Migrating From ASI PMS to an Agentic Hotel AI: A Complete Playbook

TL;DR

Migrating from ASI PMS to CallSphere's agentic hotel platform takes 3–10 days depending on property size. The playbook below covers data export, rate plan import, policy ingestion, number porting, and go-live QA.

Phase 1: Inventory Your ASI Data (Day 1)

Pull the following exports from ASI PMS:

flowchart TD
    START(["Migrating From ASI PMS to an Agentic Hotel<br/>AI: A Complete Playbook"])
    S0["TL;DR"]
    START --> S0
    S1["Phase 1: Inventory Your ASI Data<br/>(Day 1)"]
    S0 --> S1
    S2["Phase 2: Map Data to CallSphere<br/>(Days 2–3)"]
    S1 --> S2
    S3["Phase 3: Ingest Policies for the<br/>AI Agents (Days 3–4)"]
    S2 --> S3
    S4["Phase 4: Provision Voice Numbers<br/>and Test (Days 4–6)"]
    S3 --> S4
    S5["Phase 5: Parallel Run (Days 6–8)"]
    S4 --> S5
    S6["Phase 6: Cutover (Day 8–10)"]
    S5 --> S6
    S7["Common Migration Gotchas"]
    S6 --> S7
    DONE(["Key Takeaways"])
    S7 --> DONE
    style START fill:#4f46e5,stroke:#4338ca,color:#fff
    style DONE fill:#059669,stroke:#047857,color:#fff

• Room type catalog (standard, deluxe, suite, connecting)
• Rate plans (BAR, weekend, corporate, government, AAA)
• Cancellation + deposit policies per rate plan
• Historical reservations (last 12 months)
• Guest profile database (loyalty flags)
• OTA channel mappings (Booking.com, Expedia, Agoda)

ASI exposes most of this via its REST API. For older deployments, use the CSV export from the admin console.

Phase 2: Map Data to CallSphere (Days 2–3)

CallSphere's hotel data model uses these core tables:

• rooms (room_number, type, floor, amenities)
• rate_plans (code, name, base_rate, restrictions)
• policies (cancellation_hours, deposit_percent, pet_policy)
• guests (profile, loyalty_tier, preferred_language)
• reservations (check_in, check_out, rate_plan, status)

Mapping takes a few hours per property — CallSphere ships an import wizard that handles ASI's common schema.

Phase 3: Ingest Policies for the AI Agents (Days 3–4)

The Concierge, Reservation, Check-In, and Check-Out agents need to know your cancellation rules, pet policy, early check-in fees, late check-out policy, rollaway rates, and parking rates. Upload these as structured policy documents — the agents use RAG to cite them accurately.

Phase 4: Provision Voice Numbers and Test (Days 4–6)

CallSphere provisions a Twilio number in your area code (or ports your existing number). Run 20 test calls covering:

flowchart TD
    HUB(("Migrating From ASI PMS<br/>t"))
    HUB --> L0["TL;DR"]
    style L0 fill:#e0e7ff,stroke:#6366f1,color:#1e293b
    HUB --> L1["Phase 1: Inventory Your ASI<br/>Data (Day 1)"]
    style L1 fill:#e0e7ff,stroke:#6366f1,color:#1e293b
    HUB --> L2["Phase 2: Map Data to<br/>CallSphere (Days 2–3)"]
    style L2 fill:#e0e7ff,stroke:#6366f1,color:#1e293b
    HUB --> L3["Phase 3: Ingest Policies for<br/>the AI Agents (Days 3–4)"]
    style L3 fill:#e0e7ff,stroke:#6366f1,color:#1e293b
    HUB --> L4["Phase 4: Provision Voice<br/>Numbers and Test (Days 4–6)"]
    style L4 fill:#e0e7ff,stroke:#6366f1,color:#1e293b
    HUB --> L5["Phase 5: Parallel Run (Days<br/>6–8)"]
    style L5 fill:#e0e7ff,stroke:#6366f1,color:#1e293b
    HUB --> L6["Phase 6: Cutover (Day 8–10)"]
    style L6 fill:#e0e7ff,stroke:#6366f1,color:#1e293b
    style HUB fill:#4f46e5,stroke:#4338ca,color:#fff

• Basic reservation booking
• Pet policy question
• Cancellation request
• Group sales inquiry
• After-hours emergency
• Spanish language handling
• Mandarin language handling

Any agent-level hallucinations or handoff gaps get flagged in the Agents Console and corrected before go-live.

Phase 5: Parallel Run (Days 6–8)

Run CallSphere in parallel with ASI for 2–3 days. Forward 50% of inbound calls to CallSphere, keep 50% on the legacy system. Compare:

• Answer rate (target: 100%)
• Booking conversion
• Average handle time
• Guest satisfaction
• Direct booking vs OTA ratio

Phase 6: Cutover (Day 8–10)

When parallel-run metrics exceed baseline, cut 100% of inbound calls to CallSphere. Keep ASI running as the PMS system of record if you want to preserve it, or decommission it for a full replacement on boutique properties.

flowchart LR
    S0["Phase 1: Inventory Your ASI<br/>Data (Day 1)"]
    S1["Phase 2: Map Data to<br/>CallSphere (Days 2–3)"]
    S0 --> S1
    S2["Phase 3: Ingest Policies for<br/>the AI Agents (Days 3–4)"]
    S1 --> S2
    S3["Phase 4: Provision Voice<br/>Numbers and Test (Days 4–6)"]
    S2 --> S3
    S4["Phase 5: Parallel Run (Days<br/>6–8)"]
    S3 --> S4
    S5["Phase 6: Cutover (Day 8–10)"]
    S4 --> S5
    DONE(["Done"])
    S5 --> DONE
    style S0 fill:#4f46e5,stroke:#4338ca,color:#fff
    style DONE fill:#059669,stroke:#047857,color:#fff

Common Migration Gotchas

1. Rate plan codes drift — ASI's internal codes don't always match OTA codes. Normalize before import.
2. Cancellation policies are written in prose — convert them to structured rules the agent can enforce.
3. Historical guest preferences — loyalty, language, room preferences need to carry over for VIP recognition to work.
4. Night audit reports — keep running the ASI report for a few weeks until accounting signs off.

FAQ

Q: Will I lose historical data? A: No. Historical reservations, guest profiles, and folios are imported during Phase 2.

Q: Can I run both systems long-term? A: Yes, and most operators do. ASI remains the PMS, CallSphere runs the voice/chat layer.

Q: What if something breaks after cutover? A: CallSphere's rollback switch reverts all inbound routing to ASI in under 30 seconds.

---

Related: ASI PMS comparison | Hotel industry page

#HotelMigration #ASIPMS #CallSphere #HotelAI