Every modification command the AI assistant understands — shift, scale, redistribute, pin-and-fill, shape change, and more — with examples and syntax.
How the AI Assistant Processes Commands
When you type a command into the AI conversation panel in Modify Plan, the assistant interprets your intent, maps it to the KR’s plan structure (check-in dates, From/To range, metric type), and generates new plan values. The preview appears in the plan table before anything is saved.
The assistant understands natural language, so you don’t need to use exact syntax. But some phrasing patterns produce faster and more accurate results. This reference guide documents every command category with the clearest phrasing for each.
Tip: the AI command text is recorded in the audit trail. Clear, specific commands create better documentation. “Shift 15% from Feb into March due to dependency delay” is a better audit entry than “make Feb lower.”
1. Shift Commands
Move target allocation from one time period to another. The quarterly total remains unchanged — you’re redistributing, not reducing.
| Command | What It Does |
|---|---|
| “Shift 20% from January into February.” | Reduces January’s incremental targets by 20 percentage points total, adds the same amount to February periods. |
| “Shift 15% from Feb into March equally.” | Reduces February by 15pp, distributes equally across all March check-in periods. |
| “Move $500K from month 1 to month 3.” | Works with absolute values too. Reduces month 1 targets by $500K total, adds to month 3. |
| “Shift the first two weeks’ targets into weeks 5–8.” | Moves specific week ranges. Zeros out weeks 1–2, adds their combined total to weeks 5–8 equally. |
| “Shift everything after Feb 15 forward by two weeks.” | Displaces the plan timeline. Targets for Feb 16+ shift to start at Mar 1. Useful for gate dependency slips. |
Shift commands preserve the quarterly total. If you need to reduce the total, use a target change command instead.
2. Scale Commands
Multiply targets for specific periods by a factor. Useful for acceleration or deceleration without manually recalculating every cell.
| Command | What It Does |
|---|---|
| “Double the pace for the first four weeks.” | Multiplies incremental targets for weeks 1–4 by 2. Rebalances remaining weeks to keep the total unchanged. |
| “Reduce March targets by 30%.” | Multiplies each March incremental target by 0.7. The freed-up target is not redistributed — the total decreases. Combine with a shift to redistribute. |
| “Scale February up by 1.5x.” | Multiplies February incremental targets by 1.5. Rebalances other months proportionally. |
| “Halve the last two weeks.” | Multiplies weeks 12–13 incremental targets by 0.5. Useful for expected end-of-quarter slowdown. |
| “Triple the pace after the launch date (Mar 1).” | Multiplies all post-March-1 incremental targets by 3. The AI references specific dates when provided. |
Scale commands that increase one period automatically decrease others to preserve the total, unless you explicitly say “and increase the overall target.” If the AI rebalances when you didn’t want it to, follow up: “Don’t rebalance. Let the total increase.”
3. Redistribute Commands
Reallocate the remaining unachieved target across a specified period. Commonly used mid-quarter when early periods have actuals and the remaining plan needs adjustment.
| Command | What It Does |
|---|---|
| “Redistribute the remaining target equally across the last 5 weeks.” | Calculates: To target minus current actual. Divides equally across weeks 9–13. |
| “Redistribute remaining 40% evenly across March.” | Divides 40 percentage points equally across all March check-in periods. |
| “Redistribute what’s left with a linear ramp.” | Calculates remaining target. Creates a linearly increasing distribution across remaining periods (smallest first, largest last). |
| “Redistribute remaining target with heavier weighting in the final two weeks.” | Allocates more to weeks 12–13, less to earlier remaining weeks. The AI uses an approximately 60/40 split between final and earlier weeks. |
| “Redistribute from current actuals to target.” | Reads the current check-in value as the starting point. Builds a plan from that point to the To target across remaining periods. |
4. Pin-and-Fill Commands
Lock specific periods at their current values and let the AI handle the rest. One of the most powerful command types for combining manual precision with AI efficiency.
