Privacy Policy

Effective Date: November 9, 2026

What changed: On November 9, 2026, Section 3.4 added PostHog, which we use for website analytics and for replays of some visits. On November 6, 2026, Section 3.4 added what the LinkedIn Insight Tag records on its own (the address and title of the pages you visit, how long you stay, what you click on and a random browser ID) and the hashed copy of your email address it sends LinkedIn when you request access, and Section 18 added that turning advertising off also stops the Meta Pixel and the Insight Tag on the pages you have open. On November 5, 2026, Sections 3.4, 9 and 18 added the LinkedIn Insight Tag and LinkedIn’s Conversions API on worker.ai, the sources we store with a request for access (campaign tags, ad click identifiers and the first page you visited), and the optional question about how you heard about Worker. On November 4, 2026, Sections 3.4, 6.3, 9 and 18 added the Meta Pixel and Meta’s Conversions API on worker.ai, including the hashed email addresses they send Meta, and how to opt out. Later the same day, Section 3.4 added the events that worker.ai’s own server sends Meta. On November 3, 2026, Sections 3.4, 6.3 and 18 added Google Analytics. Nothing changed for the Services or for data from connected accounts.

At Icos AI Inc., the company behind Worker (“Worker”, “we”, “our”, or “us”), we build software that turns commitments made on calls and in correspondence into commitments that get executed. Doing that means we handle information about you, about your colleagues, and about people at the organizations you work with. This Privacy Policy explains what we collect, why, how long we keep it, who sees it, how we use it, and what choices you have.

This Privacy Policy is incorporated into and subject to our Terms & Conditions. Terms used here without definition have the meaning given in the Terms & Conditions. If you have questions, or need this policy in another format, contact us at privacy@worker.ai.

1. Scope of this Privacy Policy

This Privacy Policy applies to our websites, applications, and services (together, the Services), and to every category of individual described in Section 3, including individuals who are named in a customer’s content, but who have never signed up for Worker.

It does not apply to the data practices of organizations we do not own or control, including the providers of the email, messaging, calendar, meeting, and CRM systems that our customers connect to the Services. Where our customers use the Services to record commitments involving people who are not Worker users, we act as a processor on that customer’s instruction, and Section 3.3 and Section 14 explain how we handle that data and how those individuals can exercise their rights.

2. Who we are, and who is responsible for your data

Icos AI Inc., 757 Third Avenue, 24th Floor, New York, NY 10017, United States, is the controller of personal data described in this Privacy Policy, except where we state that we act as a processor.

Where a customer instructs us to ingest, reconcile, and store content from its connected systems, that customer is the controller of the personal data in that content and Worker is its processor.

You can reach our data protection contact at privacy@worker.ai. If you are located in the European Economic Area, the United Kingdom, or Switzerland and wish to contact our representative for your region, email privacy@worker.ai and we will put you in touch.

3. Categories of individuals and personal data we collect

3.1 Customers and authorized users

Contact and profile details (name, business email, telephone number, job title, organization, business address); account and authentication data (username, hashed credentials, single sign-on identifiers, session and device information, IP address); billing information (billing address and payment card details, processed by our payment providers, we do not store full card numbers on our servers); records of your use of the Services, including which surfaces you open and which commitments you act on; and records of your correspondence with us, including support tickets and feedback.

3.2 Meeting and Commitment Data

When you connect a source or upload content, we process the following, together Meeting and Commitment Data:

• Meeting artefacts produced by your own tools: transcripts, meeting notes, summaries, participant lists, meeting titles, times, and calendar metadata.
• Correspondence and workspace content from the sources you connect: email messages, chat messages, calendar invitations, and CRM records, limited to the mailboxes, channels, and objects you authorize. Connected accounts, and the provider-specific terms that apply to them, are described in Section 4.1.
• Extracted commitments: the commitment itself, the person identified as owner, the deadline, the verbatim source excerpt that evidences it, the organization on each side, and the status history of the commitment over time.
• Reconciliation data: the record of how a commitment was matched across your connected sources, including message identifiers, timestamps, and the confidence signals we used.
• Derived surfaces: pre-meeting briefs, weekly receipts, and per-account records assembled from the above.

