Orchestrator
The Orchestrator is Fload's AI brain — it coordinates all other agents, generates daily briefings, and decides what actions to take based on what's happening across your apps.
Think of it as an operations manager that never sleeps. It observes your metrics, reviews agent outputs, makes decisions, and dispatches work — multiple times a day.
How it works
The Orchestrator runs on a configurable schedule (default: every 6 hours) and follows a four-phase cycle:
1. Observe
Collects results from all recent agent runs (review replies, growth audits, ad performance, anomalies). Pulls the latest metrics from every connected data source.
2. Review
Analyzes everything together. Identifies patterns and issues that span multiple areas — for example, a spike in negative reviews combined with a drop in retention might signal a bug in your latest release.
3. Decide
Prioritizes what needs attention. Considers what worked (and what didn't) from past actions using its built-in memory system. Ranks actions by impact and urgency.
4. Dispatch
Creates pending actions for your approval — or auto-executes them if you've configured it that way. Queues specialized agent runs as needed (e.g., "run a growth audit" or "generate new ASO recommendations"). Produces the daily briefing.
Daily briefing
Every cycle produces a briefing that includes:
- Key metrics summary — what changed since the last briefing
- Agent activity — what each agent did and found
- Anomalies & alerts — anything unusual that needs attention
- Recommended actions — what the Orchestrator thinks you should do next
- Pending actions — actions waiting for your approval
Briefings are delivered via your preferred channel — email, Slack, or Discord.
Memory and learning
The Orchestrator remembers past decisions and their outcomes:
- After dispatching an action, it schedules a check-back to evaluate results
- Successful actions inform future decisions ("pausing that campaign last time improved ROAS")
- Unsuccessful actions are flagged so the same mistake isn't repeated
- Over time, the Orchestrator gets better at understanding your specific apps
Event-triggered mini-cycles
Beyond scheduled runs, the Orchestrator reacts to urgent events:
- Anomaly detected — a metric spiked or dropped significantly
- Negative review surge — multiple 1-star reviews in a short period
- Ad spend anomaly — unexpected budget burn or campaign failure
- Submission rejection — your app was rejected by App Review
Mini-cycles follow the same observe-review-decide-dispatch flow but focus on the triggering event.
Draft vs Agent mode
The Orchestrator operates in two modes:
Draft mode
The default. The Orchestrator generates recommendations and queues them as pending actions for your review. Nothing changes until you approve it. This is ideal when you're getting started or managing high-stakes apps.
Agent mode
The Orchestrator acts on its own — executing decisions without waiting for approval. It still logs everything and respects safety guardrails (like budget caps), but it moves faster. Best for teams that trust the system and want hands-off operations.
You can switch between modes at any time in Settings → Orchestrator.
Agent write actions
When the Orchestrator decides something needs to happen, it coordinates write actions across agents:
- Review Agent — post a reply to a review
- ASO Agent — update promotional text or descriptions
- Ads Agent — pause a campaign, adjust a bid, or reallocate budget
All write actions are tracked in a unified activity log. In draft mode, they appear as pending actions. In agent mode, they execute and you're notified after the fact.
What it can do
The Orchestrator's available actions depend on your connected platforms:
| Connected platform | What the Orchestrator can coordinate |
|---|---|
| App Store Connect | Review replies, ASO updates, submission checks |
| Google Play | Review replies, description updates |
| Apple Search Ads | Pause/resume campaigns, adjust bids and budgets |
| Google Ads | Pause/resume campaigns, adjust bids and budgets |
| Meta Ads | Pause/resume campaigns, adjust bids and budgets |
| TikTok Ads | Pause/resume campaigns, adjust bids and budgets |
| Stripe / RevenueCat | Revenue monitoring, tier alerts |
| Slack / Discord | Notifications, briefing delivery |
The more platforms you connect, the more the Orchestrator can do.
Configuration
Run frequency
- Default: every 6 hours (4 times daily)
- The schedule is configurable
- Mini-cycles run on-demand regardless of schedule
Approval mode
- Draft mode — all actions require your manual approval before executing
- Agent mode — all actions execute automatically (requires explicit opt-in)
Notifications
- Choose which events trigger notifications
- Pick your delivery channel (email, Slack, Discord)
What's free vs. paid
| Feature | Free | Pro ($50/month) |
|---|---|---|
| View daily briefings and agent activity | Yes | Yes |
| View pending actions queue | Yes | Yes |
| View activity log and history | Yes | Yes |
| AI-powered Orchestrator cycles | — | Yes |
| Automated agent dispatch | — | Yes |
| Event-triggered mini-cycles | — | Yes |
See Pricing for the full breakdown.