Products Lemma APIProof issuance & verification platformTrust402Delegate to agents, and transactSealSign-in for the agent era — no keys handed over
Use cases Manufacturing & Critical InfraInspection Record AssuranceFinance & FinTechCounterparty Record VerificationPublic SectorCertificate-less ProceduresHealthcareQualified Worker AssuranceProcurement & Supply ChainSupplier Credential MonitoringMedia & ContentContent AuthenticityService & RetailCross-group IdentityAI Adoption (cross-industry)AI Run GovernanceDevelopers & Agent OpsAgent Authority Control ▸ Browse the use-case index
Pricing
Resources Critical BriefThe frontier of AI × trustBlogThinking and implementation notesDocumentationAPI & specsVerification CenterReal verification & issuance countsAbout usFRAME00, Inc.ContactSales & press inquiriesGlossaryDefinitionsFAQFrequently asked questions
Dashboard ↗ JA
Home / Critical Brief / No. 032
No. 032 Critical Brief · Regulatory Attribute Proof

Polaris Holdings / Booking.com: A hotel's payout account was switched to a third party's inside the legitimate admin screen, and one hotel lost about ¥9 million (company disclosure)

≈ ¥9million
Loss at one hotel (as of disclosure)
Severalhotels
Had their payout account rewritten. At all but one, the fraudulent payout was prevented
05-23→05-28
Five days from spotting the anomaly to disclosure
CountryJapan
Incident date2026-05-23
Published2026-06-08
Revised v2.1 · 2026-11-02
AuthorsLemma Critical Team
Contents
  1. 01Summary
  2. 02What got through
  3. 03Pre-execution check
  4. 04What Lemma does
  5. 05Facts and timeline
  6. 06Sources

Summary

Polaris Holdings says payout accounts at several hotels were switched, and that an early response prevented fraudulent payouts at the others. The switch was made in the legitimate admin screen, as a logged-in operation. To the system it looks like an ordinary settings change. A change to where the money goes, once confirmed on a logged-in session alone, cannot be told apart from a correct setting.

Exhibit 1Detection vs pre-execution check

A pre-execution check means matching, before a payout-account change takes effect, whether another person approved it, and keeping a record. Detection means noticing after an anomaly surfaces. In between, the switched account is treated as a correct setting.

Pre-execution check · not known from the disclosuresMatch the change to another person's approval before it takes effect
What was confirmedPayout accounts switched at several hotels
About ¥9M lost at one
Detection · workedNotice the account anomaly and stop the other fraudulent payouts
STEP 1
Login
STEP 2
The payout-account change takes effect
STEP 3
The anomaly is noticed
STEP 4
Other payouts are stopped

What got through

What got through
A change to the bank account that receives a hotel’s sales proceeds, in Booking.com’s hotel admin site (the extranet). Polaris Holdings, which runs the hotels, managed its group’s hotels through a group-level account. Payout accounts at several hotels were switched to a third party’s.
Loss
At one hotel, part of its receivables (money it was owed) was paid into the fraudulent account, a loss of about ¥9 million (at the time of disclosure). At several other hotels an early response, in the company’s words, prevented fraudulent payouts.
What reached the receiving side
In general, an operation made on a logged-in session (an authenticated session) looks to the system like a legitimate one. A payout-account change and a contact change arrive as the same kind of settings change. The payout itself is routine: money to the registered account.
Spillover to guests
Phishing messages reached guests booked at some hotels. The company is looking into whether booking data taken from the admin screen was used, or whether it leaked from an external booking-management system.
Not disclosed
How the attacker got in (under investigation at disclosure), and whether the payout to one hotel came before or after the anomaly was detected.

Pre-execution check

Whatever the route, the condition for telling them apart is that whoever applies a payout-account or bank-account change checks, before it takes effect, whether someone other than the person who made the change approved it through a path separate from the login.