We do not create voiceprints, speaker templates, or any other biometric identifier, and we do not perform speaker identification that produces a reusable voice profile.

3.3 Individuals identified in customer content who are not Worker users

Our Services record commitments made in meetings and correspondence. Those records name individuals who do not have a Worker account, including people who work at other organizations. For those individuals we may process: name, business contact details, role and employer, the commitment attributed to them, the verbatim source excerpt, and the dates on which the commitment was made, updated, or completed.

We process this data as a processor on the instruction of the customer whose account holds the record. That customer is responsible for having a lawful basis for the processing and for providing any notices and obtaining any consents that the law requires. Section 14 explains how a person in this category can access, correct, or delete a record about them.

3.4 Website visitors

Device and browser data, including IP address, browser type, operating system, referring URLs, pages viewed, and time spent on pages. Our Cookie Policy explains the technologies involved. If you join the waitlist on worker.ai, we also collect your work email address, the address of the page that referred you, the campaign tags in the links that brought you to worker.ai (from your first and your latest visit in the last 90 days), the first page you visited, a random ID for your request and, where advertising is allowed, the ad click identifiers that Meta and LinkedIn add to their ad links. If you choose to answer our question about how you heard about Worker, we store your answer with your request. We use all of this to contact you about Worker, to set up your company and to measure our ads.

We use Google Analytics to collect website data, including your approximate location (from your IP address, which Google Analytics does not store), device and browser details, and how you use the pages: clicks, scrolling, film plays and form use. Google signals is turned on, so Google can associate this data with your Google account if you are signed in to Google and have allowed ad personalization, and can use it for advertising (Section 18). We do not send your email address or what you type into forms to Google Analytics.

We also use PostHog, from PostHog, Inc., which works only for us, to collect website data: the pages you view, what you click on, scrolling, film plays, form use, your approximate location (from your IP address), device and browser details, and the campaign that brought you. PostHog also records replays of some visits, which show how the page looked and where you moved, clicked and scrolled. Anything you type is hidden before it leaves your browser, and so is the company name you type into the demo. We do not send PostHog your email address; when you request access, it receives the ID of the request only. PostHog does not use this data for advertising.

We also use the Meta Pixel and Meta’s Conversions API, from Meta Platforms, Inc., to measure our ads and to show ads to people who have visited worker.ai. They send Meta the pages you view, clicks, scrolling, film plays, form events (such as a request for access, or an address we decline), device and browser data, your IP address, a random visitor ID that we create in your browser, and a hashed (scrambled) copy of any email address you enter in our forms, including addresses we decline. Meta’s automatic event detection and automatic advanced matching are turned on, so Meta’s code can also record clicks and page details on its own and can read the email field in our forms to send Meta a hashed copy. Meta can match this data to a Facebook or Instagram account and use it for advertising (Section 18). We never send Meta your email address in readable form, anything else you type, or the company name you type into the demo. Our own server also sends Meta a copy of these events through the Conversions API, with your IP address, browser details, approximate location (country, region, city and postal code, estimated from your IP address) and hashed identifiers, so that Meta receives them even when a browser blocks the Pixel. In the European Economic Area, the United Kingdom and Switzerland, neither the Meta Pixel nor our server sends anything to Meta until you accept.

