# 02 — Competitor AI and B2B Distribution Matrix

**Step:** 3  
**Prepared:** 2026-07-26  
**Subject:** SatuSatu only. TipTip ticketing is context only, not an opportunity target.  
**Evidence base:** copied from `research/raw/teardown-klook-gyg.md`, `research/raw/teardown-viator-kkday-headout.md`, and `research/raw/teardown-connectivity-indonesia.md`. Full URLs are in `99-sources.md`.

## Executive Read

The category is not yet agentic at the transaction layer. Klook, Tripadvisor/Viator, and GetYourGuide have shipped GenAI in discovery, content, supplier onboarding, review summarisation, and internal productivity, but their B2B APIs remain conventional REST/catalog/availability/booking surfaces. The most transferable AI pattern for SatuSatu is not "AI books trips"; it is **AI compresses supplier/catalog operations and makes exclusive Pool B inventory usable at scale**.

## AI Maturity Summary

Scale: 0 absent, 1 rules, 2 classical ML, 3 GenAI-assisted, 4 agentic.

| Company | Strongest shipped AI evidence | Weakest area | Overall maturity | Strategic thesis |
|---|---|---:|---:|---|
| **Klook** | GenAI localisation cut production time >80%; K.AI trip planner; review summaries for merchants; AI shopping agent in closed beta | Public outbound distributor API absent; pricing/fraud AI unverified | **3** | Own discovery and conversion, then route booking back to Klook-controlled checkout |
| **GetYourGuide** | Supplier GenAI product wizard auto-completes 8 of 16 steps; hybrid semantic search; transformer ranking; AI image optimisation | Itinerary, pricing, payments/fraud, CS | **3** | Use AI to improve marketplace liquidity: better listings, search, ranking, and supplier productivity |
| **Viator / Tripadvisor** | AI-native MVP scaled to half English web traffic; Viator app in ChatGPT; AI-enabled prebooking chat; AI-assisted sign-up doubled conversion, segment attribution unclear | B2B API, partner onboarding, pricing, payments/fraud | **3** | Escape paid CAC by moving into AI discovery channels and direct/app repeat usage |
| **KKday / rezio** | No material Tier-A AI artifact found in rezio API/site | AI generally absent; API is programmable but not intelligent | **1** | Productise operator backend to win system-of-record data and feed KKday distribution |
| **Headout** | AI-powered support and smarter pricing claimed; Dabble acqui-hire for CV/spatial computing | No named shipped AI feature; public API docs stale since 2024-07-29 | **1–2** | Cautionary case: AI narrative and API publication do not equal maintained product |
| **GlobalTix** | Funding release announced AI/predictive analytics; Basic Chatbot covers 5 questions | Partner API exposes no AI, ranking, recommendation, or forecasting endpoints | **1** | Competent connectivity + reseller marketplace, not an AI partner |

## Value-Chain Matrix

| Stage | Klook | GetYourGuide | Viator / Tripadvisor | KKday / rezio | Headout | GlobalTix |
|---|---:|---:|---:|---:|---:|---:|
| Supply sourcing | 1 | 2 | 1 | 1 | 1 | 1 |
| Onboarding & catalog production | 3 | **3** | 3 | 1 | 1 | 1 |
| Distribution-in / feed ingestion | 1 | 1 | 1 | 1 | 1 | 1 |
| Catalog scale ops | 3 | **3** | 2 | 1 | 2 | 1 |
| Inventory & availability | 1 | 1 | 1 | 1 | 1 | 1 |
| Pricing & packaging | 1 | 1 | 1 | 1 | 2 | 1 |
| B2B rate & contract management | 1 | 1 | 1 | 1 | 1 | 1 |
| Partner acquisition & onboarding | 1 | 1 | 0–1 | 1 | 1 | 1 |
| Partner dashboard | 3 | 1 | 1 | 1 | 1 | 1 |
| API & connectivity | 0 | 0 | 1 | 1 | 0–1 | 1 |
| Demand generation | 3 | 2 | 3 | 1 | 1 | 1 |
| Discovery & conversion | **3 / beta 4** | 3 | 3 | 1 | 1 | 1 |
| Pre-trip planning & itinerary | 3 | 0 | 3 | 0 | 1 | 0 |
| In-destination execution | 1 | 0–1 | 3 | 0 | 1 | 0 |
| Payments, FX & fraud | 0–1 | 0 | 1 | 1 | 1 | 1 |
| Post-trip | 2 | 1 | 1 | 1 | 1 | 1 |
| Customer service | 3 | 1 | 3 for TheFork, Viator unquantified | 0–1 | 2–3 claimed | 1 |
| Supplier operations | 3 | 3 | 1 | 1 | 1 | 1 |
| Trust & safety | 1 | 1 | 2 | 1 | 1 | 1 |
| Internal ops | 3 | 3 | 3 | 1 | 1 | 1 |

