Artificial intelligence can now assist with almost every stage of financial planning, from gathering client information and modelling scenarios to generating strategies and triggering reviews. But each use brings its own oversight demands – and professional judgment remains central to the advice ultimately given to the client.
New guidance from the Financial Planning Standards Board (FPSB) maps potential AI applications across all six elements of the financial planning process and sets out the oversight expected of financial planning professionals at each stage.
The Practice Guidance Note: Use of Artificial Intelligence in Financial Planning, released on 24 June 2026, also warns that the challenge is not confined to catching obvious factual errors or “hallucinations”. AI outputs may be incomplete, biased, or shaped by assumptions, interaction patterns, and a restricted range of alternatives. The FPSB says planners should also consider whether conflicts of interest may be embedded in an AI model.
The guidance is intended to help financial planning professionals understand their professional and ethical obligations when using AI. It complements the FPSB’s global professional standards but does not replace local laws or regulations.
The FPSB is the non-profit global standards-setting body for the financial planning profession and owns the international CFP certification programme outside the United States.
AI across the planning process
The guidance identifies potential uses for AI from the beginning of the client relationship through to ongoing review. But it does not treat all uses as equivalent or prescribe a single form of oversight.
When establishing and defining the relationship with a client, AI may assist with drafting communications, conducting background research, and preparing document templates. The FPSB expects professional review and clear disclosure of AI use.
During the collection of client information, AI may support data gathering, document analysis, and preliminary risk profiling. Here, planners should verify that the information is complete and conduct qualitative verification through discussion with the client.
For analysing and assessing a client’s financial position, potential uses include scenario modelling, risk assessment, and strategy research. The FPSB says calculations should be validated, possible biases in the AI tools detected, and professional judgment applied.
AI may also assist with strategy generation, presentation materials, and sensitivity analysis when recommendations are developed and presented. But professional judgment should determine the final recommendations, taking account of any AI-tool bias and tailoring the advice to the client.
During implementation, AI may support document preparation, product research, and progress tracking, subject to professional oversight and appropriate product due diligence. In ongoing reviews, it may be used for performance monitoring, environmental scanning, and review triggers, with planners expected to assess the results professionally, validate calculations, and evaluate whether adjustments are needed.
The guidance’s central proposition is that AI should support rather than substitute for professional expertise, critical thinking, and human oversight. Regardless of how sophisticated the technology becomes, the financial planning professional remains responsible for the final work product communicated and delivered to the client.
“AI is reshaping the practice of financial planning, but trust, professional judgement, and accountability remain essential,” Dante de Gori, the FPSB’s chief executive, said in a media release.
A plausible answer may still be a poor answer
The FPSB distinguishes between assisted AI, where a tool supports a specific task while the professional retains direct control; augmented AI, which enhances professional capabilities through deeper collaboration; and autonomous AI, which operates with minimal human intervention, such as automated rebalancing or robo-advice platforms.
It also distinguishes explainable AI, which provides reasoning that can be assessed, from “black-box” AI, whose opacity creates accountability challenges.
These distinctions matter because different tools have different limitations. Generative AI can draft communications, summarise documents, and support preliminary research, but the risk of hallucinations requires human validation. Analytical AI can assist with cash-flow analysis, retirement projections, and portfolio stress testing, yet its reliance on historical patterns may leave it unable to anticipate unprecedented events.
AI-enhanced planning software may support tax optimisation, rebalancing recommendations, and opportunity alerts, but the FPSB says advice based on AI outputs requires validation by a competent professional. Practice-management tools, such as AI-enabled client relationship management and scheduling systems, may pose lower risk because they have less direct impact on advice.
The guidance identifies hallucinations, bias, data dependencies, context limitations, over-reliance, incomplete outputs, and confirmation bias among the risks associated with current AI systems.
Some of these risks are less obvious than a fabricated fact or incorrect calculation. AI may omit material information, assumptions, or relevant considerations without signalling that anything is missing. It may produce answers aligned with a user’s expectations or interaction patterns rather than independent, objective analysis.
Under the Objectivity principle in the FPSB’s professional framework, planners should critically evaluate AI outputs and consider whether conflicts of interest may be embedded in a model, including in the assumptions it makes and the range of alternatives it considers.
