Privacy Policy

Last updated: August 30, 2026

株式会社ファネルAi (formerly 株式会社ロゴラボ; the “Company,” “we,” “us” or “our”) establishes this Privacy Policy (the “Policy”) concerning the handling of personal information we obtain.

Unless otherwise specified, terms in this Policy have the meanings given in Japan’s Act on the Protection of Personal Information (Act No. 57 of 2003; “APPI”) and related laws.

This Policy describes the handling of information about users, including personal information and user information, in Funnel Ai (the “Service”).

The Service complies with APPI, other applicable laws and guidelines, and the Google API Services User Data Policy.

Use of the Service is also subject to the Terms of Service, guides, help materials, notices and other provisions (the “Supplementary Provisions”), which form part of this Policy.

1. Collection and purposes of use of personal information

We obtain users’ personal information appropriately and use it within the purposes set out in this section.

1.1 Information collected

Examples of information collected and collection methods are listed below. We collect only the minimum necessary information.

• Account information: names, email addresses and profile images, obtained through Google authentication.
• Authentication and permission information: OAuth tokens, permission scopes and integration status, obtained based on user actions.
• Google user data: obtained within OAuth-approved scopes. Data types, purposes and retention periods are described in Section 7. This includes Gmail messages, threads, labels and drafts; Google Calendar events and booking information; Google Meet meeting metadata and transcripts; and Google Drive Doc files generated by Google Meet.
• Zoom integration data: Zoom account information, OAuth tokens, meeting information, cloud recordings, transcripts and Webhook events. See Section 8.
• Usage and device information: IP addresses, cookies, browser types, device information, referrers, operation histories and error logs, collected automatically.
• Support information: inquiries, log extracts and attachments voluntarily provided by users.

1.2 Purposes of use

The Service uses collected information for the following purposes:

• Service provision: providing, maintaining and improving the Service, identity authentication and saving user settings.
• Email classification and reply drafts: AI classification and creation of drafts when expressly instructed by the user.
• Scheduling and calendar synchronization: creating, updating and canceling calendar events and automatically generating Google Meet links.
• Meeting summaries: automatically generating summaries from Meet transcripts.
• Support: responding to inquiries and outages and sending important notices.
• Analysis and improvement: usage statistics and analysis, including processing into a form that does not identify individuals. The bodies and contents of Google user data are not used for analysis.
• Legal compliance: fulfilling legal obligations and responding to laws or requests from administrative authorities and similar bodies.
• Transactions and communications: communicating with parties necessary to provide the Service and performing work necessary for smooth and appropriate transactions with users.
• Related purposes: purposes incidental to those above.

The Service may integrate with External Services such as Gmail based on user instructions. Their availability and specifications depend on their providers, and the Service does not guarantee their continued provision.

2. Third-party provision and outsourcing

We do not provide personal information to third parties except:

• With the user’s consent;
• As required by law;
• Where necessary to protect life, physical safety or property and obtaining consent is difficult;
• Where particularly necessary to improve public health or promote the sound development of children and obtaining consent is difficult;
• Where cooperation is necessary for a national or local government body, or its entrusted party, to perform legally prescribed duties and obtaining consent may impede those duties; or
• Where our business, including the provision of personal information, is succeeded to through a merger, company split, business transfer or other reason.

We may outsource infrastructure, delivery, analysis, customer support and other activities as necessary to operate the Service. We contractually require appropriate security measures and supervise these providers.

Provision and outsourcing involving Google user data and Zoom data are governed by Sections 7 and 8, respectively.

3. Requests for disclosure and other actions

When an individual or their representative requests disclosure, correction, addition, deletion, cessation of use or cessation of third-party provision of retained personal data, we respond appropriately under APPI. We verify identity using our prescribed procedures.

We may withhold all or part of the information if disclosure could harm the life, physical safety, property or other rights of the user or a third party; materially interfere with our business operations; or violate law. We will promptly notify the requester in such cases.

If personal information disclosed to a user is incorrect, we will promptly correct it upon the user’s request.

If we receive a claim that personal information is handled outside its purposes of use or was obtained improperly, we will promptly investigate as necessary and, based on the results, cease use or erase the information (“Cessation of Use”) and notify the user. If Cessation of Use would entail substantial expense or is otherwise difficult, we will take alternative measures necessary to protect the user’s rights and interests.

Users may also revoke OAuth access through their Google account settings.

4. Changes to this Policy

We will announce changes in advance by posting them on our website or sending email. Minor changes that do not materially disadvantage users, including changes resulting from legal amendments, corrections of errors and clarifications, may be announced afterward.

For material changes, such as additional categories of collected data or fundamental changes in purposes of use, we will obtain users’ consent again.

The amended Policy applies when posted on our website, unless we specify a different effective date.