## B2B Distribution Teardown

| Company | Outbound partner / reseller API | What it exposes | Onboarding | Commercial model | AI in B2B surface | Implication for SatuSatu |
|---|---|---|---|---|---|---|
| **Klook** | **No public outbound API found.** Public GitBook is merchant/supply inbound | Merchant content, availability, pricing, holds, price locks | Sales-gated | Undisclosed; Klook controls checkout surfaces | Merchant review feedback loop | Do not reason from Klook merchant API as if it were a reseller API. Klook is more threat than pipe, especially with Klook Bali Pass |
| **GetYourGuide** | **Yes.** Public Partner API OpenAPI spec, active repo | Search, catalog, categories, supplier, reviews, availability, price breakdown, cart, booking, cancellation | Commercially sales-gated, technically self-serve docs and sandbox | Commission or MasterBill/invoiced path, terms gated | None exposed | Copy the posture: own catalog record and spec, let others integrate. Do not copy the build scale |
| **Viator** | **Yes.** Partner API v2.0, 33 endpoints, active docs | Product ingestion, search, recommendations, availability, booking, hold, cancellation, amendments, payments | Human qualification, certification, live booking/cancellation test | Affiliate or merchant; merchant invoiced by Viator, IDR not merchant invoicing currency | None exposed | Sets the industry bar for D3. Certification complexity is much higher than SatuSatu can staff now |
| **rezio** | Not marketplace outbound; operator backend API | Product creation, prices, sessions, seats, orders, redeem list, webhooks | Self-serve plan upgrade | SaaS at roughly US$50–217/month, no commission disclosed | None | If SatuSatu builds supplier tools, the value is Pool B system-of-record data, not SaaS revenue |
| **Headout** | Public API docs stale since 2024-07-29 | Unreliable as current evidence | Unclear | Unclear | No maintained B2B AI evidence | Warning: publishing an API is cheap; maintaining it is the product |
| **GlobalTix** | **Yes.** Partner API 3.0 public spec | Catalog, changes, new products, availability, reserve/confirm/release, cancellation, ticket PDFs, webhooks | Credentials via account manager; production after test booking/cancel | Prepaid credit, nett price, minimum selling price floor, 1.5% FX markup on non-SGD | None | Keep as Pool A utility. It is also a licensed travel agent and competes with D2/D3 |
| **Prioticket / ETG** | Yes, public Distributor API and OCTO-native surfaces | Products, availability, reservations/cart, orders, payments, contacts, notifications | Public docs, commercial gate | Undisclosed | None evidenced | Best second-aggregator candidate if Indonesia/Bali SKU depth checks out |
| **Bókun** | Yes, public docs and OCTO support | Availability, pricing, booking, amend, cancel; channel-manager features | Public docs, commercial gate | Supplier SaaS / marketplace terms | None evidenced | Technically strong but owned by Tripadvisor, so telemetry exposure must be accepted explicitly |

## Shipped vs Announced vs Speculation

| Claim | Status | Note |
|---|---|---|
| Klook GenAI content/localisation production time >80% reduction | **SHIPPED / self-reported** | Klook + Google Cloud official source |
| Klook AI shopping agent completes bookings and post-sale service | **BETA** | Select users per Skift; not GA |
| GYG GenAI supplier wizard | **SHIPPED** | Full rollout after failed first experiment and UX correction |
| GYG semantic/hybrid search and transformer ranking | **SHIPPED** | Engineering blog evidence |
| Viator app in ChatGPT | **SHIPPED** | Discovery handoff, not autonomous booking |
| Tripadvisor AI-native MVP on half English web traffic | **SHIPPED / self-reported** | Q1/Q4 earnings call |
| TheFork 40% B2C support handled by AI | **SHIPPED / self-reported** | Do not transfer the number to Viator |
| GlobalTix AI/predictive analytics | **ANNOUNCED** | No product endpoint or follow-up evidence |
| Headout Dabble-derived product | **UNVERIFIED** | Acqui-hire, no named product after announcement |

## Direct Implications for SatuSatu

1. **Do not build an agentic booking layer now.** The market leaders do not expose one in B2B, and OpenAI retreated from Instant Checkout in March 2026.
2. **Prioritise catalog and supplier operations.** GYG's supplier wizard and Klook's content/localisation workflow are the clearest transferable AI patterns.
3. **Treat D3 as proof of exclusive Pool B supply, not a platform business.** Viator and GlobalTix already own the generic API/reseller pattern at scales SatuSatu cannot match.
4. **Use GlobalTix defensively.** It solves coverage, but its resale price floor, FX markup, and agent marketplace make it structurally unsuitable as the B2B profit engine.
5. **If adding a second aggregator, shortlist Prioticket/ETG first.** Its public Distributor API and OCTO posture fit the no-dedicated-engineering constraint, but only if Bali SKU depth is real.