The implication is that checking whether an AI-generated answer is factually correct may not be enough. A planner may also need to interrogate what the system assumed, what it omitted, whether bias affected the output, and whether relevant alternatives were adequately considered.
The FPSB further cautions that AI cannot be depended on to fully grasp nuanced personal circumstances or the emotional dimensions of financial decisions.
Adoption has run ahead of governance in some firms
The guidance follows global FPSB research into AI use by 6 206 financial planners across 24 territories, including South Africa.
As Moonstone previously reported, the FPSB’s survey found that nearly two-thirds of financial planners were already using, piloting, or planning to adopt AI, although only 45% said their firms had an AI policy or guidance in place and 18% said one was being developed.
The research also found substantial optimism about the technology: 78% of respondents believed AI would help them better serve clients and 60% believed it would enhance the quality of financial advice. At the same time, data privacy and cybersecurity, cited by 47%, and the accuracy and reliability of AI outputs, cited by 42%, ranked among the leading concerns.
The new guidance translates these broad opportunities and risks into expectations linked to the FPSB’s Financial Planner Code of Ethics and Professional Responsibility, including Client First, Integrity, Objectivity, Fairness, Professionalism, Competence, Confidentiality, and Diligence.
Disclosure and client data
Under the Integrity principle in the FPSB’s professional framework, financial planning professionals should disclose to clients any use of AI in the financial planning process or otherwise in providing client services, including client communications.
The FPSB says disclosure should occur as early as possible, ideally before engaging the client, and should be open, honest, and understandable. It should cover benefits, known limitations, risks, and any additional costs to the client arising from the use of AI.
The guidance also gives a direct warning about client information. Financial planning professionals should not enter or upload client data into public-facing AI tools, such as publicly accessible generative AI tools, or into other AI tools where privacy cannot be assured.
When in-house or vendor-provided AI is used, the FPSB says robust data-protection measures, access controls, and compliance with local privacy laws are essential. Planners should also consider applicable data-residency and sovereignty requirements.
Competence is another recurring theme. Financial planning professionals should develop the knowledge and technology skills needed to investigate AI tools, understand their ethical implications, and distinguish between uses aimed mainly at efficiency and those supporting complex decisions that may demand greater human judgment.
The FPSB says AI outputs and associated advice should be reviewed by someone with the appropriate knowledge and skill to do so, whether that is the financial planning professional or another person.
The guidance also urges planners to cross-check calculations, look for patterns of bias, and question the reasoning behind outputs. A polished presentation should not be mistaken for accuracy, and planners should be alert to AI producing agreeable or convenient answers based on interaction patterns rather than objective analysis.
A separate South African regulatory track
The FPSB guidance is not part of South Africa’s regulatory framework, and its specific expectations should not be conflated with requirements imposed under local financial-sector or privacy law.
But it arrives as South African financial regulators continue developing their own response to the use of AI in the financial sector.
As Moonstone reported on the FSCA’s 2026–2029 Regulation Plan, AI remains a strategic focus area for the Financial Sector Conduct Authority.
Following earlier research by the FSCA and Prudential Authority into AI adoption and associated risks in the financial sector, the authorities have initiated further research and intend to develop a discussion paper for stakeholder engagement on key regulatory and supervisory questions. High-level AI governance principles are also intended to be incorporated into the planned Joint Standard on Culture and Governance requirements.
The FPSB and South African regulatory work therefore operate on separate tracks. Even so, their areas of concern overlap in important respects, including governance, accountability, and risks arising from the use of AI in decisions and processes that can affect clients.
For financial planning practices, the FPSB guidance raises practical questions that go beyond whether an AI tool saves time: which tools are being used and how they operate; what assumptions they rely on; whether and how client data is processed; what privacy safeguards apply; how outputs are checked; whether reviewers are competent to assess them; and where professional judgment must determine the final outcome.
The FPSB states that responsible AI integration can improve service quality, expand access to financial planning, and reduce costs. But it says realising these benefits while maintaining professional standards requires thoughtful implementation.
“Financial planning remains fundamentally a human endeavour; clients seek understanding, wisdom, and partnership to achieve their personal goals,” the guidance states.