The core of this case
A change to where the money goes, once confirmed on a logged-in session alone, cannot be told apart from a correct setting.
Who approves
A separate approver decided in advance. The approval is given in a form that the changing account’s own path cannot produce. If it could be approved through the same path as the taken-over account, it would mean nothing.
Where it is checked
The part of the system that applies payout-account and bank-account changes. Changes from the admin screen and from a supplier’s request are both checked at one point before they take effect. A change that does not match an approval is not activated, and payments to that account are held until the check is done. The new account’s holder is matched against the bank’s own record.
What is recorded
When, on whose approval, and to which account a change was let through, kept in a form a third party can verify later.

The anomaly was noticed after the change had taken effect. Other cases where “arrived through a legitimate route” stood in for “is a correct request” are Revolut, which answered disclosure requests that came through a national certified-mail network, and SpeeCAN RAIDEN, where a fake email arrived from the genuine delivery domain.

What Lemma does

Lemma builds the pre-execution check into the side that applies payout-account and bank-account changes. It does not look for intrusions; it confirms, before a change takes effect, that it rests on another person’s approval. Because this change happened inside the booking site’s admin screen, the place to build it in is the booking or payment service. What a user can add is a stop after the change and before the first payout (the hold in 03, and the checklist below).

Industries and operations to watch

IndustryServices & retail
Lodging, restaurant and retail businesses that receive sales proceeds through booking sites, e-commerce or payment providers. The payout-account setting sits in an external service’s admin screen.
IndustryFinance & fintech
Banks and payment providers that accept registration and changes of payout accounts. A payout account rewritten on the customer’s side also arrives as a valid instruction.
OperationBank account changes
Registering or changing your own or a counterparty’s payout account. The result stays hidden until the next payment.
OperationAccounts & permissions
Managing accounts at external services and who can change what. The switch here was made from a taken-over account.

Current controls vs Lemma

The route in this case is column 1 (a change to the payout account, made on a logged-in session in the admin screen). Column 2 is a supplier’s request to change an account; column 3 is where the approver was fooled or the approval path was taken over. Whether either applied here has not been disclosed.

Control Logged-in change in the admin screen A supplier’s request to change an account Approver fooled, or approval path taken over
Multi-factor authentication (MFA) Partial Not relevant Not effective
Endpoint monitoring (EDR) Partial Not relevant Not effective
Call-back confirmation Not effective Effective Not effective
Email sender authentication (DMARC etc.) Not relevant Partial Not relevant
Anomaly detection and early response Partial Partial Partial
Lemma (pre-execution check) Effective Effective Not effective

MFA works at the door, but because the route is undisclosed it is not known whether it mattered here, and once a device or a logged-in session is taken over it does not look at what is done afterwards. Endpoint monitoring might catch an infection like the one in Sekoia’s example, but it cannot see settings changes made inside a booking site or external service. Call-back confirmation works when a supplier asks to change an account; when your own setting is rewritten, nobody gets a call. Detection works only once an anomaly surfaces.

Note: Lemma’s check works only when it is built into the system that applies payout-account changes, changes from the admin screen and from requests pass the same check, and the approval path is separate from the system that receives the change. It passes if the approver is fooled or the approval path itself is taken over. Holding changes until checked also delays legitimate ones. The table compares what each control checks; it does not show whether this incident would have been stopped. Lemma does not detect or prevent intrusions.

CheckChecklist

Three things to confirm about bank-account changes

  • Who can change a payout or bank account, and who other than the person who made a change is notified?
  • Can you hold the first payment until someone other than the changer approves the change through a different channel from the one used to register it?
  • Who matches the first payment after a change, within how many days, and can you stop the next transfer if it does not match?

