> For the complete documentation index, see [llms.txt](https://help.filed.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.filed.com/reference/products/ai-tax-planning.md). # AI Tax Planning ## What AI Tax Planning does AI Tax Planning reads a client's finalized tax return and produces forward-looking tax strategies with estimated savings. Where AI Review finds errors in a draft, AI Tax Planning looks ahead and surfaces opportunities to present during a planning conversation. * Each strategy cites the specific figures on the return that make it applicable. * Savings estimates are calculated from arithmetic on the return's actual numbers, not guessed. * Strategies are grouped by planning domain and sorted by estimated savings. * The plan also includes a written narrative overview and a built-in advisor chat aware of the strategies and underlying return data. ## What kinds of strategies it surfaces The planning domains vary by entity type. The AI surfaces only strategies the return's facts support, and stays silent on domains the data does not support. **Individuals (Form 1040):** retirement contributions and Roth conversions; investment gain and loss timing; equity compensation; charitable giving vehicles; real estate depreciation and deferral; business-owner deductions; health and medical accounts; education funding; income timing and bracket management; credits and incentives; family and dependent planning; Social Security claiming strategy; estate and gift considerations. **Business entities (Form 1120-S, 1065, 1120, 1041):** reasonable compensation; the QBI and pass-through deduction; basis and loss limitations; distributions management; retirement plan design; SALT workarounds; exit and succession planning; entity structure; R\&D credits. **Tax-exempt organizations (Form 990):** unrelated business income considerations and related compliance planning. ## Before you run AI Tax Planning * AI Tax Planning analyzes a finalized return, not a draft. Fully prepare the return before triggering a plan. * It does not run automatically after AI Review; trigger it separately from the client's page. * The richer the return (more schedules, K-1s, and supporting forms), the more domains the advisor can evaluate. ## Running AI Tax Planning **Walkthrough: run AI Tax Planning** 1. Open **AI Tax Planning** from the client hub. On the single-step **Upload the final 1040 return** screen, add the finalized return PDF via the drop zone or **Browse files**, then click **Start tax planning**. ![AI Tax Planning: upload the finalized return](https://storage.googleapis.com/filed-prod-public-assets/walkthroughs/ai-tax-planning/image1.webp) 2. AI Tax Planning analyzes the return and shows an in-progress loader ("Building your tax plan") while it finds planning strategies. This usually takes a few minutes and updates automatically when ready. ![The AI Tax Planning in-progress loader](https://storage.googleapis.com/filed-prod-public-assets/walkthroughs/ai-tax-planning/image2.webp) 3. When the run completes, the plan appears with the savings band and the strategy cards sorted by estimated savings (see [Reading the tax plan](#reading-the-tax-plan)). ## Reading the tax plan When the plan is ready, the page shows the client name, tax year, and strategy count, followed by: * **Savings band** - estimated savings broken down by horizon (This year, Multi-year, Lifetime, Event-driven), so the total opportunity is visible without combining figures from different time frames. * **Global narrative** - a summary of the key opportunities. * **Strategy cards** - sorted by estimated savings. Savings figures are computed from arithmetic on the return's actual numbers. The calculation behind each figure is verifiable and shown on the card in a dedicated Calculations section. ### Understanding savings horizons Each strategy is tagged with one of four savings horizons: * **This year** - A one-time or current-year benefit. For example, an HSA contribution made before the filing deadline. * **Multi-year** - A benefit that recurs over several years. For example, adjusting withholding or accelerating depreciation on a recurring asset. * **Lifetime** - A benefit that applies indefinitely. For example, optimizing Social Security claiming age or restructuring required minimum distributions. * **Event-driven** - A benefit triggered by a future transaction, such as a business sale, property disposition, or stock option exercise. Savings figures are never summed across horizons. A lifetime benefit and a current-year benefit represent different types of value and cannot be meaningfully added together. ## Working through strategy cards Each card shows the strategy title, estimated savings figure, savings horizon badge, and planning domain badge. If a dollar amount cannot be computed from the return, the savings field shows a dash with a tooltip explaining why. Expanding a card reveals: * **Applicability** - The specific facts from the return that make this strategy relevant to this client. * **Implementation plan** - An ordered list of steps to act on the strategy. * **Document references** - Source documents cited as evidence. Clicking a reference opens that document in the binder. * **Calculations** - The arithmetic behind the savings figure, shown as a verifiable formula. Appears only when a dollar estimate is present. * **Assumptions** - Any conditions or caveats the estimate depends on. ### Selecting and dismissing strategies Each strategy is in one of three states: * **Proposed** (default) - No action has been taken yet. * **Selected** - The firm intends to pursue this strategy with the client. * **Dismissed** - The strategy is not applicable or not being pursued for this client. Use the Select and Dismiss buttons to record the firm's decision; a dismissed strategy can be restored at any time. Selections and dismissals persist across re-runs by domain and strategy identity, so a previously selected or dismissed strategy retains its status when a new plan is generated. ## Using the advisor chat The plan includes a built-in advisor chat that knows the client's plan, the full set of strategies, and the underlying return data. Use it to ask follow-up questions, request client-facing summaries, explore edge cases, probe a strategy's assumptions, or ask how strategies interact. ## Re-running AI Tax Planning Re-run when a revised or amended return becomes available, or when planning for a new tax year. Each run is stored in the sidebar history. Selections and dismissals carry forward by domain and strategy identity, preserving the firm's working decisions. ## Which return types are supported AI Tax Planning supports: * **Form 1040** - individuals * **Form 1120-S** - S corporations * **Form 1065** - partnerships * **Form 1120** - C corporations * **Form 1041** - trusts and estates * **Form 990** - tax-exempt organizations The planning domains and strategy vocabulary differ by entity type; the AI selects the appropriate domain set automatically based on the client's configured return type. ## Pages to add later Planned future pages in this section: * **Understanding strategies** - Strategy card anatomy, savings calculations, and applicability evidence in depth. * **Presenting the plan with clients** - Workflow for using the plan during a client meeting. * **Planning habits** - How workspace and user-level planning overlays customize strategy generation. * **Planning by return type** - Domain coverage reference per entity type. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://help.filed.com/reference/products/ai-tax-planning.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.