Privacy Policy

“Pannoesis Technologies”, “Pannoesis”, “we” and “us” refer to Pannoesis Technology One-Member Private Limited Company, a company registered in Ethiopia with its head office at Summit, Lemi Kura Sub-City, Addis Ababa, Ethiopia.

This policy covers:

• this website and the Markdown and text copies of it published for search engines and assistants;
• our products and services — the CaseFlow platform, its managed agents and its connectors, including the Pannoesis Google Workspace connector (together, the “Services”);
• our sales, support, recruitment and other business communications.

It does not cover services run by others that you connect to the Services — Google Workspace, a model provider your organization selects, or any other third party. They handle data under their own policies, which you should also read.

A customer-hosted deployment sends nothing to Pannoesis. We receive information from it only when the Customer sends it to us — for example, in a support request.

If you use the Services through your employer or another organization, that organization’s privacy notice explains how it uses your information. Send requests about Customer Data to it; we will help it respond.

We do not decide what Customer Data contains. The Customer is responsible for having a lawful basis to process it. Sensitive categories of personal data — such as health, biometric or criminal-record data — are processed only where the Customer’s agreement and applicable law allow it and the Customer has configured the corresponding access controls.

Where an audit record concerns Customer Data or Google user data, it records the tool, the time, the outcome and a summary, not the content itself.

• No cookies or local tracking. The site sets no cookies. It stores one item in your browser — your light or dark display choice — and only if you change it.
• No analytics, advertising or trackers. We do not measure visits with any analytics tool and load no third-party scripts, pixels or embeds.
• Fonts served from our domain. Typefaces are served by this site, so your browser makes no request to a font service.
• Request data. Our hosting and network provider, Cloudflare, processes your IP address, browser user agent, the page requested and the time, to deliver the pages and protect them from abuse. We do not combine it with anything else to identify you.
• Strict browser security policy. The site tells your browser to load only its own resources and never to embed it in another site.
• Links elsewhere. Links to other sites — such as Google’s policies — take you to services with their own privacy practices.

We do not use personal information for any purpose incompatible with the one it was collected for.

1. Your instructions only. We process Customer Data only to provide, support and secure the Services, as the Customer’s agreement and documented instructions require.
2. No standing access. Customer Data is processed inside the Customer’s Deployment. Pannoesis personnel have no standing access to it and do not view it in the ordinary course of business. Access happens only under the procedure in section 8; where strictly necessary to investigate an attack on or abuse of the Services, with the access logged and the Customer told; or when the law compels it under section 18.
3. No training. Except as point 4 allows, we do not use Customer Data — content, files, prompts, outputs or connected records — to train, fine-tune, evaluate or otherwise improve any artificial-intelligence or machine-learning model, whether ours or a third party’s.
4. Customer-owned adaptation. At the Customer’s own written instruction, a model may be adapted with its Customer Data inside its Deployment. The result serves only that Customer, stays in its Deployment and is deleted with its data. Google user data is never used for this.
5. No sale, no advertising. We do not sell, rent or trade Customer Data, use it for advertising, or share it with data brokers.
6. Confidential personnel. Everyone who works for us is bound by written confidentiality obligations that outlast their engagement.
7. Export and deletion. Customers can export their data at any time during the agreement and for 30 days after it ends. Within 30 days after that export window closes, we delete Customer Data held in any environment we operate — or return it, if the Customer asks — and purge it from backups within a further 60 days. If the law requires us to keep something, we keep it isolated and use it for nothing else.

1. You ask. An administrator of the Customer opens a support request that cannot be solved without looking at Customer Data.
2. You grant in writing. The administrator authorizes access in writing, naming the request, the data or screens concerned, the people who may look and how long the grant lasts. If the request involves Google user data, the Authorized User whose account it is must also agree in writing, naming the specific messages or files.
3. Least access. We use only what the grant names, with an individual account for each person — never a shared one.
4. Everything is logged. Each access is recorded in the Customer’s audit log: who, when, what, and why.
5. It ends on time. Access is removed when the request is resolved or the grant expires, whichever comes first.
6. You can stop it. The Customer can revoke a grant at any moment.

• Model choice. The Customer chooses which model deployments its agents may use, from the governed catalogue its Deployment offers. The providers available to a Deployment are listed in that catalogue, and we provide the list on request.
• Provider terms. Where the Customer selects a third-party model provider, data sent to it to answer a request is governed by the terms between the Customer and that provider, or by the terms we disclose for providers we contract. Google user data is sent only to model deployments that are self-hosted through Model Studio, or run by providers contractually prohibited from retaining it after responding or using it for training.
• Proposals, not decisions. Outputs from models and agents may be incomplete or wrong. They are presented as proposals and evidence, with their sources, and are not a substitute for professional judgment.
• A person approves effects. Agents use only the tools a Customer attaches to them. Tools that send, change, publish or approve wait for an authenticated person’s approval.
• Untrusted content stays untrusted. Text in a connected email or document is treated as data, so instructions hidden in it cannot grant an agent new authority.
• Synthetic stays labelled. Simulated or synthetic data carries its label on every screen and in every record that uses it.
• No automated decisions about you. Pannoesis makes no decision about you that is based solely on automated processing and produces legal or similarly significant effects. Customers use the Services to support their own decisions, and our Terms require meaningful human review for such decisions.

