AI Programme Management – Adapting Programme Leadership for AI Delivery

1. Insight

Artificial Intelligence is moving rapidly from experimentation into coordinated delivery across organisations. In many cases, that delivery is expected to be led through existing experienced programme delivery teams.

Experienced Senior Responsible Owners, Programme Managers and Business Change Managers are increasingly being asked to lead or support AI delivery initiatives.

The fundamentals of good programme management still apply. Existing portfolio and programme management approaches remain essential for prioritisation, governance and delivery. However, AI introduces characteristics that need to be understood and taken into account for successful delivery.

In many cases, this creates both a capability gap and an orientation gap.

Experienced delivery leaders understand programme delivery, but may not yet have a clear frame of reference for how AI behaves in practice and how this impacts delivery.

This Insight is intended to help experienced programme leaders understand how AI changes the programme management environment, what this means for delivery, and how to adapt with confidence.

Within the Orr Consulting AI Transformation Process, AI Programme Management sits within the Deliver stage, following completion of Discovery and Design.

2. Why This Matters

AI delivery often feels more complex and less predictable than traditional digital or business transformation delivery.

This is not because delivery discipline is weaker, but because the conditions in which AI delivery takes place are different.

AI programmes typically operate in conditions of:

• higher uncertainty

• stronger dependency, particularly on data

• greater iteration

• increased sensitivity to early decisions

• significant human and organisational impact

For experienced programme leaders, ambiguity and risk are not new. Established approaches such as Managing Successful Programmes (MSP) are designed to manage uncertainty and support progressive delivery.

However, in AI these conditions are amplified.

• outcomes may not be fully known upfront

• solutions may need to be proven rather than assumed

• dependencies may emerge during delivery

• plans may need to evolve as understanding improves

As a result, programmes can appear to be under control on the surface, while underlying risks build and can materialise as issues quickly.

A critical realisation is this:

Many AI delivery challenges are symptoms. The causes are often introduced upstream during the Discovery and Design stages of the AI Transformation Process.

3. What This Means for AI Programme Management

AI does not require a new programme management discipline.

It requires experienced programme leaders to apply existing discipline in a delivery environment characterised by greater uncertainty, dependency and sensitivity to early decisions.

There are four realities of AI delivery that programme leaders need to recognise, and four corresponding implications.

3.1 Reality 1 – AI Is Now Part of the Programme Landscape

AI delivery is no longer optional or experimental. It is becoming a standard part of organisational change portfolios and programmes.

What This Means:

Programme leaders need to develop sufficient understanding and orientation in AI to lead effectively.

This includes:

• awareness of AI capabilities and where they are best applied

• understanding of how AI initiatives fit within the broader AI Transformation Process

• recognition of how AI differs from traditional technology delivery

  
  

This is not about becoming a technical expert. It is about being able to lead confidently in a new delivery context.

The AI Universe provides a structured view of AI capabilities and where they are best applied. The AI Transformation Process provides the structure for how AI moves from Discovery to Design and then into Delivery.

3.2 Reality 2 – AI Delivery Is More Uncertain Than Most Leaders Are Used To

AI delivery operates beyond traditional waterfall and even beyond many agile implementations.

It is best understood as iterative delivery under conditions of high uncertainty, dependency and learning.

What This Means:

Programme leaders should recognise this as an extension of their existing role.

Programme management disciplines already exist to:

• manage uncertainty

• deal with ambiguity

• adapt to new information

• support progressive delivery

These now need to be applied more deliberately and rigorously.

This includes:

• accepting that not everything can be known upfront

• structuring delivery to allow for iteration and learning

• maintaining control while allowing plans to evolve

• avoiding false certainty in early plans and commitments

AI does not replace programme management discipline — it increases the need for it.

AI delivery also does not occur in a single uniform way. In practice, organisations typically encounter different delivery modes — including custom AI development, configuration of AI-enabled vendor solutions and generative AI adoption — each with distinct delivery and risk characteristics. These differences become particularly important at project delivery level and are explored further in the AI Project Management Insight.

3.3 Reality 3 – Much of the Risk Sits Upstream Before Delivery Begins

Many of the challenges that emerge during delivery are not created during delivery itself. They are introduced earlier through weaknesses in the Discovery and Design stages of the AI Transformation Process, including:

• AI Capability and Maturity – overestimation of organisational readiness

• AI Use Case Discovery – weak or misaligned use case selection

• AI Strategy and Roadmap – unclear prioritisation and sequencing

• AI Governance and Assurance –inadequate governance and assurance controls

• AI Business Case Development – untested assumptions and unrealistic expectations

  
  

If these stages are weak, delivery becomes an exercise in managing the consequences rather than delivering outcomes.

What This Means:

Programme leaders should not act purely as recipients of work entering delivery.

They should act as gatekeepers of delivery, ensuring that what enters delivery is:

• sufficiently well defined

• feasible

• aligned with strategy

• supported by realistic expectations

This may involve:

• influencing or contributing to Discovery and Design activities

• challenging assumptions before delivery begins

• ensuring appropriate governance and assurance is in place

• delaying or reshaping initiatives that are not ready for delivery

This reflects the role of the AI Transformation Process in strengthening how opportunities are defined and prepared before entering portfolio prioritisation and programme delivery.

Programme management manages uncertainty. Discovery and Design determine how much uncertainty there is to manage.

3.4 Reality 4 – Successful Delivery Does Not Guarantee Adoption or Realised Value

Even where AI solutions are successfully delivered from a technical perspective, this does not guarantee:

• adoption by users

• integration into business processes

• realisation of expected benefits

The human and organisational dimension is critical.

While this is a cross-cutting concern across all transformation, it is particularly significant for AI due to:

• changes in how work is performed

• trust in AI outputs

• perceived impact on roles and responsibilities

What This Means:

Programme leaders must place strong emphasis on:

• stakeholder engagement and communication

• business change management

• user adoption and trust

• alignment between delivery outputs and operational reality

• active benefits realisation management

This requires close collaboration with Business Change Managers and a clear focus on outcomes, not just outputs.

4. Risks of Not Adapting

Organisations and programme leaders who do not adapt to the characteristics of AI delivery often experience:

• initiatives entering delivery before they are properly defined

• over-optimistic expectations of benefits and timelines

• data and feasibility challenges emerging late

• fragmented or poorly prioritised use cases

• solutions delivered but not adopted or used effectively

These are rarely delivery failures in isolation. They are often the result of upstream weaknesses.

5. What Good Looks Like

Programme leaders who recognise and adapt to the realities of AI delivery can:

• improve the quality of what enters delivery

• reduce uncertainty and risk earlier in the lifecycle

• increase confidence in programme decisions

• enable more effective and controlled delivery

• improve adoption and the realisation of benefits

6. Final Thoughts

AI does not replace the fundamentals of programme management.

It changes the conditions in which those fundamentals are applied.

For experienced programme leaders, the key is not to reinvent programme management, but to apply it with greater awareness of uncertainty, stronger attention to upstream Discovery and Design, and a clear focus on adoption and realised value.

Programme management ensures delivery is controlled. Discovery and Design ensure that what is delivered is worth delivering.

This Insight is part of the Orr Consulting AI Insights Library — structured thinking for AI transformation leaders and decision makers.

7. Call to Action

If your organisation is preparing for or currently undertaking AI delivery and would like to strengthen how opportunities are defined, governed and delivered, we would be pleased to support you.

Discuss your next AI steps

Subscribe to Orr Consulting to receive occasional emails with practical AI Insights and updates.

Subscribe to Orr Consulting