Incremental AI Formula Updates

Feature Request: Incremental AI Formula Updates

Summary:
Enable the AI to perform incremental, repetitive update passes that identify and generate only new or missing formulas, without reprocessing or modifying formulas that have already been created and confirmed. This allows users to safely rerun the AI as content evolves while preserving existing work.

Problem Statement:
Currently, when users rerun the AI to generate or refine formulas, the AI may reprocess existing formulas, potentially overwriting or altering logic that has already been validated, tested, or approved. This behavior makes users hesitant to re-invoke AI assistance as requirements change, leading to manual workarounds, duplicated effort, and risk of regressions in previously correct formulas.

Proposed Solution:
Adjust the AI behavior to support an “incremental update” mode for formulas that:

• Scans the current artifact (e.g., BOE, worksheet, or configuration) and detects which items already have formulas.

• Compares the latest context (new line items, changed structure, or additional data) to the existing formulas.

• Generates formulas only for new items or clearly missing formula fields, leaving existing formulas untouched by default.

• Provides an option or toggle (e.g., “Only add missing formulas”) when invoking AI so users can choose incremental vs. full regeneration.

• Highlights newly added formulas and, optionally, flags any potential inconsistencies with existing formulas without automatically changing them.

• Logs which items were skipped because they already had formulas, for traceability.

Benefits:

• Reduces the risk of unintentionally overwriting validated formulas during subsequent AI runs.

• Saves time by allowing users to repeatedly use AI as the model evolves, without manually protecting existing formulas.

• Encourages broader AI adoption by making reruns safer and more predictable.

• Improves data quality and consistency through incremental, low-risk updates instead of full regenerations.

• Streamlines workflows when large models or BOEs are built iteratively over time.