5. Contact

For inquiries about this Policy or personal information handling, contact:

• Operator: 株式会社ファネルAi (formerly 株式会社ロゴラボ)
• Person responsible for personal information protection: Nozaki Tomohiro (野﨑 智裕)
• Email: info@funnel-ai.jp

6. Security measures

We implement organizational, personnel, physical and technical security measures to prevent leakage, loss, damage, alteration and unauthorized access.

• Encryption at rest: data in PostgreSQL (Cloud SQL), Firestore and Google Cloud Storage is encrypted using Google Cloud Platform’s default encryption (Google-managed encryption keys, AES-256).
• Encryption in transit: all communications between users and the Service and between the Service and Google APIs use TLS 1.2 or later.
• Access control: staff access to production data is restricted using IAM roles, and all access is recorded in audit logs.
• OAuth tokens: refresh_token and access_token are encrypted at the application level using AES-256-GCM before storage in Firestore, which is also encrypted at rest as described above. Plaintext tokens are not left in logs, code, error output or AI inference prompt logs.
• No content logging: Google user data content, including Gmail bodies, Calendar event details and Meet transcripts, is not written to Cloud Logging, application logs or AI inference prompt logs.
• Backups: we use Google Cloud SQL automatic backups. Backups include the entire database, including Google user data content, and are stored encrypted at rest using Google-managed encryption keys (AES-256). Retention follows GCP’s standard operational settings, and GCP automatically deletes backups after the retention period.
• Cache: Google user data content is not stored in short-term cache layers; it is processed only in memory.

7. Google user data collected

7.1 Data types and OAuth scopes

Through Google account integration, the Service collects the following data within scopes approved by the user through OAuth.

Gmail data

• Scopes: https://www.googleapis.com/auth/gmail.readonly (Restricted), https://www.googleapis.com/auth/gmail.modify (Restricted), and https://www.googleapis.com/auth/gmail.compose (Restricted).
• Data: email bodies, subjects, senders and recipients, thread information, metadata such as message_id, thread_id and history_id, labels and drafts.
• Why narrower scopes are not used: gmail.metadata alone cannot retrieve email bodies needed for AI intent and priority classification and reply-context analysis. gmail.labels alone cannot apply labels to messages (messages.modify). gmail.send alone cannot manage the draft lifecycle (create, update and delete). The combination gmail.readonly + gmail.modify + gmail.compose is therefore the minimum permission set for the Service’s email workflow: retrieving incoming email bodies, AI classification, label management, draft creation and sending. The broader full-access scope https://mail.google.com/ is not used.

Google Calendar data

• Scope: https://www.googleapis.com/auth/calendar (Sensitive).
• Data: calendar events (titles, descriptions, start and end times and attendees), created event information and calendar metadata.
• Why narrower scopes are not used: calendar.readonly cannot perform events.insert or events.delete required for scheduling. calendar.events cannot access calendar metadata (calendarList.get / colors.get) needed for the interface. calendar.freebusy cannot create or delete events. The calendar scope is therefore currently necessary.

Google Meet meeting metadata

• Scopes: https://www.googleapis.com/auth/meetings.space.readonly (Sensitive) and https://www.googleapis.com/auth/meetings.space.settings (Non-sensitive).
• Data: meeting transcript metadata and Meet space settings (automatic transcription settings only).
• Why narrower scopes are not used: meetings.space.readonly retrieves transcript metadata; meetings.space.settings is used only to patch autoTranscriptionGeneration. Together these are the minimum permissions for retrieving transcripts and enabling automatic transcription.
• Updates through spaces.patch using meetings.space.settings are limited to the single field config.artifactConfig.transcriptionConfig.autoTranscriptionGeneration. Recording, participant controls and other Meet settings are not changed.

Google Drive data

• Scope: https://www.googleapis.com/auth/drive.meet.readonly (Restricted; Meet-generated files only).
• Data: Doc files in Drive created or edited by Google Meet, such as transcripts and meeting notes.
• Why this scope is used: drive.readonly grants access to all of Drive and would be excessive for a Service that only accesses Meet-generated transcript Docs. We intentionally use drive.meet.readonly as the narrowest option.
• Important: the Service does not access arbitrary Drive files. Access is limited to files created or edited by Google Meet. The Service does not have a Drive picker or other interface for selecting arbitrary files.

Google account information

The Service obtains the Google account email address for identity verification and communication, and the display name and profile image for display within the Service.

7.2 Purposes of use

Collected Google user data is used only for:

• Email classification and organization: AI determination of intent and priority and automatic labeling;
• Reply draft generation: Gmail data is referenced only when expressly initiated by the user; sending requires user action and is not automatic;
• Scheduling and booking management: creating, updating and canceling calendar events and automatically generating Google Meet links;
• Generating meeting summaries from Meet transcripts;
• Account identity verification and authentication state management; and
• Minimum necessary processing state management, including message_id, thread_id, event_id, processing status and timestamps.

