AI-Ready Travel Infrastructure

The data layer travel has been missing for 65 years.

TripForge gives travel companies one authoritative record of every booking - every document, every version, updated at every change event. That record is what AI requires to operate in travel. It has not existed until now.

Request early access See how it works

Every system holds a fragment. No system holds the whole.

Every functional capability in travel was automated independently, by different vendors, using different data models. The result: the same booking exists in five places simultaneously, each with its own version of the truth.

Reservations
Accounting
T&Cs
CRM
Operations
AI sees: nothing coherent  —  90% of enterprise AI projects fail. In travel, the reason is structural.
The solution

TripForge: one unified container per booking.

TripForge intercepts documents at the event layer - at creation or change - and builds one authoritative record per booking. It starts as a parallel layer alongside existing systems. Over time, it takes on more of the data lifecycle: writing updates back to legacy systems in their native formats, then becoming the record of origin. The legacy estate winds down on the customer's timeline, not ours.

Stored as received. No normalization required.

Documents are stored exactly as received - EDIFACT, XML, plain text, PDF. No normalization required before storage. Transformation happens when it is needed, not before.

Every version retained. Nothing overwritten.

Documents are never overwritten. Every version, every change, every communication is retained - a complete audit record of the booking's entire lifecycle.

No ETLs. No data lakes. No export pipelines.

Every analytics and AI operation executes inside the instance. There are no pipelines exporting data to external stores, no ETL processes creating copies, no data lakes accumulating exposure. The attack surface that does not exist cannot be breached.

Partners get answers. Not access.

A certified partner can run a propensity model, receive a recommendation, and act on it - without ever touching the underlying booking data. The commercial model that makes third-party AI viable without compromising customer data.

Document transformations execute inside the instance.

Transformation Functions run inside TripForge against data in place. Check-in arrives; boarding passes generate automatically. Event-driven document manufacturing without custom integrations.

Each system sees exactly what it needs. Nothing more.

Per-booking-type, per-document-type, per-operation access control. Each system sees exactly what it needs. Nothing more. Designed for the reality of travel operations.

A $13B market. We target 1% of it.

$13B
SAM: AI infrastructure and automation in travel by 2030 · 28% CAGR
$5M
Seed raise underway. 30 months to cash-flow positive on current model.
$130M
5-year revenue target · 1,277 customers · English-speaking and Western European markets only
80%
Target EBITDA margin by 2031. IaaS model: near-zero marginal cost per customer at scale.

What changes when the data layer is fixed.

Every AI initiative in travel is blocked by the same problem. Fix it once, and the downstream effects compound.

01

Development starts on the product. Not the plumbing.

Every AI project in travel currently begins with the same undifferentiated work: extracting data from multiple systems, normalizing inconsistent formats, reconciling conflicting records, and building pipelines to keep everything current. That work produces no competitive advantage. It is the cost of having the wrong architecture.

TripForge eliminates it. The data is already consolidated, versioned, and structured. Development time - human or LLM-assisted - starts on the product. New customer experiences that previously took quarters to scope now take weeks to build. That gap widens as AI coding tools improve.

02

AI can see a complete booking. For the first time.

Personalization, disruption management, dynamic pricing, proactive service recovery - every meaningful AI application in travel requires a complete, real-time view of the booking. Today that view does not exist in any single system. AI cannot operate on data it cannot see.

TripForge gives AI exactly what it needs: every document, every version, every change event - alongside the metadata and reference data that give those documents meaning. Everything in one place, queryable in real time. The AI use cases that have been theoretical become operational.

03

Back-office complexity stops scaling with the business.

For distributors, every new supply channel added today increases back-office complexity. More suppliers means more confirmation formats, more change notification patterns, more exception handling, more manual reconciliation. The operational cost grows with the supply chain.

TripForge decouples them. The persistence layer absorbs the complexity of the supply chain. Back-office processes read from one place regardless of how many suppliers feed into it. Growth in supply no longer means proportional growth in operational cost.

04

Content connections without the normalization tax.

Connecting a new supply channel carries a cost that is less visible but equally real: the normalization layer. Every supplier's content must be translated into a canonical format before it can be used - a process that is expensive to build, lossy by nature, and requires ongoing maintenance every time a supplier updates their format.

TripForge removes the requirement to normalize at ingestion. Each supplier's documents are persisted in their native format. Transformation Functions generate the normalized display view when it is needed - for customer-facing booking summaries, for AI consumption, for reporting. The supplier-specific detail is preserved underneath. Connecting a new supply channel becomes a question of defining how documents are persisted and what the display TF produces - not building a bidirectional translation layer from scratch. Time-to-market for new supply relationships drops. The supplier's native data, preserved intact, remains available for any future use case that requires it.

05

Legacy infrastructure becomes optional.

The long-term cost argument in travel is not about AI. It is about the legacy systems that have accumulated over 65 years. Every major travel company carries significant licensing, maintenance, and integration overhead for systems that exist primarily because no alternative data layer existed.

TripForge is that alternative. As it becomes the record of origin - on the customer's timeline - legacy systems can be wound down selectively. Not as a forced migration. As a deliberate cost reduction, delivered at the pace the customer chooses.

We have done this before — in this industry.

Founder-market fit is not the pitch. It is the résumé.

WP
William Phillipson
Founder & CEO
Co-founder and President/COO of SilverRail Technologies. $75M raised; $150M sale to Expedia Group (2017), founding team retaining ~20% equity. 25 years in travel tech infrastructure across airlines, rail, holidays, and navigation.
BC
Brendan Collins
Co-Founder & Architecture / AI Engineering
Founding System Architect at SilverRail. Built SilverCore from the first line of code. 30+ years; four exits including Expedia and Salesforce ($2.8B).
JD
Jean-David Oladele
Co-Founder & Product
VP Product at SilverRail (2012-2025). Led global product and design team of 21. Yale MBA, MIT AI certification.
MH
Murtagh Hunt
Co-Founder & Customer Success
Director, Platform and Reliability Engineering at Mastercard. Director of System Operations at ITA Software through the $700M Google acquisition.

Advisor: Oliver Heckmann — former VP Engineering at Google, leading Google Travel and Google Shopping (1,500+ engineers). Previously Head of Engineering at Coda (acquired by Grammarly, 2024).

Writing on the data layer problem.

Published weekly on LinkedIn. The architecture argument, made in full.

Get involved.

We are raising our seed round and working with a small number of initial customers. If you're building in travel or investing in infrastructure, reach out.

No spam. Direct contact only.

Message received.

We will be in touch directly.