With increasing pressure to build AI literacy and capability across the organisation, many programme sponsors push for implementation before having mapped competencies and capabilities against a target state. While speed and agility in digital transformations are important, neglecting crucial steps in the process can result in employee resistance, operational disruptions and technical debt. The trick is to design the programme so that it is able to deliver quick wins, but ultimately realises tangible business value in the medium to long term. Ensuring that each design, adoption and application cycle is guided by a Responsible AI framework is crucial to prevent governance breaches.
This approach comprises several components which can be implemented within an agile framework.
- Responsible AI framework
Building AI literacy and capability should include understanding the principles of Responsible AI and how it applies in every phase of design, development, and deployment. Integrating frameworks such as the Responsible AI Wheel (below) into the programme ensures that innovation and integrity evolve hand in hand.
- Stakeholder input and buy-in
Getting stakeholder alignment is one of the most critical components of a programme’s success. The objective is to consolidate business outcomes, understand pain points and identify areas of mutual interest which is why cross-functional workshops are recommended. This is so that the convergence of capabilities is agreed upfront, and joint business cases can be developed. Very often one function might drive the implementation of a technology but depends on other business units for adoption and integration.
An impact assessment also helps to understand what barriers exist, whether technical, operational or cultural, that may impede programme buy-in and outcomes. Targeted interventions can then be scoped to address these barriers which may include introducing new operating models and workflows, change management programmes or skills training.
Key outcomes from a stakeholder engagement process should include agreement on priority use cases and the allocation of roles documented in a RACI chart.
- Capability mapping and benchmarking
A capability is the demonstrated, consistent ability to perform against a desired objective and requires the appropriate combination of people, processes and technology. A function is a grouping of related capabilities. Each capability should have a clear definition that articulates the future vision for that capability which helps in clarifying role descriptions and core competencies. A capability model is a helpful way to outline what capabilities each function requires to achieve its goals and how they relate to each other. Capabilities can, for example, be mapped across categories, lifecycles or customer journeys.
A maturity assessment of the current state of capabilities against target state establishes where the gaps are but also where strengths are. Maturity scores can be depicted via scales, colour coding or spider charts. This step is important in determining where upskilling, outsourcing or insourcing is required and where to prioritise budget and time. Skipping this exercise could result in misaligned investments, ineffective change management, and difficulty measuring progress.
- Talent competency model and benchmarking
Similar to capability models, competency frameworks are necessary to ensure that the right people are given the right learning opportunities, and to avoid generic learning pathways that may cause confusion, anxiety or disengagement.
Frameworks are useful in:
- Internally benchmarking performance for both capabilities and job roles
- Setting timely learning and development interventions
- Giving employees autonomy in their career
- Encouraging and establishing a practice of lifelong learning
- Creating a universal set of expectations
- Benchmarking employee’s application of skills against industry standards
Once competency assessments have been done, learning paths can be customised along the continuum of current and required competencies. Companies can develop learning personas to help with personalising communication and material.
- Initiatives cataloguing, prioritisation and implementation roadmap
During the stakeholder engagement process, use cases were formulated and prioritised based on significance of pain points in relation to operational efficiency and customer service. These are then further mapped within an impact versus complexity matrix. Those that fall within the high impact and low complexity quadrant are naturally the quick wins that can be implemented immediately. These are great for proof-of-value and can be showcased as successful outcomes of the programme in an effort to garner further stakeholder buy-in and build employee trust. Those that are in the high impact and high complexity quadrant will require more planning and investments and most likely proof-of-concepts.
Initiatives are then catalogued and mapped on a roadmap as an easy visual reference. Each use case will have its own goal, KPIs, business requirements and project timeline.
This step prevents the following programme failures from happening:
- Fragmented and overlapping efforts
- Without cataloguing, multiple teams may run similar initiatives in silos, duplicating effort and wasting resources.
- Critical gaps can also be overlooked because no-one is tracking the full portfolio.
- Misaligned priorities
- Without prioritisation, “whoever shouts loudest” often gets resources,
- Low-value or politically driven projects may be funded while high-impact initiatives stall.
- Unclear sequencing and dependencies
- Some initiatives must happen before others such as data governance before AI.
- Without a roadmap, initiatives can be launched out of order, causing rework, delays or even failure.
- Resource strain and burnout
- Teams may be spread too thin across too many initiatives at once
- This creates delivery bottlenecks, missed deadlines and low morale.
- Lack of accountability and visibility
- Leaders won’t have a clear line of sight into what’s happening, why it matters or how it links to strategy.
- This erodes stakeholder confidence and weakens support for the programme.
- Difficulty demonstrating value
- Without a structured roadmap, it’s nearly impossible to track progress, measure ROI or show quick wins.
- The programme risks being seen as “busy work” rather than strategic.
- Operating model
Designing an operating model to implement the programme provides structure and governance because it defines who owns which deliverables, how decisions are made and how performance is tracked. It also prevents the programme becoming fragmented or too centralised. An operating model enables scaling and repeatability making it easier to roll it out across business units, geographics or functions.
Things to consider in an operating model is having a governing coalition to oversee the direction, budget, resources and sign-offs, and a centre of excellence headed up by the programme lead to guide the design, delivery, measurement, content and learning requirements. To co-ordinate regional implementation, a programme management office should be set up supported by an organisation’s shared services, such as communication, tech enablement, learning and development, working together in communities of practice. This is what is referred to as a “hub and spoke” operating model where the spokes are empowered to implement the programme in a way that is compatible with their own contexts, cultures and infrastructure.
- Programme measurement framework
Having a measurement framework is critical because it turns abstract goals into concrete, trackable progress. Without it the programme risks being seen as “training for training’s sake” instead of being a strategic enabler with measurable improvements in organisational capability.
Below is an example of a measurement framework that could be applied to a programme.
A complex programme such as building AI capability across a large organisation with multiple regional offices requires a very structured approach. The temptation to implement a generic playbook without diving deeper into what initiatives will drive the most value, may mean that the programme is seen as an expensive and time-consuming tick box exercise. Embedding Responsible AI principles across an AI capability building programme signals that ethical awareness is not an afterthought but is foundational to how teams will engage with AI. The seven components outlined above act as important guardrails for strategic programme design and value-based implementation.