Facts and timeline

Party
Polaris Holdings (hotel operator). Its group hotels managed the booking site Booking.com through a group-level account.
What was changed
The account into which Booking.com pays a hotel’s sales proceeds. A setting the hotel registers in the admin screen.
Industry background
The admin screens of hotels listed on Booking.com have long been targeted. Security firm Sekoia’s November 2025 research found emails posing as Booking.com being used to get hotel staff to run malicious programs and steal admin-screen logins; guests were then lured, from the taken-over account’s email or WhatsApp and with genuine booking details, to pages posing as a payment re-confirmation. Whether this case’s intrusion route matches has not been disclosed.

Glossary
Extranet — The admin site Booking.com provides to hotels.
Authenticated session — The state after a successful login. Anything done in it looks like a legitimate action to the system.

Timeline

2026-05-23
Polaris spots an anomaly in its group Booking.com account and starts investigating.
Investigation
Payout accounts at several hotels had been rewritten. At one hotel part of the receivables was paid into the fraudulent account, a loss of about ¥9 million; at the others an early response prevented fraudulent payouts.
2026-05-28
Polaris discloses the unauthorised access. The travel-industry outlet TRAICY reports it the same day.
2026-06-02
through 06-03, security outlets ScanNetSecurity and Security NEXT report it.

The facts here rest on Polaris Holdings’ disclosure of 28 May and the reporting on it. How the attacker got in, and whether guests’ personal data leaked, were still under investigation at disclosure. What has not been disclosed does not mean there were no controls. The disclosure document has since been taken down from the company’s site, so this Brief cites the press coverage.

After disclosure

  • Polaris: changed the passwords of every hotel’s Booking.com-related accounts, requested the access logs, and acted to contain the damage. No credit-card data leak has been confirmed. It expects a minor impact on earnings.
  • Industry: attacks on Booking.com’s hotel admin site keep being disclosed in Japan as well. Most target guests with phishing; this case went a step further and rewrote the hotel’s own money flow, turning the attack into a direct loss.

Sources

  1. Primary, company disclosurePolaris Holdings “Notice of unauthorised access (Booking.com account)” (Japanese)2026-05-28 · the document has since been taken down from the company’s site
4 more (reporting, analysis)
  1. PressSecurity NEXT “Payout account details at lodging booking service tampered, fraudulent transfer — Polaris HD” (Japanese)2026-06-03 · security-next.com/185176
  2. PressTRAICY “Polaris HD suffers fraudulent transfer after unauthorised access to Booking.com; phishing messages to guests also confirmed” (Japanese)2026-05-28 · traicy.com/posts/20260528371641
  3. PressScanNetSecurity “Unauthorised access to Booking.com, sales payout account tampered; about ¥9 million lost” (Japanese)2026-06-02 · scan.netsecurity.ne.jp/article/2026/06/02/55403
  4. Industry researchSekoia “Phishing campaigns ‘I Paid Twice’ targeting Booking.com hotels and customers”2025-11-06 · sekoia.com/blog/…

References: Why after-the-fact detection is not proof — “The last layer left for cyber defense in the age of AI”.

The same method elsewhereSee all 16 "Payee & bank-account changes" cases

This Brief sets out, from disclosed facts, what organisations in a similar position can check. It is not guidance on individual losses or procedures.

This material is a structured analysis of public information; it is not an audit, diagnosis, or recommendation for any specific organization.

Cite this Brief

Lemma Critical Team. (2026).
"Polaris Holdings / Booking.com: A hotel's payout account was switched to a third party's inside the legitimate admin screen, and one hotel lost about ¥9 million (company disclosure)".
Lemma Critical Brief No.032. Lemma / FRAME00, Inc.
https://lemma.frame00.com/critical/briefs/032-booking-payout-account-tampering/
Read nextNext
All Briefs →
Revision History

Current version: 2.1

Lemma

If it can't be verified,
it doesn't enter your operation.

Lemma is a pre-execution check: logins, payments, records and publishing that look genuine are checked before they run, with cryptographic proofs the receiving side verifies without asking the issuer. Detection stays; the check sits in front of it.