| Command | What It Does |
|---|---|
| “Keep January as-is. Distribute the rest evenly across Feb and March.” | Locks all January targets. Calculates remaining target (To minus January endpoint). Distributes equally across Feb and March periods. |
| “Keep everything through Feb 10. Ramp the rest to 100%.” | Locks all periods through Feb 10. Builds a linear ramp from the Feb 10 value to the To target across remaining periods. |
| “Keep actuals as-is. Fill the rest with an S-curve.” | Locks periods that already have check-in data. Generates an S-curve distribution for the remaining periods. |
| “Don’t touch weeks 1–6. Redistribute weeks 7–13 to reach the target.” | Explicit period locking by week number. Calculates shortfall from week 6 endpoint to target. Fills weeks 7–13. |
| “Keep February 3 and February 10 targets. Adjust everything else.” | Locks two specific dates. Recalculates all other periods to ensure the plan sums to the target while preserving those two values. |
Pin-and-fill is the ideal workflow for mid-quarter modifications. Your early periods reflect reality (pin them). Your late periods need to absorb the change (let the AI fill them). This produces plans that are grounded in actual data for the past and intelligently distributed for the future.
5. Shape Commands
Change the overall distribution curve without changing specific periods. Shape commands regenerate the entire plan distribution from From to To using the specified pattern.
| Command | What It Does |
|---|---|
| “Make it linear.” | Equal incremental targets across all periods. The simplest distribution. |
| “Make it an S-curve.” | Slow start (~15% of target in month 1), steep middle (~50% in month 2), gradual finish (~35% in month 3). |
| “Back-load into the last month.” | Minimal incremental targets in months 1–2 (~30% combined), heavy targets in month 3 (~70%). |
| “Front-load the first three weeks.” | Heavy incremental targets in weeks 1–3 (~45%), tapering through the rest of the quarter. |
| “Flat at 20% through February, then ramp to 100% in March.” | Plateau-and-ramp. All periods through February set to maintain 20%. March periods ramp linearly from 20% to 100%. |
| “Slow start, fast finish.” | Exponential-style curve. Very small increments early, accelerating sharply toward the end. |
| “Steady ramp with a plateau in week 6–8.” | Linear ramp with a flat period inserted. Useful when a known pause (testing phase, holiday, dependency wait) interrupts progress. |
Shape commands replace the entire distribution. If you’ve manually edited specific cells, a shape command will overwrite them. If you want to preserve manual edits, use pin-and-fill instead: “Keep [manual periods] as-is. Make the rest an S-curve.”
6. Set-Specific Commands
Set an exact value for one or more specific check-in periods. The most precise command type — useful when you know exactly what a particular date’s target should be.
| Command | What It Does |
|---|---|
| “Set the January 19 target to 25%.” | Changes the To (Target) value for the Jan 19 check-in to 25%. Recalculates the incremental target for that period and the next. |
| “Set all February targets to 10% incremental.” | Sets each February period’s incremental target to exactly 10%. The cumulative To values are recalculated accordingly. |
| “Set the first check-in to 0% and the last to 100%.” | Anchors the endpoints. Other periods remain unchanged unless you combine with a fill command. |
| “Set weeks 5–7 to 8% incremental each.” | Sets three specific weeks to the same value. Other periods unchanged. |
| “Set March 31 to 100% (the final target).” | Ensures the plan reaches exactly 100% on the last check-in date. Useful for confirming the plan endpoint. |
7. Target Change Commands
Modify the KR’s From (Baseline) or To (Target) value and redistribute the plan accordingly. These commands use the inline From/To editing capability within Modify Plan.
| Command | What It Does |
|---|---|
| “Change the target to 85%.” | Sets the To value to 85%. Redistributes all incremental targets proportionally to fit the new range. |
| “Reduce the target by 15 percentage points.” | Calculates new To = current To minus 15. Redistributes proportionally. |
| “Increase the target to $12M.” | Sets the To value to $12M. Proportional redistribution scales all periods up. |
| “Change the baseline to 5% (we’re not starting from zero).” | Sets the From value to 5%. The first period’s starting point shifts; all incremental targets recalculate. |
| “Reduce the target to 85% and back-load the reduction into March.” | Changes the target and specifies where the reduction lands. January and February stay close to original; March absorbs the full reduction. |
Target change commands modify the KR’s actual From/To values, not just the distribution. These changes are persisted to the KR entity on save, keeping the plan and KR synchronized.
