A CO-AUTHORED PERSPECTIVE
The strategic role of the Chief People Officer (CPO) in AI implementations
Five questions for the CPO
KSIB has written extensively about the need for collaboration to drive successful enterprise AI implementations at scale. It is important that AI is not seen as “another technology” that sits within the remit of the CIO/CTO. In addition to the solutions needing to be business driven/led, there are significant roles to be played by the CFO to manage risks and costs and the CPO to ensure that the entire system, people and agents, is working effectively. Given the newness of this technology and its application at an enterprise level, this presents an opportunity for CPOs to play a strategic role in designing the future workforce to capture the benefits of AI investments for both the organisation and for individuals. This can be done by understanding the system as it operates today, what it needs to look like going forward, and the key ingredients for the transition. Five key questions CPOs need to consider are outlined in this article:
- How do you baseline what your people are doing now, at the work activity / task level?
- What do you need humans to do and agents to do to deliver your strategy?
- How are you designing the system to achieve the optimal outcome and manage the risk (in collaboration with the CFO, the CRO and the business owner)?
- How are you measuring success and designing for future change in the system going forward – both humans and agents?
- What is the workforce strategy you need to put in place to make all of this work?
What are your people doing now?
Research from MIT considers the root cause of those generative AI pilots that have failed and or delivered no measurable P&L impacts and points to “the learning gap”.1 This is the failure of organisations and their people to absorb AI into how work gets done. In the experience of KSIB and Beamible, who have worked together on foundational workforce design to support AI transformations, one of the key challenges is that AI attaches to tasks, not roles or job descriptions. Many organisations do not understand the type or extent of tasks that are currently being performed, or the new activities which will be needed with AI in place.
The case study outlined below provides a practical example of how current state tasks can be understood and visualised, in order to inform workforce design implications, prior to an AI transformation commencing.
1. Challapally, Aditya, Chris Pease, Ramesh Raskar, and Pradyumna Chari. 2025. “The GenAI Divide State of AI in Business 2025 MIT NANDA.” cloudelligent.com/.../v0.1_State_of_AI_in_Business_2025_Report.pdf
Case study: baselining current workforce tasks
KSIB and Beamible have worked together with ASX 50 clients to prepare and analyse foundational tasks as the baseline to enable AI transformations. Key elements of this work include:
Key learnings from some of KSIB and Beamible’s work to understand and visualise current tasks include:
What do you need humans to do and agents to do?
The most important organisational design decision to be made in an AI implementation is what you require humans to do and what you would like agents to do. This is a different question to “what are agents capable of doing” and requires strategic focus. Given that most employees will already be using AI to augment their work, whether by design or not, and the fact that there are significant risks inherent in AI systems that need to be mitigated, the importance of organisational and role design is paramount. MIT’s research found employees using personal AI tools in over 90 percent of the companies studied, despite only 40% of those companies having an official LLM subscription.1
Some of the key criteria impacting the organisation design question include:
The agent or the human needs to be capable of performing the task assigned. The new risks created by the changed environment need to be monitored for both agents and humans.
Understanding the true costs of the AI enabled system and the relevant unit costs for key design decisions are important, including the cost of “shadow AI”.
These need to be understood and designed for in both the old and the new systems. It is particularly important to understand where judgement is required.
There must be human oversight and accountability, regardless of the agent contributions. This is already included in the EU AI Act and is an expectation of Australian Boards.
Some interactions require human beings to build critical trust and relationships. This is not possible with an agent. Similarly, trust needs to be maintained or built within the team to drive outcomes.
Conscious workforce investment decisions are being made by thoughtful organisations to recognise the skills that will be needed for the future. Development pathways and ongoing training need to be redesigned in this new era. Leadership training will be important as we are entering a completely new paradigm.
1. Challapally, Aditya, Chris Pease, Ramesh Raskar, and Pradyumna Chari. 2025. “The GenAI Divide State of AI in Business 2025 MIT NANDA.” cloudelligent.com/.../v0.1_State_of_AI_in_Business_2025_Report.pdf
How are you designing the system?
It is our belief that the dominant work pattern will be neither human nor agent but humans leading agents, which means the real design task is defining the interaction: who checks, who decides, who is accountable.
This really does require the innovation system to be working well and considers work allocations carefully. A simple allocation framework might look like this, but will be different for each organisation:
Human retains
When it applies
Accountability, trust, novelty or capability-building value is high
Examples
- Final credit decisions
- Sensitive client conversations
- People management and leadership
Design task
Protect the task explicitly; document why it must remain human
Agent performs
When it applies
High volume, repeatable, verifiable, low blast radius when wrong
Examples
- First draft documents
- Data assembly
- Routine enquiries
- Scheduling
Design task
Define the agent’s role, KPIs, escalation triggers and owner
Human leads agent
When it applies
Judgement plus scale; most knowledge work
Examples
- Analysis
- Forecasting
- Research
- Drafting for review
Design task
Specify who initiates, who verifies, who decides, who owns the outcome
Note: the above allocation framework addresses some of the CPO considerations but does not deal with risk management or costs which need to be carefully understood and managed and will often be led by the CFO (in collaboration with the CRO where applicable).
