Briefings and forecasts¶
A process forecast helps leaders discuss how much recorded work activity may occur in a future week. An executive briefing combines observed organization information, a forecast, assumptions and, when available, a recommendation. Use them to prepare questions and decisions about a process.
These features do not measure an employee's productivity, predict an individual's performance, establish a completion deadline or automatically change work. A recommendation still needs human judgment and the appropriate business action.
Availability¶
The backend implements aggregate forecasts, briefing schedules, previews and saved briefing occurrences. Final Screens contains the intended forecast conversation, a “More history needed” outcome, a preview and a produced briefing. The inspected frontend still disables HiveAI entry and has no standalone HiveAI application. These designed journeys are not evidence of deployment.
The guide explains the intended conversation and the current behavioral limits. Numeric forecast delivery through conversation has an unresolved implementation gap, so illustrated numbers are examples. Schedule creation, editing and disabling have backend support, but their complete screen journeys have not been verified.
Three things to distinguish¶
| Item | Purpose | What it changes |
|---|---|---|
| Process forecast | Estimate recorded activity in a future weekly period from historical activity | Reads evidence; does not alter the underlying work |
| Briefing preview | Inspect what the next scheduled briefing could contain using information available now | Does not create a saved occurrence, advance the schedule or run the scheduled recommendation generator |
| Produced briefing | Read the saved result for one due occurrence | Records the result of scheduled processing, including partial or failed outcomes |
A preview can differ from the produced briefing. More evidence may arrive, access or policy may change, and paid recommendation generation may fail even when the preview was available. A successful preview does not prove that scheduling or delivery will succeed.
HiveOrbit must be ready for current use¶
The aggregate forecast depends on HiveOrbit being enabled and ready for the workspace at the time of the request. Saved historical activity alone is not enough. Incomplete setup, a paused or scheduled operating period, changed prerequisites or temporarily unavailable access checks can prevent a new forecast from using that history.
Ask an authorized administrator to review HiveOrbit setup and current operation and recovery. More weeks of history will not, by themselves, resolve an access or setup restriction. An unavailable forecast is not a forecast of zero activity, and an existing saved briefing does not establish that a new forecast is currently available.
Who the feature serves¶
The direct forecast and briefing interfaces require advanced HiveAI use, permission to read organization metrics, and organization-head or delegated governance authority. Current tenant, licensing, policy and feature controls still apply. A management title or membership of a recipient list does not grant access.
The conversational forecast tool has a different authority path; its current implementation has not established parity with the direct interface. The guide describes the intended authorized executive use and does not promise access based on role names alone.
Briefing schedules require an active organization-head owner and active organization-head recipients when created or changed. Authorized briefing readers currently read across their organization; the recipient list is not a private readership boundary. Ordinary members and external participants should use the information available within their own authorized scope.
What a forecast means¶
The current calculation counts distinct recorded contribution events within a supported work domain and fits a straight-line trend to weekly buckets. It does not calculate duration, even if an internal process name mentions cycle time. It also does not count every live request, record or project in the platform.
At least eight complete weekly buckets and a minimum contributor cohort are required. The current source uses a minimum of five distinct contributors. Missing, stale or unsupported evidence can make the forecast unavailable. Empty weeks are represented as zero; this alone does not prove that every contributing service supplied complete evidence.
The point estimate and interval describe a future weekly count. A request about the 30-day point is not a request for a 30-day cumulative total. Forecast horizons are grouped into weeks, and the interval is model-based uncertainty rather than a guarantee of what will happen.
What a briefing contains¶
The observed information comes from the organization-intelligence dashboard. Its adoption definitions, unavailable quality measures, mixed freshness and credit-information limitations continue to apply inside a briefing.
A forecast section adds the modeled weekly estimate. Assumptions explain the method and evidence limits. An uncertainty section describes the available observations or the reason a forecast could not be produced. A recommendation, when present, suggests something to examine; it does not prove a root cause or authorize a change.
The overall “ready” or completed result does not establish that every nested measure is available or current. Read each section and its caveats. A partial briefing can contain useful facts while omitting a recommendation or withholding a forecast.
Scheduling and credits¶
Schedules run weekly or monthly at a fixed UTC hour. Monthly schedules use the last day of a shorter month when the selected day does not exist. They do not store a local time zone or automatically follow daylight-saving changes.
Scheduled narrative generation uses tenant prepaid credits and applicable owner-level budget scopes. A failed reservation prevents the model call. Provider or settlement problems can leave a fact-only result or require operational investigation. Previewing does not invoke that scheduled paid generation path, but an enclosing HiveAI conversation can still consume credits.
The backend saves briefing occurrences for authorized reading. This audit did not establish automatic email, messaging or push delivery to recipients. A configured recipient and a scheduled due time are not delivery receipts.
Retention and changes¶
Editing a schedule changes future settings and re-enables it. Disabling removes its next due time; it does not erase already produced briefings. There is no dedicated schedule-delete or regenerate-occurrence interface in the inspected surface.
User-data deletion handling retains shared schedule and briefing history with the affected participant identifiers replaced; affected schedules are disabled. This is not a promise to erase all produced content. Tenant-wide data export and deletion follow the platform lifecycle controls.
Continue with using forecasts and briefings, interpreting organization intelligence, and reviewing approval requests.
Understand held and settled briefing credit¶
The credit and budget guide explains why missing recommendations can coexist with held or settled credit, how UTC monthly limits apply, and how support should reconcile the original work before another paid attempt.