We also use the LinkedIn Insight Tag and LinkedIn’s Conversions API, from LinkedIn Corporation, to measure our ads on LinkedIn, to show ads to people who have visited worker.ai, and to see aggregate information about the professional characteristics of our visitors, such as job function and industry. The Insight Tag sends LinkedIn the address and title of each page you view (including any campaign tags and ad click identifiers in the address), how long you stay on it, what you click on, including the visible text and attributes of the link or button (with any email address or phone number in that text masked), a random ID it keeps in your browser, device and browser data and your IP address, and tells LinkedIn when you click Request access or send a request. When you send a request, the Insight Tag also sends LinkedIn a hashed copy of your email address, computed in your browser before it is sent, and our server sends LinkedIn a hashed copy of your email address and, if you arrived from a LinkedIn ad, LinkedIn’s click identifier, so LinkedIn can match the request to a LinkedIn member and count it once. LinkedIn can use this data for advertising (Section 18). LinkedIn never receives your email address in readable form, anything else you type into our forms or your answer to our question about how you heard about Worker. The Insight Tag and our server send nothing to LinkedIn where the Meta Pixel would not load: not when your browser sends Global Privacy Control, not after you turn advertising off, and in the European Economic Area, the United Kingdom and Switzerland not until you accept. If you turn advertising off, the Insight Tag stops at once, also on pages you already have open, and we remove the identifiers it stored for worker.ai in your browser.

3.5 Job applicants

Application materials you send us, including CV, cover letter, work history, education, and references, and, where permitted by law and with appropriate notice, eligibility and background information.

3.6 Suppliers and partners

Business contact information, and the contract, banking, and tax details needed to manage the relationship.

4. Where the data comes from

1. From you directly, when you register, complete a form, contact us, or use the Services.
2. From your organization, when an administrator provisions your access or connects a source.
3. From systems you authorize us to read. We receive Meeting and Commitment Data only from sources a customer connects and authorizes, using the permissions that customer grants. Worker does not join meetings, does not deploy a bot or a recorder, and does not itself create an audio or video recording of any conversation. Where a meeting artefact reaches us, it was produced by a tool the customer already operates. Connected accounts are described in Section 4.1.
4. From automated technologies on our websites, as described in the Cookie Policy.
5. From third parties, including service providers, business partners, recruiters, and public sources.

4.1 Connected accounts

Worker works with the tools you already use. You choose which accounts to connect, and we request only the permissions needed for the features you turn on. The table below lists each supported provider, what we access, and what, if anything, we write back. We add a row when we add a provider.

The following commitments apply to every connected account.

What we use it for. Only to provide the features you turn on: today, signing you in, preparing pre-meeting briefs, linking meetings to your weekly meeting agendas, and showing your upcoming meetings inside Worker. We use it for nothing else.

Retention and deletion. Data from a connected account is stored on Worker systems and retained under Section 10: calendar events and invitations as ingested for no more than 30 days, and the briefs and commitment records created from them for the periods stated there. Disconnecting an account stops all further access, and the events and invitations we hold from that account are deleted within 30 days. Briefs and commitment records already created remain available to you under Section 10 and can be deleted by you or your account administrator at any time. Deleting your Worker account deletes all of them.

No model training, no other use. Notwithstanding Section 6, including Sections 6.1, 6.2, and 6.3, data Worker receives from a provider listed in this Section 4.1, and any de-identified or aggregate data derived from it, is used only to provide and improve the features described above. It is never used to develop, train, evaluate, or improve generalized or non-personalized artificial intelligence or machine learning models, whether our own or a third party’s, on any plan and regardless of your account settings, and it is not included in the benchmarks, industry reports, or other de-identified data products described in Section 6.2. We do not sell it, do not use it for advertising, and do not disclose it other than to the sub-processors that process it on our behalf, listed at worker.ai/legal/subprocessors, each of which may use it only to provide services to Worker. Our model providers are contractually prohibited from training on it (Section 6.1).

Revoking access. You can disconnect any account at any time in Worker at My Connections, then Disconnect, or revoke Worker’s access directly with the provider. For Google, go to https://myaccount.google.com/permissions. For Microsoft 365, go to https://myapps.microsoft.com and revoke Worker’s permissions; where your organization approved Worker centrally, your Microsoft administrator can remove its access for your whole organization.

