Our Portland, Maine org has a long tail of fields nobody can explain. Which ones get retired, and on what evidence?
We check each field's fill rate, when it was last updated and whether any report, automation, integration or page layout references it. Fields that are empty, unreferenced and unowned become candidates for removal, which we confirm with the teams affected. Before deleting anything, we hide candidates from layouts for a trial period and archive their data, so a field someone truly needs can be restored without loss.
Finance and sales in our Southern Maine office quote different revenue numbers. Can optimization reconcile them?
Usually, once the definitions are settled. We sit down with finance and sales leadership to agree what counts as booked, won or recognized, and when. Then we trace where Salesforce data diverges, whether stage misuse, missing close dates or duplicate opportunities, fix the data and adjust validation so it stays clean. The rebuilt dashboards use those agreed definitions and are reconciled against finance reports before leaders rely on them.
Would tidying the org first make an Agentforce pilot more meaningful for our service team?
Generally yes. AI features draw on the same records, fields and knowledge articles your people use, so clutter and duplicates make their answers worse. An optimization pass cleans the objects an agent will read, trims outdated knowledge content and clarifies case categories. For Southern Maine service teams, that groundwork makes a first Agentforce pilot far easier to judge fairly, because poor results will reflect the tool rather than messy data.