head of Learning and Development, PSNI. Delivering leadership and organisational development across government, public safety, energy, and aviation internationally.