Who, what, when, where, why — in one line: In April 2026 the U.S. Securities and Exchange Commission (SEC) and Department of Labor (DOL) issued joint draft guidance proposing operational and disclosure rules for automated and AI-driven retirement-advice tools; the public comment period closed in June 2026 and, as of Sept. 15, 2026, the agencies have not issued a final rule, leaving industry and savers to plan for likely but not yet finalized changes.

Context: why this matters now

Automated planners and robo-advisors are increasingly used for high-stakes retirement decisions — Roth conversions, 401(k) rollovers, pension lump-sum offers and RMD management — where tax timing, longevity and benefit interactions can change outcomes materially. The April proposal would not outlaw algorithmic advice but would require documented individualized suitability checks, transparent tax modeling (including the treatment of RMDs and Social Security taxation), model governance and conflict disclosures.

The 60-day notice-and-comment window the agencies opened in April closed in June 2026, meaning regulators are now reviewing the public record. That record — and voluntary changes firms made while the proposal was pending — will shape whether the final guidance becomes prescriptive rulemaking or remains higher-level principles with phased implementation.

What changed between July and Sept. 2026

  • Comment themes consolidated. Public filings show recurring industry themes: (1) requests for clearer definitions of “automated” vs. “human-supervised” advice; (2) concerns from smaller fintechs about compliance costs for backtesting and model governance; (3) consumer-advocate calls for explicit safeguards for near-RMD-age clients.
  • Voluntary industry moves. Several large custodians and robo-advisors announced interim steps to address the proposal’s core concerns. Common changes include adding RMD-aware Roth-conversion scenarios to user-facing screens, publishing model-assumption summaries, and offering one-click options to download PDFs with tax-projection assumptions.
  • Standards and tech guidance surfaced. Providers began citing the National Institute of Standards and Technology’s (NIST) AI Risk Management Framework and industry “model-governance playbooks” as templates for audits and documentation.
  • Regulatory posture. Neither the SEC nor the DOL has issued a date for final action. Agency staff have said in public statements that they are weighing trade-offs between consumer protection and preserving low-cost digital access to advice.

New practical examples and why they matter

Updated scenario: a 68-year-old with a $800,000 traditional IRA receives an automated recommendation to convert $200,000 to a Roth. Under the draft guidance such a platform would need to display side-by-side scenarios showing:

  • Current-year tax impact: estimated taxable income delta and likely marginal federal and state tax-bracket exposure.
  • RMD interplay: how partial conversion changes RMD base in subsequent years and whether full conversion before the RMD year is feasible.
  • Benefit interactions: whether the conversion pushes provisional income over Social Security taxation thresholds or increases Medicare Part B/D IRMAA surcharges.

That level of modeling frequently changes the recommendation. For example, showing a three-year conversion schedule often reduces immediate tax spikes and minimizes IRMAA exposure — a detail many earlier “one-click” recommendations omitted.

Industry and adviser ramifications — what changed in practice

Firms are taking different tactical approaches while the final rule remains pending:

  • Large custodians are expanding model-governance teams and publishing more detailed assumption summaries in client portals.
  • Robo-advisors have begun adding explicit RMD and Social Security modules to Roth-conversion flows and are saving model-output PDFs automatically to client records.
  • Small fintechs report higher compliance burdens and are forming consortia to develop shared testing frameworks to spread costs.
  • Advisers who use automated tools are increasingly documenting human oversight steps and suitability checks to preserve fiduciary defense-in-depth.

Impact on savers — what to do today

If you are actively making Roth, rollover, or pension-lump-sum decisions, take these four immediate steps:

  1. Ask for the assumptions and outputs. Request the full tax-model output (PDF or screenshot) that shows taxable income, marginal tax rate, RMD projections and Social Security/IRMAA impacts for any recommendation.
  2. Check for individualized suitability. Confirm the tool used your actual age, planned retirement date, expected longevity assumptions, and other income sources — not just an age band or “typical retiree” profile.
  3. Document human review. If an adviser signs off, ask how they reviewed the algorithm’s recommendation: what checks did they run, and can they supply a written rationale?
  4. Watch for conflicts. Demand disclosure of revenue-sharing, annuity referral fees, or proprietary-product incentives tied to the recommendation.

Reactions — who said what

Consumer groups emphasized stronger guardrails; industry trade associations pressed for clearer technical definitions and phased compliance. Firm statements during the comment window generally welcomed transparency goals while warning that overly prescriptive backtesting rules could raise costs and reduce access for smaller providers. Regulators have acknowledged both concerns in public remarks and are reviewing the comment record.

What’s next — timeline and signals to watch

Key milestones for readers to monitor:

  • Final rule or guidance — watch agency bulletins and the Federal Register for any notice of final action. As of Sept. 15, 2026 no final rule has been published.
  • Implementation window — industry participants are budgeting for a 12–18 month phased compliance period if the guidance becomes final and prescriptive.
  • Market behavior — expect more providers to publish assumption summaries and to add RMD-aware scenarios even before a final rule, as an interim consumer-protection measure.

How to interpret the uncertainty

Uncertainty benefits no saver. Treat algorithmic recommendations as a starting point, not a final decision. Preserve decision records (screenshots, PDFs), ask for human verification for complex tax interactions, and if you’re near RMD age (typically 72–75 depending on your birth year), give extra scrutiny to any conversion advice.

Frequently asked questions

Has the April 2026 proposal become a final rule?

No. The draft guidance opened a 60-day comment window in April and that window closed in June 2026. As of Sept. 15, 2026, the SEC and DOL had not published a final rule. Watch the Federal Register and agency press releases for formal notice.

Do robo-advisors already follow the proposed disclosure practices?

Some large providers have voluntarily adopted portions of the proposal — for example, adding downloadable tax-projection reports, RMD-aware scenarios, and clearer conflict disclosures. Practices vary by firm; always ask your provider for the specific outputs and documentation they produce.

What should I insist on when offered a Roth conversion or rollover by an automated tool?

Insist on: (1) a downloadable tax-projection report showing immediate and multi-year taxable-income impacts; (2) explicit treatment of RMD timing and Social Security taxation; (3) a written suitability determination (or adviser sign-off) that accounts for your full-income picture.

Will these rules make advice more expensive?

Possibly. Firms face costs for enhanced model governance, backtesting and disclosures. Large custodians can amortize those costs more easily than small startups. That said, many providers are adopting partial changes now to keep services competitive while preparing for final rules.

For retirement planners and savers, the bottom line: transparency and documentation are the immediate wins you can demand today. Whether the agencies finalize prescriptive rules or keep to principles, well-documented modeling and human oversight will be the best protection against costly automated missteps.