Provider-specific commitments. Some providers require specific commitments as a condition of access. We list them here and add to this list as we add providers.

• Google. Worker’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. The policy is available at https://developers.google.com/terms/api-services-user-data-policy.

5. How we use personal data, and our legal bases

Where the law requires a legal basis, for example in the European Economic Area and the United Kingdom, we rely on the bases set out below. In Switzerland, private controllers do not rely on a statutory list of legal bases; our processing is justified by your consent, by an overriding private interest, or by law, as provided in Articles 30 and 31 of the Swiss Federal Act on Data Protection.

Where we rely on legitimate interests, we have assessed the balance between those interests and your rights and freedoms. You can request a summary of that assessment at privacy@worker.ai.

6. Artificial intelligence: how we use content to train and improve our models

We state this plainly, because it is the question customers ask us most often, and because the answer differs by plan and by whose data it is.

6.1 What we use to train and improve our own models

1. On our free and self-serve plans, we use your content to train, evaluate, and improve our own models by default. This includes Meeting and Commitment Data and the outputs our models generate from it. You can turn this off at any time in your account settings, and turning it off applies from that point onwards. For this purpose Worker acts as a controller, not as your processor.
2. On our paid plans, we do not use your identifiable content to train or improve any model unless you opt in in writing. We do use de-identified and aggregate data derived from your content, as described in Section 6.2.
3. We never use identifiable content about an individual who is not a Worker user to train a model. Before any content enters a training set, records naming individuals outside the account holder’s organization are reduced to organization level or removed. Cross-organization patterns are learned at the level of the company, not the person.
4. We never use voice or other biometric data to train anything, because we do not create it. See Section 3.2.
5. We do not permit our third-party model providers to train on your content. Our contracts with them prohibit it. Every provider that processes customer content is listed, with what it does, at worker.ai/legal/subprocessors, and we will answer questions about it at privacy@worker.ai.

Connected accounts. None of the above applies to data received from a provider listed in Section 4.1. That data is never used to train, evaluate, or improve any model, on any plan, regardless of your account settings. See Section 4.1.

6.2 De-identified and aggregate data

We create de-identified and aggregate data from your content and from your use of the Services, and we use it for any lawful business purpose, including developing, training, evaluating, and improving models, features, and products, and producing benchmarks and industry reports. Our rights in de-identified and aggregate data continue after your account is closed and after your content is deleted, because that data is no longer about you. De-identified or aggregate data derived from a provider listed in Section 4.1 is used only to provide and improve the features described there, and not for the other purposes in this Section 6.2.

Our commitment. Worker maintains and uses de-identified data only in de-identified form. We do not attempt to re-identify it, except solely to test whether our de-identification is effective. We contractually require every recipient of de-identified data to do the same. We publish no statistic drawn from fewer than ten organizations. If we ever combine de-identified data with information that identifies you, we treat the combined data as personal data under this policy.

6.3 Statement for the purposes of state privacy laws

Worker uses personal data to train and improve large language models and other artificial intelligence models, on the plans and within the limits described in Section 6.1, and excluding data received from the connected accounts described in Section 4.1. Worker does not sell personal data for that purpose. Section 18 describes the website analytics and advertising tools that some state laws treat as selling or sharing personal data.

6.4 If we change this

If we decide to use content for a materially different artificial intelligence purpose than the one described here, we will publish the change on this page and notify account owners by email before it takes effect, and we will apply it only to content processed after the effective date of the change. We do not apply new data-use rights retroactively to content collected under an earlier version of this policy.

7. Automated decisions, and what we do not do with commitment records

Worker does not make decisions about you based solely on automated processing or profiling in a manner that produces a legal effect concerning you or that similarly significantly affects you. Our models propose owners, deadlines, and follow-ups; a person decides.

Commitment records are not a performance-management instrument. Worker does not use them to evaluate, score, rank, or rate any individual, and does not provide any feature whose purpose is employee performance measurement or disciplinary evidence. Customers deploying the Services where works council or employee-representative consultation obligations apply are responsible for meeting them.

