Reconcile sales forecast categories with AI before the review

Record the mapping and reporting basis

A sales stage and a forecast category serve related but different purposes. Before comparing deals, collect the administrator-approved category mapping, pipeline definitions, forecast period, amount basis and an authorized opportunity snapshot. Keep a version and extraction time so a later setting change does not make the review impossible to reproduce.

HubSpot supports mapping stages to forecast categories and manually changing a deal’s category. Its forecast configuration also distinguishes weighted and total amounts. A category that differs from the usual mapping may therefore be intentional. Ask AI to identify the difference and request the owner’s evidence, rather than assuming that the system or salesperson is wrong.

Build an exception table with explanations

Use deal ID, pipeline, stage, mapped category, current category, documented override, amount basis and review question. Preserve missing mappings and categories that your organization has customized. Do not replace your actual definitions with generic labels the model remembers from another CRM. A tool’s category name is not a statistically calibrated probability.

In a fictional pipeline, a proposal-stage deal normally maps to Best case but appears in Commit after a manager-approved override. The review should retain the approval and ask whether its supporting condition remains current. Another deal with no override evidence belongs in an unresolved queue. Neither case authorizes AI to change the category or create a new revenue prediction.

Resolve definitions before comparing totals

Have sales operations and the account owner review exceptions. If the mapping changed, distinguish the configuration effect from a real change in deal evidence. Save the reviewed mapping and decisions alongside the forecast snapshot. Keep rep judgment visible where your forecasting process permits it instead of disguising judgment as a generated fact.

HubSpot requires administrative permissions for forecast configuration, and changing the forecast period can reset forecast goals and submissions. This preparation task should not modify those settings. Compare a few fictional exceptions first and verify the calculations separately in a spreadsheet or report. The useful outcome is an understandable forecast basis, not a claim that reconciliation improves accuracy or validates every projected close date.

Review checklist

• Preserve mapping version, snapshot and forecast period.
• Record total versus weighted amount basis.
• Keep documented overrides distinct from missing evidence.
• Do not translate category labels into invented probabilities.
• Leave setting and record changes to authorized owners.

Source context and limits

HubSpot documents stage mapping, manual categories, amount choices and configuration permissions. The exception review is our proposed workflow, not a tested forecasting model or a forecast-accuracy claim.