For information and reference only – not to be submitted as your own work.

This extract shows how a Masters-level report on agentic AI and change management can be structured and argued. It follows the three parts of the MGMT902 Assignment 3 agentic AI change management brief, and our planning guide to the MGMT902 agentic AI report explains the thinking behind each section. The organisation is a hypothetical NZX Main Board retail bank, used so that no real company is misrepresented; a real submission must anchor every claim in that company’s own annual report and NZX announcements. Earlier assessments in this course are covered separately in our MGMT902 Assignment 1 ICT strategy sample and MGMT902 Assignment 2 cyber security governance sample.

Planning table

Part 1: Implementing agentic AI-enabled change

What makes agentic AI different

Conventional robotic process automation executes fixed rules on structured inputs; when the process changes, a person must rewrite the script. Generative AI responds to a prompt and returns content, but a person decides what happens next. Agentic AI differs because the system is given a goal, plans a sequence of steps, calls other enterprise systems and takes actions with limited step-by-step instruction (Russell & Norvig, 2021). That shift in who initiates and completes work is why the change is organisational rather than purely technical.

For the hypothetical bank, the first use case is mortgage application triage. An agent reads submitted documents, checks them against the credit policy, requests missing items from the customer and routes only exceptions to a lending analyst. The analyst’s job moves from assembling files to judging borderline cases and auditing the agent’s reasoning.

Applying ADKAR to the workforce shift

ADKAR is well suited because it treats change as an individual journey through Awareness, Desire, Knowledge, Ability and Reinforcement (Hiatt, 2006). Table 1 maps each element to the analyst group.

Kotter’s (1996) eight steps complement this at the organisational level, particularly creating urgency and generating short-term wins, which keep sponsorship visible while individual ADKAR gaps are closed.

Innovation, agility and competitive advantage

Davenport and Ronanki (2018) argue that the most valuable AI projects enhance existing processes rather than chase novelty. For a bank in a competitive market, faster and more consistent decisions can improve customer experience and free staff for advice work. Advantage is sustainable only if the organisation builds the oversight capability that competitors find harder to copy than the technology itself.

Part 2: Overcoming barriers to agentic adoption

Barriers in New Zealand organisations

Six barriers recur. Each is paired with the evidence a strong answer would cite.

• Trust and explainability. Employees over-rely on or reject automation depending on how reliable it appears (Parasuraman & Riley, 1997). Trust is built on perceived ability, benevolence and integrity (Mayer et al., 1995), so agents must expose their reasoning.
• Data quality. An agent acting on inaccurate customer records scales errors faster than a human would. Information Privacy Principle 8 requires reasonable steps to ensure personal information is accurate before use (Privacy Act 2020, s 22).
• Legacy systems. Core banking platforms often lack modern interfaces, so agents depend on brittle integrations.
• Unclear accountability. When an agent errs, responsibility between the business owner, technology team and vendor must already be defined.
• Privacy and regulation. The Privacy Commissioner expects organisations to consider necessity, transparency and human review when using AI with personal information (Office of the Privacy Commissioner, 2023).
• Cultural inertia. Risk-averse cultures may treat any change in decision rights as a compliance threat.

Governance and assurance

The NIST AI Risk Management Framework organises risk work into govern, map, measure and manage functions (National Institute of Standards and Technology [NIST], 2023), while ISO/IEC 42001 sets requirements for an AI management system (International Organization for Standardization [ISO] & International Electrotechnical Commission [IEC], 2023). New Zealand’s national AI strategy favours light-touch, risk-based approaches built on existing law (Ministry of Business, Innovation and Employment [MBIE], 2025), so the bank’s own governance carries much of the burden.

Recommendations

1. Establish an AI governance committee with named executive accountability before any pilot.
2. Run the triage agent in shadow mode, comparing its recommendations with analyst decisions, before it acts on live files.
3. Complete a privacy impact assessment and data quality remediation for the fields the agent uses.
4. Require decision logs that a non-technical supervisor can read.

Part 3: Strategic change management plan

Use cases and business alignment

Use case one is mortgage application triage; use case two is fraud alert triage in which the agent gathers transaction context and recommends hold or release. Both align with a strategy of faster service and lower operational risk. Selecting two bounded cases keeps the plan testable and keeps the oversight model specific.

Stakeholder analysis

Communication, training and oversight model

Communication starts with a sponsor message explaining why the bank is adopting agents and what will not change, followed by team-level briefings. Training covers reading decision logs, challenging agent output and escalation. Decision rights follow a graded model: the agent may act alone on low-value, low-risk steps such as document requests; it recommends and a human approves for credit decisions and account holds; and humans retain sole authority over declines and customer complaints. The accountable executive owns outcomes, and every agent action is logged and attributable.

Reinforcement and measurement

Reinforcement includes recognition for high-quality oversight, role titles that reflect supervisory expertise and retrospectives after each release. Table 2 shows a measurement set, with targets that are illustrative figures only.

If the agent’s override rate rises or the survey falls, the plan pauses expansion until the cause is understood, which demonstrates that governance is operating rather than merely documented.

If you need a structured report of your own, our report writing help and management assignment help pages explain what we offer, or you can submit your assignment brief for a quote.

References

Davenport, T. H., & Ronanki, R. (2018). Artificial intelligence for the real world. Harvard Business Review, 96(1), 108–116.

Hiatt, J. M. (2006). ADKAR: A model for change in business, government and our community. Prosci Learning Center Publications.

International Organization for Standardization & International Electrotechnical Commission. (2023). Information technology – Artificial intelligence – Management system (ISO/IEC 42001:2023).

Kotter, J. P. (1996). Leading change. Harvard Business School Press.

Mayer, R. C., Davis, J. H., & Schoorman, F. D. (1995). An integrative model of organizational trust. Academy of Management Review, 20(3), 709–734.

Ministry of Business, Innovation and Employment. (2025). New Zealand’s strategy for artificial intelligence: Investing with confidence. https://www.mbie.govt.nz/business-and-employment/economic-growth/digital-policy/new-zealands-ai-strategy-investing-with-confidence

National Institute of Standards and Technology. (2023). Artificial intelligence risk management framework (AI RMF 1.0) (NIST AI 100-1). https://doi.org/10.6028/NIST.AI.100-1

Office of the Privacy Commissioner. (2023). Artificial intelligence and the information privacy principles. https://www.privacy.org.nz/assets/New-order/Resources-/Publications/Guidance-resources/AI-Guidance-Resources-/AI-and-the-Information-Privacy-Principles.pdf

Parasuraman, R., & Riley, V. (1997). Humans and automation: Use, misuse, disuse, abuse. Human Factors, 39(2), 230–253. https://doi.org/10.1518/001872097778543886

Privacy Act 2020 (NZ). https://www.legislation.govt.nz/act/public/2020/31/en/2024-07-01/

Russell, S., & Norvig, P. (2021). Artificial intelligence: A modern approach (4th ed.). Pearson.