8. Who can see your data

1. Inside your organization. Account administrators may access, export, and delete content in the account, and may see logs and usage information about authorized users. A pre-meeting brief prepared for a specific host is private to that host. Where you connect a calendar, Worker may add a link to the Worker meeting page to the description of your calendar events; the link itself is visible to the event’s attendees, and the meeting page remains behind sign-in.
2. Outside your organization. We do not send any message to any person outside your organization unless an authorized user of your account specifically approves that message. There is no automatic external outreach, and there is no setting that enables it.
3. Inside Worker. Access to customer content is limited to personnel who need it to operate, support, or secure the Services, is role-based, and is logged.

9. How we share personal data

We share personal data with: service providers and sub-processors that perform functions on our behalf, such as hosting, artificial intelligence processing, analytics, payment processing, and customer support; Google, Meta and LinkedIn, which receive data about visits to worker.ai for advertising and ad measurement as described in Section 18; your own organization and its administrators; our professional advisers; legal and regulatory authorities where required by law, subpoena, or other legal process, or to protect rights, property, or safety; and a successor in connection with a merger, acquisition, financing, or sale of assets, as described in Section 17. A current list of our sub-processors, with the purpose and location of each, is published at worker.ai/legal/subprocessors and is also available on request at privacy@worker.ai.

10. How long we keep data

What deletion means for a commitment record. A commitment record is assembled from more than one source. When you delete a record, or when we act on a deletion request, we delete the source artefacts we hold for it, remove the derived record, and remove it from the per-account view. Where we are required to keep a minimal note that the deletion occurred, we keep only that. De-identified and aggregate data already derived is not affected, because it no longer identifies anyone.

We keep personal data for longer where the law requires it, where we need it to resolve a dispute or enforce our agreements, or while a suspected breach of our Terms is being investigated.

11. Security

We maintain administrative, technical, and physical safeguards designed to protect personal data. Our security program covers security policies and governance, access controls, personnel screening and training, encryption in transit and at rest, endpoint protection, vulnerability management, data management, logging, monitoring and incident response, and third-party vendor management. We will share current details, including the status of any certification, on request at privacy@worker.ai.

No method of transmission or storage is completely secure. If a personal data breach affects your data, we will notify you and, where required, the relevant authority, within the time the law requires and in any event without undue delay.

12. International transfers

Worker is headquartered in the United States, and we and our sub-processors may process personal data in the United States and in other countries. Where personal data is transferred out of the European Economic Area, the United Kingdom, or Switzerland to a country without an adequacy decision, we rely on the European Commission’s Standard Contractual Clauses, the United Kingdom International Data Transfer Addendum, and a Swiss addendum recognizing the Federal Act on Data Protection. You can request a copy of the relevant transfer mechanism at privacy@worker.ai.

Security certifications such as ISO 27001 are assurances about how we protect data. They are not, by themselves, a lawful cross-border transfer mechanism, and we do not present them as one.

13. Your rights

To exercise any right below, email privacy@worker.ai. We respond within the time applicable law requires. We do not discriminate against anyone for exercising a right.

13.1 California residents

• Know what personal information we collect, use, disclose, and sell or share.
• Access and obtain a portable copy of your personal information.
• Correct inaccurate personal information.
• Request deletion of personal information.
• Limit the use and disclosure of sensitive personal information.
• Opt out of the sale or sharing of personal information. See Section 18.

13.2 Residents of other United States

Residents of states including Virginia, Colorado, Connecticut, Utah, Texas, Oregon, Montana, and Maryland have rights to access, correct, delete, and port personal data, and to opt out of targeted advertising, sale, and certain profiling. Where a state law gives you the right to appeal a decision we make about your request, we describe the appeal process in our response.

