Atlas · Control Tower Agent
At a glance
- Live at: Multiple LSPs (control tower deployments)
- Channels: API-driven (autonomous monitoring), voice (outbound calls to drivers/customers)
- Languages: EN, HI (locale-dependent)
- Owner: AI pod
- Key stat: Processes incident tickets automatically — vehicle stationary, suspicious delivery patterns, delayed consignments
The problem it solves
A logistics control tower monitors thousands of shipments in real time. When something goes wrong — a vehicle stops moving, a delivery takes suspiciously long, a consignment misses its SLA — someone needs to investigate and act. At scale, human operators can’t triage fast enough. Incidents pile up, response times slip, and preventable losses grow.
Atlas automates the triage-investigate-act loop. It picks up incident tickets, analyzes the situation, contacts the relevant parties (drivers, dispatchers, customers), and takes action — all without a human in the loop unless the situation is genuinely complex.
How it works
Step by step
- Ticket intake — Incident ticket arrives from LIA (Query Builder) or ProjectX monitoring.
- Supervisor analysis — The
control_tower_supervisoragent reads the ticket, identifies the incident type, and decides on the action plan. - Outbound action — Routes to the appropriate voice agent:
- Driver agent — calls drivers about trips (vehicle stationary, delays)
- Dispatch agent — coordinates with dispatch for reallocation
- Customer agent — verifies delivery, confirms receipt
- Ticket update — Posts findings as comments on the LIA ticket, updates status.
- Follow-up — If unresolved, schedules a follow-up call/action using the
schedule_followuptool. - Close — Marks the ticket as resolved when the issue is cleared.
Agent templates in the codebase
| Template | ID | Type | Model | Role |
|---|---|---|---|---|
control_tower_supervisor | 2005 | sub_agent | Gemini 2.5 Flash | Processes incident tickets, decides action plan |
control_tower_driver | 3002 | voice_agent | Gemini 2.5 Flash | Calls drivers about trips (via ElevenLabs) |
control_tower_dispatch | 3003 | voice_agent | Gemini 2.5 Flash | Dispatch coordination calls |
control_tower_customer | 3004 | voice_agent | Gemini 2.5 Flash | Customer delivery update calls |
Tools used
| Tool | What it does |
|---|---|
fetch_information | Fetch trip/consignment details from ProjectX |
lia_post_comment | Post findings as comments on the incident ticket |
lia_update_ticket_status | Update ticket status (acknowledged → actionTaken → closed) |
lia_update_ticket_custom_field | Update custom fields with agent findings |
schedule_followup | Schedule follow-up calls/messages |
object_action | Generic ProjectX object actions (fetch, update) |
Atlas vs Maya
| Atlas | Maya | |
|---|---|---|
| Scope | General-purpose control tower — multiple incident types | BDO-specific — card delivery monitoring |
| Deployment | Multiple LSPs | BDO Unibank |
| Incident types | Vehicle stationary, suspicious delivery, delays, address issues | Card delivery SLA, delivery confirmation, exception handling |
| Customization | Generic templates, configured per deployment | Deeply customized for BDO’s card delivery workflow |
Atlas is the productized version of the control tower pattern. Maya was the first deployment (BDO-specific); Atlas generalizes it for any LSP.
Tech stack
| Layer | What’s used |
|---|---|
| Models | Gemini 2.5 Flash (cost-optimized for high-volume incident processing) |
| Memory | Short-term: conversation + ticket context. Long-term: incident resolution history. |
| Tools (MCP) | ProjectX fetch, LIA ticket management, follow-up scheduling. |
| Guardrails | Content safety on voice calls. PII masking in ticket comments. |
| Evals | Scenario-based per incident type; escalation rate tracking; follow-up effectiveness. |
Sources
- agent-platform repo:
data/agent/agents.json— control_tower templates - Atlas/Maya video walkthrough
- Agent video playlist: YouTube
- See Maya (BDO) for the BDO-specific control tower deployment
Changelog
- 26 May 2026: Full content from GitHub repo analysis (control tower templates) and video library.