How are you monitoring the system?
In practice, the CPO will need to work with business leaders and the CTO/CIO, CFO and CRO to effectively design the system. From a CPO’s perspective, there are 3 critical principles to follow:
Accountability and delegations mapping
“The system did it” must never be an available answer.
Escalation by design
Agents need defined conditions to hand over and humans need the time, skill and authority to receive the hand over.
Role redesign, formalised
If a human role changes dramatically, this needs to be reflected in a genuine role redesign.
How do you monitor the system going forward, humans and agents both?
Agents need to be monitored continuously for both performance and risk outcomes. Every production agent should have a defined role and scope, an owner, measurable performance indicators, a review cadence and offboarding criteria. This should dovetail with the monthly CFO reporting pack. Organisations should consider independently monitoring and managing agents from a performance and risk perspective. AI enabled tooling is available to help with this.
From a human perspective, recognition of the higher value work that is to be performed, the verification burden that may arise, the time it takes to understand the system and the agents, all have performance and wellbeing implications. Humans will continue to be a key differentiator for organisational performance – driving innovation, strategic direction and low-risk execution. Considering the work which energises people, and getting them in the right role, with the right skills, will therefore be critical.
What workforce design do you need?
Workforce design in the AI era is a rolling, funded, board visible discipline. The World Economic Forum’s Future of Jobs Report 2025 draws on over 1,000 employers and projects 170 million jobs created and 92 million displaced globally by 2030, with employers expecting 39 percent of core skills to change in the same period.1
Key considerations in any workforce plan dealing with humans and agents include:
Flexibility and sprints
Given the evolution of technology and organisations, a high-level plan is important, but the detailed supporting plans need to be sprint based, flexible and dynamic. The evolution of roles can be considered as part of the overall plan.
Reskilling
Gartner has found that acquiring and developing AI and digital talent is a near term challenge for finance chiefs and the same is true across all functions.2 The realistic answer for Australian organisations is to invest in building capability through reskilling of key talent. This requires careful planning and investment, including a capability gap analysis. Leadership skills will be critical in this time of transition.
Funding
Reskilling and workforce redesign requires a budget.
Talent pipeline
If agents absorb much of today’s “junior” work, tomorrow’s mid level capability must be built deliberately.
Role clarity
Once the system is designed, role clarity, including accountabilities and delegations, is critical for success.
Critical thinking and judgement
This will be the key skill of the future – a culture that supports and encourages this will be a differentiator. This type of culture will only emerge through deliberate design, focus and leadership.
1. Zahidi, Saadia. 2025. World Economic Forum | Future of Jobs Report 2025.
weforum.org/publications/the-future-of-jobs-report-2025
2. “Gartner Survey Reveals Acquiring and Developing AI and Digital Talent Is CFOs’
Top Near-Term Challenge.” 2026. In Gartner.
gartner.com/en/newsroom/press-releases/2026-03-23-gartner-survey-reveals-acquiring-and-developing-ai-and-digital-talent-is-cfos-top-near-term-challenge
A word on the people impacts of transformational change driven by AI:
This article has been written from the perspective of people who are involved in large scale, enterprise AI implementations. We have tried to avoid hype and focus on our own experiences and broader evidence that supports our views. We know that many team members we speak to at our clients are really worried about job security and are fearful of making mistakes or inadvertently breaching company policies when it come to AI.
In our view, there is a leadership responsibility, and indeed opportunity, to clearly communicate intentions and process with respect to AI implementations. An organisation that focuses on its customers and its people will emerge as market leaders as business models change. We have not found one example where focusing purely on cost cutting or implementing AI in a silo (such as IT) has delivered the benefits as promised. We know of several Australian market leaders focusing on customers, people and risk management, where the potential of AI is being realised.
To discuss how to get started on this leadership opportunity, contact us at www.ksib.com.au or email Kristin or Sarah directly at kristin@ksib.com.au or sarah@ksib.com.au. If you would like to get in touch with workforce design experts at beamible.com, email Stephanie at stephanie@beamible.com.
We work in partnership where strategy and workforce design, analytics and future workforce modelling are needed in order to make critical decisions in an AI enabled world.
×
Liability limited by a scheme approved under Professional Standards Legislation.