Your AI team catches what you miss, stores what matters, and gets sharper with every session.
The five-stage vigilance cycle
Expand any step. Each one has a plain-English explanation for context, and a technical breakdown for the engineers in the room.
Think of this as your AI team running a morning sweep. They check the places work usually slips: CRM, email, tasks, calendar, and your knowledge graph. Then they produce a structured briefing: what is stalling, what went quiet, what follow-up is missing, and what risk is emerging. You did not have to ask.
The Radar aggregates a tenant-scoped snapshot from CRM, email, tasks, calendar, and Memoria. A tier-aware model resolver picks the allowed model rung for that tenant and intent, then returns a structured JSON envelope with five sections: top_opportunities, stale_opportunities, next_followups, content_gaps, risks_and_blockers. The runner is read-only and can be triggered manually or on a schedule.
The results appear as a briefing in your AEGIS dashboard. Each catch has a title, a short summary, and a suggested next action. This is an information layer: AEGIS is showing you a picture, not making decisions. Nothing has changed in your CRM, inbox, or other systems.
The structured envelope is stored as a skill_runs record. Setup gaps such as disconnected integrations or empty sources are filtered before display so the dashboard only shows business signals.
For each catch, you decide whether it belongs in your business's permanent knowledge. Click Accept and it goes into Memoria. Ignore it and it does not. This gate is non-negotiable: AEGIS will not decide on its own what is strategically important. That judgment stays with you.
Clicking Accept calls the approval flow for the Radar item. Items are blocked from acceptance if they are setup gaps, missing evidence, or lack a valid section classification. Only items that pass validation proceed to Memoria writeback.
What you approved does not disappear when you close the dashboard. It becomes part of your business's living knowledge graph: Memoria. Every agent can read from it, so the context you build now is available the next time the team briefs you.
Accepted items are written to kb_nodes with trust_level=human_approved. Nightly jobs apply confidence decay and crystallization so the graph stays current and useful.
Every time you rate a skill output with 👍 or 👎, you're teaching the system. AEGIS tracks which procedures work and which do not. Overnight, it rewrites the weaker ones and notifies you when a better version is active.
Every skill run writes a skill_runs row with a rating field. The nightly auto-learn job looks for underperformers, generates an improved prompt version from recent feedback, saves it to prompt_history, and applies it immediately.
AEGIS runs two categories of background jobs. One you configure; one runs automatically for all tenants.
System jobs — always running
| Job | Runs at | What it does |
|---|---|---|
| job_syncer | Every 60 s | Picks up new tenant schedules from the database and registers them with APScheduler |
| job_nightly_memoria | 02:00 AM | Confidence decay on all kb_nodes + crystallization tier promotion |
| job_auto_learn | 03:00 AM | Scans skill_runs feedback, generates improved prompts for underperformers |
| job_nightly_pulse | 02:00 AM | Business Pulse CRM refresh — pipeline health metrics recalculated |
Tenant schedules — you configure in Workflows
Set a cron expression (e.g. 0 8 * * 1-5 = weekdays 8am), choose a task prompt and an agent. Within 60 seconds, the job_syncer registers it. When it fires: APScheduler → execute_task() → openclaw → LiteLLM → agent response → result stored → Telegram notification (if enabled). If the LLM is temporarily unavailable (quota, timeout), a neutral delay notice is stored and the error is never surfaced to your Telegram. The schedule runs again automatically at its next interval.
| Raw AI (one-shot) | AEGIS Loop |
|---|---|
| Forgets everything after the session | Memoria graph grows with every approved catch |
| You bring all context manually | CRM, email, tasks auto-injected before every LLM call |
| Same outputs every time | Skills improve from your feedback — every run is logged |
| You have to remember to ask | Radar surfaces what you didn't ask about |
| No audit trail | Every run, approval, prompt version, and decision logged |
| Platform owns your data | Schema-per-tenant — your data never touches another account |
auto_learn_mode setting enabled, available from Pro (€49/mo) upward. Model selection is governed by tier and intent.8 agents. Persistent memory. Skills that improve every run. Free to start.