8. Query and Explain Commands
Not all interactions with the AI are modifications. You can also ask questions about the current plan without changing anything.
| Command | What It Does |
|---|---|
| “Explain the current distribution.” | The AI describes the plan shape: monthly allocation percentages, where the steepest ramps are, and the overall pattern name (S-curve, linear, etc.). |
| “What percentage of the target is allocated to February?” | Calculates and reports the sum of February’s incremental targets as a percentage of the total target. |
| “Which week has the highest incremental target?” | Scans the plan and identifies the period with the largest single-period target. |
| “If I shift 10% from Feb to March, what does March look like?” | Simulates the shift and describes the result without applying it. Say “apply it” to confirm. |
| “Is the plan front-loaded or back-loaded?” | Analyzes the distribution and reports: “The plan is moderately back-loaded: 25% of the target is allocated to month 1, 33% to month 2, and 42% to month 3.” |
| “How much is left to distribute after current actuals?” | Calculates the gap between the current check-in value and the To target. Reports the remaining amount and how many periods are left. |
Query commands are free — they don’t modify the plan or create audit trail entries. Use them to understand the current state before deciding what to change. A common workflow: query first (“explain the distribution”), then modify (“shift 10% from Feb into March”).
9. Compound Commands
Chain multiple operations in a single instruction. The AI processes them sequentially, left to right.
| Command | What It Does |
|---|---|
| “Make it linear, then double the first two weeks.” | First generates a linear distribution, then scales weeks 1–2 by 2x and rebalances the rest. |
| “Reduce the target to 85% and make it an S-curve.” | Changes the To value, then generates an S-curve distribution within the new range. |
| “Keep January as-is, redistribute Feb and March equally, then set the last week to exactly 100%.” | Three operations: pin January, redistribute mid-periods, anchor the endpoint. |
| “Shift 10% from Feb to March, then scale March’s last two weeks by 1.5x.” | Shift first, then selectively scale within the shifted period. |
| “First back-load into March, then set Feb 17 to 40%.” | Generates a back-loaded distribution, then overrides one specific date with a precise value. |
Compound commands are executed strictly in order. Each operation sees the result of the previous one. If the sequence doesn’t produce the expected result, break it into separate commands and review the intermediate state after each step.
Command Writing Best Practices
The AI assistant is flexible, but these practices produce the fastest and most accurate results:
| Practice | Example |
|---|---|
| Use specific time periods, not vague references. | “Shift 15% from February into March” not “make the middle part smaller.” |
| Specify whether values are percentages or absolute. | “Reduce by 10 percentage points” vs. “reduce by 10%.” These mean different things: 10pp from 50% = 40%. 10% of 50% = 45%. |
| Name the shape when you want one. | “S-curve,” “linear,” “back-loaded” are recognized patterns. “Make it curvy” is not. |
| Reference dates when precision matters. | “Set the Feb 10 target to 35%” is unambiguous. “Set the second February check-in” works too. |
| Include the “why” for better audit trail. | “Shift 15% from Feb into March due to API delay” documents the rationale automatically. |
| Ask for explanation before complex changes. | “Explain the current distribution” before “make it an S-curve.” Understanding the baseline prevents surprises. |
| Use “undo” if a command doesn’t look right. | Type “undo” or “revert to previous.” The AI restores the plan table to its state before the last command. |
Quick Reference Card
Bookmark this section for fast reference during plan modifications:
| I Want To… | Command Pattern |
|---|---|
| Move targets between months | “Shift [X]% from [month] into [month].” |
| Speed up or slow down a period | “Scale [period] by [factor]x.” or “Double/halve [period].” |
| Spread remaining target evenly | “Redistribute remaining across [period].” |
| Lock some periods, fill the rest | “Keep [period] as-is. Distribute the rest [shape/equally].” |
| Change the overall curve shape | “Make it [linear/S-curve/back-loaded/front-loaded].” |
| Set one specific date’s target | “Set the [date] target to [value].” |
| Change the KR’s target | “Change the target to [value].” |
| Understand the current plan | “Explain the distribution.” or “How much is left to distribute?” |
| Chain multiple changes | “[Command 1], then [Command 2].” |
| Undo the last change | “Undo.” or “Revert to previous.” |
Say it. The plan does it.
Profit.co’s AI plan assistant understands natural language commands for every type of plan modification. Shift, scale, redistribute, reshape — all in a sentence. Start your free trial.