13.3 European Economic Area and United Kingdom

You have the right to access, rectify, and erase your personal data; to restrict or object to processing, including the right to object at any time to processing based on our legitimate interests; to data portability; to withdraw consent at any time without affecting processing already carried out; and to lodge a complaint with your supervisory authority. Where we rely on legitimate interests to train our models, you may object, and we will stop that processing of your data.

13.4 Switzerland

You have the right to information about the processing of your personal data, to access it, to have inaccurate data corrected, to have data deleted or processing restricted, to receive your data in a portable form, and to object to processing. You may contact us at privacy@worker.ai or contact the Federal Data Protection and Information Commissioner. Where our processing is likely to present a high risk to your personality or fundamental rights, we carry out a data protection impact assessment before starting it, and we consult the Commissioner where the law requires.

14. If you are named in a record but are not a Worker user

If you believe a customer of ours holds a commitment record that names you, contact us at privacy@worker.ai and ask to access, correct, or delete it. Because we hold that record on our customer’s instruction, we will:

1. confirm whether we hold a record naming you and, if we do, identify the customer account that holds it, unless we are prohibited from doing so;
2. route your request to that customer without undue delay and assist them in responding;
3. tell you what we have done and who to contact; and
4. act on the request ourselves where applicable law makes us directly responsible for doing so.

We also make this policy publicly available so that the information required by Article 14 of the GDPR and Article 19 of the Swiss Federal Act on Data Protection is accessible to individuals whose personal data we did not obtain from them directly.

15. Children

The Services are for business use by adults. They are not directed to anyone under 18, and we do not knowingly collect personal data from anyone under 18. If you believe someone under 18 has provided us with personal data, contact privacy@worker.ai and we will delete it.

16. Cookies and similar technologies

We use cookies and similar technologies on our websites. Our Cookie Policy explains which categories we use, what they do, how to control them, and which browser signals we honour.

17. Business transfers

If we are involved in a merger, acquisition, financing, insolvency, or sale of assets, personal data may be transferred as part of that transaction. We will notify you by email or by a prominent notice on our website, and we will tell you what rights you have if the new owner intends to use your data differently.

18. Do we sell or share personal data

Worker does not sell personal data for money. On worker.ai we use Google Analytics with Google signals and advertising features turned on, the Meta Pixel and Meta’s Conversions API with Meta’s automatic event detection and automatic advanced matching turned on, and the LinkedIn Insight Tag and LinkedIn’s Conversions API. They let Google, Meta and LinkedIn use data about your visit, including a hashed copy of any email address you enter in our forms, for advertising and ad measurement (Section 3.4 and our Cookie Policy). Under the laws of California and some other states, that can count as selling or sharing personal data, or as targeted advertising, so we treat it that way and give you a way to opt out: use “Do not sell or share my personal information” at the bottom of any page on worker.ai, or turn on Global Privacy Control in your browser. Either one takes effect at once for that browser: the Meta Pixel and the LinkedIn Insight Tag stop loading and our server stops sending your visits to Meta and LinkedIn. Turning advertising off with the link also stops both at once on the pages you have open. You can also set how Meta uses data for ads at facebook.com/adpreferences, set how LinkedIn uses data for ads in your LinkedIn account settings, under Advertising data, and opt out of interest-based ads from many companies at optout.aboutads.info.

We do not sell or share personal data from the Services, including Meeting and Commitment Data and data received from the connected accounts in Section 4.1, and we do not knowingly sell or share the personal data of anyone under 16. We do commercialise de-identified and aggregate data as described in Section 6.2; under California law, de-identified and aggregate consumer information is not personal information.

19. Changes to this Privacy Policy

We may update this Privacy Policy. If a change is material, we will post a notice on our website and email account owners before it takes effect. A change that expands how we use data already collected applies only to data processed after the effective date of the change, unless we obtain your consent. Your continued use of the Services after a change takes effect means you accept the updated policy.

20. How to contact us

privacy@worker.ai or by phone (415) 907-4709

If you are dissatisfied with how we have handled a request, you may contact your local data protection authority.