ObjectStackObjectStack

How to Request Features

A one-page template that turns business needs into the platform's vocabulary — hand it to your implementation team or an AI agent and get a running system

The fastest requirement conversations happen in the platform's own vocabulary. Answer these seven questions and what you hand over is not a requirements document — it is most of the configuration.

【Objects】   What data will you manage? What fields does each need?
             e.g. Opportunity = name, account (lookup), amount, stage
             (select: prospecting / qualified / proposal / negotiation /
             closed), expected close date, win probability,
             expected revenue (formula), product line items (master-detail)

【Views】    Who looks at it, and through which lens?
             e.g. reps → kanban by stage; managers → team table;
             executives → dashboard

【Rules】    What must be blocked at entry? Which stage moves are legal?
             e.g. discounts over 40% cannot save; negotiation → won/lost only

【Automation】When should the system act on its own?
             e.g. amount crosses 500k → notify the regional manager;
             30 days before contract expiry → create a renewal task

【Approvals】 Which actions need a human yes? Who? What counts as passed?
             e.g. discount > 30% → finance and sales director group
             sign-off; 24h overdue → escalate to VP

【Access】   Who sees which records — and which fields?
             e.g. regions isolated; cost price managers-only;
             deals > 1M auto-share to the VP

【Reports】  What numbers do leaders watch, and how often?
             e.g. funnel + monthly trend dashboard; Monday email digest

Two habits make this even faster:

  1. Point at HotCRM: "like HotCRM's opportunity approval, but with a second tier above 2M" — referencing a running example beats describing from scratch. Explore it via What Can It Do? or install it from the Marketplace.
  2. Hand the filled template to an AI agent: the platform is built so that agents author configuration from exactly this kind of description — see How AI Development Works.