From retirement estimates to member engagement

The power and accessibility of artificial intelligence (AI) is raising expectations among consumers in relation to retirement planning. Generative AI provides instant, personalised support, which some users may perceive as financial advice, particularly where the response includes retirement estimates or projections. To the savvy, financially literate consumer, AI is a magic lamp; for the unwary, a potential house of cards.

Super providers possess (hopefully) reliable retirement models, and limited data about their members. Members know their own circumstances, while Generative AI holds out the promise of a tailored retirement model. Regulators are rightly focused on ensuring consumers are not misled. In a world where personalised retirement estimates (accurate or not) can increasingly be generated outside of regulated channels, what role should regulated retirement estimates play?

In the Actuaries Digital article on Mandatory, Standardised Retirement Estimates , the authors proposed amendments to regulations that would improve the accuracy and usefulness of estimates to members. This article explores two ways retirement estimates can be enhanced:

• How retirement estimates could be made more relevant with the use of more personal data and preferences, and
• Applications of retirement estimates to educate members, prompt engagement and monitor retirement readiness.

Part A: Harnessing personal data and member preferences

How can personal data make retirement estimates more accurate?

Retirement estimates are currently required to use only a member's age, balance, employer contributions, insurance premiums and administration fees. While these are the most important items, additional data readily available to providers and members can significantly enhance accuracy.

Asset allocation is a clear example. The impact of investment choice in projecting superannuation benefits is substantial, particularly for those in lifecycle options where built-in derisking should be prominently shown. Similarly, voluntary contributions known to providers can be incorporated via credibility algorithms to determine which contributions warrant inclusion. Many providers are already incorporating these factors into retirement estimates.

Member circumstances beyond balance, contributions and asset allocation are equally important. Partner status, for instance, materially affects the Age Pension component of retirement income – noting whether a member has nominated a partner as a beneficiary could help determine which status to use. Homeowner status, critical to Age Pension eligibility, can sometimes be inferred from other sources but is rarely incorporated.

Many of the most valuable inputs for a holistic retirement estimate are preferences known only by members themselves – information providers do not currently hold:

• Desired retirement age and lump sum needs
• Preferences for the “shape” (including level, duration) of retirement income
• Other assets and income, and bequest goals
• Tolerance for different types of risk
• Preferences for different retirement income products

These preferences are critical drivers of retirement outcomes.  Members may be willing to share this information in exchange for a more personalised estimate that helps them understand whether they are on track. Providers can then use this information to prompt members as relevant dates or targets approach, turning retirement estimates into proactive engagement tools.

The challenge lies in ensuring information gathered remains credible for future estimates. Providers should establish protocols for collecting member preferences – whether through prompts, interactive portals, targeted questionnaires, or digital advice journeys – that maintain data integrity while respecting member privacy. By combining data held by providers with member-provided information, retirement estimates can evolve from cookie cutter projections into relevant and useful forward-looking guides better tailored to each member's circumstances and aspirations.

Part B: Using retirement estimates to educate, engage and assess readiness

How can retirement estimates educate members about risk?

Retirement estimates serve a powerful educational function. When designed effectively, they translate complex financial risks, such as investment volatility, sequencing risk, longevity risk, inflation, from abstract concepts into tangible, relatable illustrations.

A key educational application is scenario analysis. Rather than presenting a single projected deterministic outcome, retirement estimates could illustrate a range of outcomes under different courses of action, for example, additional voluntary contributions, delayed retirement, or adjusted investment strategy. By showing the impact of member decisions on retirement income rather than account balances, estimates help members distinguish between factors they can control (contribution levels, investment strategy, retirement timing) and factors beyond their control (investment returns, inflation, longevity). Retirement estimates could also illustrate a range of potential stochastic or historical outcomes, helping members understand the inherent uncertainty in their future retirement outcomes. This transforms estimates from passive statements of average outcomes into panoramic illustrations for informed decision-making.

Retirement estimates also create opportunities for timely, targeted engagement. Members projected to fall short of desired income could be prompted to review contribution strategy. Those approaching retirement should be encouraged to compare drawdown options with lifetime retirement income products. Most importantly, retirement estimates can reduce uncertainty and build the confidence that comes with visibility of potential outcomes – many Australians remain uncertain about retirement preparedness even when objective measures suggest they are on track.

How should retirement estimates change across life stages?

Members' needs vary significantly across their retirement journey. For younger cohorts (ages 25–45), when retirement is decades away, emphasising personal milestones for super balance (e.g., 1x annual salary) can catalyse early engagement and encourage behaviours like voluntary contributions without burdening members with premature decumulation complexity.

As members approach retirement (ages 46–60), the focus should shift from balance accumulation to income replacement. A replacement rate framework, calibrating target retirement income as a percentage of pre-retirement earnings (typically 60–70%), accounting for reduced work-related expenses, provides a benchmark members can directly compare against their current standard of living. Incorporating Age Pension entitlement estimates becomes critical at this stage. For members in retirement (ages 60+), showing a range of outcomes with confidence intervals (e.g., 75% or 90% certainty that a given drawdown rate will sustain to age 95) helps to explain the sequence-of-returns risk and market volatility that single-point estimates fail to capture. Showing how the Age Pension functions as an indexed buffer during market downturns and the role of flexible income and lifetime income can play is particularly powerful in shifting retirees’ focus from defensive capital preservation to confident, sustainable consumption.

Retirement estimates can also serve as a gateway to deeper engagement – guiding members toward digital guidance tools, intra-fund advice, or comprehensive financial advice where appropriate. Viewed through this lens, retirement estimates are far more than projections; they support financial literacy, invite members to engage, and provide a pathway for members to access personalised guidance suited to their circumstances and goals.

Conclusion

Regulators' concerns about consumer protection are valid, but regulation should also enable retirement estimates to evolve with members’ needs. Greater use of personal data, and reasonable assumptions about members’ circumstances and preferences could make estimates more relevant and useful. Combined with a broader role in education, engagement and measuring retirement readiness across different life stages, regulated retirement estimates could provide members with a trusted alternative to AI-generated estimates that may appear accurate and personalised without necessarily being reliable.