7.3 Data flow summary

• Gmail: receive push notification → retrieve message body using gmail.readonly → classify with Gemini API → apply labels using gmail.modify → create draft using gmail.compose. Sending requires user action; no automatic sending.
• Calendar: retrieve events through events.list / events.watch using calendar → create scheduling bookings through events.insert and delete cancellations through events.delete.
• Meet summaries: meeting ends → obtain fileId through conferenceRecords.transcripts.list using meetings.space.readonly → retrieve text through files.export(fileId, "text/plain") using drive.meet.readonly → generate summary with Gemini API.
• Meet automatic transcription: when creating a new meeting, enable automatic transcription through spaces.patch using meetings.space.settings, with updateMask limited to the autoTranscriptionGeneration field.

Prompts and responses sent to Gemini API during inference are not retained in logs. Gmail bodies, Calendar details and Meet transcripts are not retained in error logs, AI inference prompt logs or short-term cache layers. See Section 6.

7.4 Third-party provision and outsourcing

The Service does not transfer collected Google user data to third parties, including for advertising, data sales or provision to data brokers for marketing.

However, to provide the Service, we entrust the minimum necessary processing to the following providers. They process data under our instructions and contracts and do not use it for independent purposes.

The Service uses Gemini API under a paid-tier contract. Submitted data is used only to generate inference results and is never used to train AI models, in accordance with Google Cloud’s Generative AI Service Terms.

7.5 Restrictions on human access

Staff are prohibited from reading users’ Google user data, including Gmail bodies, Calendar content and Meet transcripts, except:

• With the user’s specific permission (consent) to resolve support or technical issues;
• Where a security investigation is necessary, such as suspected misuse;
• Where necessary for legal compliance, such as a court order; or
• Where urgently necessary to protect the user’s life, physical safety or property.

All human access is recorded in audit logs and is subject to internal audit.

7.6 Tenant isolation

The Service uses a multitenant design and enforces tenant_id constraints on all database queries and API calls to prevent data leakage between tenants. The technical implementation of tenant isolation has been independently verified in a CASA Tier 2 audit report.

7.7 Google API Services User Data Policy compliance (Limited Use)

The Service complies with the Google API Services User Data Policy, including its Limited Use Policy, when handling information received from Google APIs.

“Funnel Ai's use and transfer to any other app of information received from Google APIs will adhere to the Google API Services User Data Policy, including the Limited Use requirements.”

Specifically, the Service complies with these four Limited Use conditions:

1. Limited purposes: Google user data is used only to provide features expressly approved by users, such as email classification, draft generation, calendar synchronization and meeting summaries.
2. No advertising use: Google user data is not used for advertising, including targeting, profiling or advertising effectiveness measurement.
3. Restricted human access: staff do not read Google user data except in the limited circumstances in Section 7.5.
4. Restricted transfers: Google user data is not transferred to third parties other than the essential service providers described in Section 7.4.

8. Zoom integration data

With authorization from users or their organization’s administrator, the Service may obtain Zoom account information, OAuth tokens, meeting information, cloud recordings, transcripts and Webhook events through Zoom integration.

We use this information to create and manage Zoom meetings, coordinate schedules, maintain meeting records, retrieve and display cloud recordings and transcripts and associate them within the customer’s tenant, and maintain integrations, address failures and prevent misuse.

Zoom OAuth access and refresh tokens are stored encrypted using AES-256-GCM. External communications use HTTPS. Webhooks received from Zoom are verified using the Secret Token, x-zm-signature and x-zm-request-timestamp. Zoom data is stored and processed separately for each customer tenant.

When Zoom is disconnected, we invalidate and delete stored OAuth access and refresh tokens and stop acquiring new data from Zoom APIs. Business records already imported are retained until the customer tenant is deleted or the customer requests deletion. We may retain information as necessary for legal compliance, dispute resolution or audits.

We do not sell Zoom data or use it for advertising delivery. We do not provide it to third parties except with user consent, as required by law, or when entrusting minimum necessary processing to providers needed to deliver the Service.

9. Retention and deletion

Even after account deletion, we may retain information for a certain period to the extent necessary for legal compliance, dispute resolution, audits or similar needs.

Retention and deletion conditions for Google user data acquired through the Service are set out below. When OAuth access is revoked or a tenant’s contract is terminated, we stop new data acquisition and, as a rule, delete stored data within 30 days.

Send deletion requests to the contact in Section 5. Content included in Cloud SQL automatic backups is automatically removed by GCP after the backup retention period.

Retention and deletion of Zoom integration data are governed by Section 8.

10. International transfers

Data may be transferred outside Japan through cloud services and similar arrangements. We take necessary protective measures under the laws applicable in the destination jurisdiction.

11. Minors

Minors should obtain a parent or guardian’s consent before using the Service. If we learn that a user is a minor, the Service may take necessary measures.

12. Cookies

Our website uses cookies. You may reject cookies through your browser settings.

13. Access history

Our website records access history to improve services. This may include IP addresses, cookies, browser types, device information and referrers. We use this information for usage statistics, analysis and service improvement, not to directly identify users.