In Part 1 of this series, we explored the potential emergence of the FP&A Agent Manager as AI agents move into planning and decision-support workflows.

Sophia, a junior FP&A analyst, is collaborating on the quarterly revenue forecast. The agent's segment projections appear credible, but for industrial coatings, it misses recent procurement consolidation signals absent from the structured data. The model trends up, misaligned with commercial reality. The senior flags the segment, documents the divergence, and reports the data gap to the agent's team along with a request to retrain the model on more recent data. Sophia writes revised commentary and prepares for next month's review. She learns to spot signals beyond data through real-time guidance, gaining judgment by exposure rather than formal training.

FP&A used to progress through data collection, variance analysis, commentary, business partnering, and strategic advice, with each stage building skills for the next. Now, Artificial Intelligence (AI) streamlines the initial steps for greater productivity but disrupts the traditional pathway for developing judgment. As Wharton’s Cornelia Walther observes, AI is now automating entry-level tasks that once developed senior professionals, leaving gaps in talent pipelines [1].

This is already happening at an organisational scale. IBM redesigned entry-level jobs to prioritise analysis, customer interaction, and oversight of AI rather than just automating routine tasks [2]. Junior software developers now engage more with clients; HR staff review and correct AI chatbot responses. The CFA Institute warns that companies that do not adopt such changes risk lacking qualified mid-level talent in the future.

The developmental risk is becoming visible in a recent study on AI-Human cooperation [3]. While AI boosts task performance, it reduces motivation and increases boredom for later solo work. For finance, this suggests AI completes tasks rather than develops skills. WEF confirms that entry-level staff have lower confidence in their career abilities [4]. Seminal research on motivation and development from the University of Rochester suggests automating early roles without new advancement paths may weaken autonomy and hinder talent growth [5]. Where AI disrupts the traditional pathway for developing judgment, the question is how to design career progression paths in an AI world to prevent breaking the talent pipeline.

The progression sequence should reflect tasks in an agentic environment, including pattern recognition, exception triage, workflow design, and strategic advisory. Early steps involve analysing and validating outputs, then feeding failures back into the system. Wharton’s Walther introduces the GROOM framework to redesign the progression path for the AI era.

G – Gap analysis

As Walther puts it,

"Organisations cannot simply eliminate junior positions and expect skilled professionals to emerge spontaneously." 

In finance, automation and AI can assist with most tasks, but will there be enough professionals who understand data lineage, complex problems, or can navigate the next downturn? How to defend their analysis to management?

R – Redesign Pathways

Create entry-level and developmental roles that combine AI support with human learning. Let junior staff tackle complex problems while using AI for repetitive work. The job characteristics model [5] can help make roles more varied, challenging, and motivating. It suggests granting junior analysts decision-making authority, using AI to teach reasoning, and enabling career advancement by keeping employees in high-value roles. Assigning junior analysts to maintain audit trails of AI-supported decisions helps them build reasoning skills and ensures operational control.

O – Optimise Knowledge Transfer 

Wharton suggests pairing junior analysts with experienced colleagues for judgment-heavy work, while McKinsey [6] emphasises integrating apprenticeship into daily work by having seniors model duties that AI cannot handle. This is what Sophia and her senior colleague are doing in practice.

O – Organise Cross-Functional Exposure 

Offer rotation programs and cross-functional projects to broaden professional experience, instead of specialising. Employees can rotate through operational and commercial roles to build business acumen and use a sandbox environment for prototyping solutions via vibe coding in IT-led workshops.

M – Monitor and Measure

Existing HR processes, like performance reviews and skills matrices, can be extended to track AI-related capabilities rather than building parallel systems.

Monday Moves: From Insight to Action

Three actions for FP&A leaders who want to start redesigning the path this week.

1. Map your current progression sequence against what AI now handles. Take your existing analyst-to-manager development path. Mark the tasks at each level that an agent can already do competently. The pattern across most finance teams is the same: the first three rungs are largely automatable, and those rungs were where judgment was historically built. That gap is the design problem you are solving for.

2. Restructure one review meeting as a coaching conversation, not a quality check. Pick a recurring session: month-end variance review, forecast challenge, board pack sign-off. In the agentic model, the numbers arrive correctly. The senior's job is no longer to verify them; it is to teach the analyst which patterns matter and why. Run the next session on that basis and observe what changes in what the analyst learns.

3. Have the conversation with HR before you need to. Role descriptions, levelling frameworks, and progression criteria across most finance functions still describe work that AI now performs. Initiate the redesign conversation now, while the gap is small. Waiting until the first cohort of agent-managed analysts comes up for promotion is waiting until the structural problem becomes a personnel problem.

The direction is clear. As AI agents move deeper into FP&A workflows, the teams that start redesigning roles, progression paths, and development structures now will build the talent pipeline internally. The risk is not that FP&A loses analysts. It is that it loses the process through which judgment was built.

Sources

1. Walther, C. C. (2025, August 12). Is AI pushing us to break the talent pipeline? Knowledge at Wharton. https://knowledge.wharton.upenn.edu/article/is-ai-pushing-us-to-break-the-talent-pipeline/
2. Szkutak, R. (2025, February 12). IBM will hire your entry-level talent in the age of AI. Tech Crunch. https://techcrunch.com/2026/02/12/ibm-will-hire-your-entry-level-talent-in-the-age-of-ai/
3. Przegalinska, A., Ciszek, K., Mazurek, G., Joinson, A., Zukowska, E., & Bryl, L. (2025). Human–generative AI collaboration enhances task performance but undermines human's intrinsic motivation. Scientific Reports, 15, 98385. https://doi.org/10.1038/s41598-025-98385-2
4. World Economic Forum & PwC. (2026, January). How AI is changing early careers: A view from entry-level workers. World Economic Forum.
5. Deci, E. L., Ryan, R.M. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55(1), 68–78.  
6. McKinsey & Company. (2026, May 18). Rethinking early-career talent in the agentic organization.