The Pannoesis Google Workspace connector lets a Customer’s CaseFlow agents work with Gmail, Google Drive, Google Calendar, Google Docs and Google Slides for an Authorized User, after that person signs in with Google and grants access on Google’s consent screen. It acts only for the Google account that granted access.

The permissions we request, and why

We request only the permissions for the services an administrator enables; Google lists them on its consent screen before you agree. Each scope below begins with https://www.googleapis.com/auth/.

How we use it

• Only to provide the features you use in the agent interface: answering your request, preparing the change you asked for, and carrying it out once you approve.
• An administrator can run the connector in read-only mode, which blocks every change.
• Every action that changes your Google data — sending or drafting mail, changing labels, creating or editing files, documents, presentations or events, or changing sharing — runs only after you approve that action.

How we store and protect it

• Content retrieved from Google is processed to fulfil your request. It is kept only where it becomes part of a conversation or work product you keep, and then it is stored in your organization’s Deployment as Customer Data, under its retention settings.
• OAuth tokens are encrypted at rest in the Deployment’s secret store, never placed in an agent’s context and never written to logs.
• The connector’s audit log records the tool used, the time and the outcome; it summarizes message and document content rather than storing it.
• All transfers use encrypted connections (TLS).

When it is shared

We do not transfer Google user data to anyone except:

• the AI model deployment your organization has configured, only as needed to perform the task you requested, and only if it is self-hosted through Model Studio or run by a provider contractually prohibited from retaining it after responding or using it for training;
• with your explicit consent;
• where necessary for security purposes, such as investigating abuse;
• to comply with applicable law; or
• as part of a merger, acquisition or sale of assets, and only after obtaining your explicit prior consent.

What we never do with it

• Use it to develop, train or improve any AI or machine-learning model, generalized or personalized, ours or anyone else’s.
• Use, transfer or sell it for advertising, including retargeting and interest-based advertising.
• Sell it, or transfer it to data brokers or information resellers.
• Use it to determine creditworthiness or for lending purposes.

Who can read it

No Pannoesis employee reads your Google user data except:

• with your own written agreement naming the specific messages or files, given as part of a support grant under section 8;
• where strictly necessary to investigate an attack on or abuse of the Services, with the access logged;
• where the law compels it, under section 18; or
• where the data has been aggregated and anonymized for internal operations.

Keeping, revoking and deleting it

• Disconnect the account in CaseFlow, or remove access at any time at myaccount.google.com/permissions. After revocation, the connector can no longer reach your Google account.
• OAuth tokens are deleted when you disconnect, or when Google tells us access was revoked.
• Google Workspace administrators can also restrict or block the connector for their whole organization from the Google Admin console.
• To have Google user data deleted from conversations and work products, delete them in CaseFlow, ask your administrator, or email contact@pannoesis.com.

Limited Use

Pannoesis Technologies’ 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. Our use of information received from Google Workspace APIs also adheres to the Google Workspace API User Data and Developer Policy, and we do not use Google Workspace API data to develop, improve or train AI or machine-learning models.

• We keep a current list of sub-processors for Customer Data, provide it on request, and notify Customers in advance of adding or replacing one, so they can object.
• In a merger, acquisition or sale of assets, information may pass to the successor with notice to you — and, for Google user data, only with your explicit prior consent.
• We do not sell personal information, and we do not share it for advertising.

• We are based in Ethiopia. Account Information is processed in Ethiopia and by the providers in section 11, some of which operate internationally.
• Customer Data is processed where the Customer’s Deployment runs, which the Customer chooses.
• Where Ethiopian law requires personal data to remain in Ethiopia, we comply, and we help Customers configure their Deployments to comply.
• Where personal data crosses borders, we rely on the safeguards applicable law requires, such as contractual protections.

Technical controls

• Encryption in transit (TLS) for every connection to the Services and this website.
• Authenticated access with roles and per-person clearance, checked on every read.
• Isolation between organizations in shared Deployments; the same ownership checks in dedicated ones.
• Connectors installed as immutable releases pinned by cryptographic digest; a mismatch fails closed.
• Secrets injected only at launch — never packaged with a connector or exposed to an agent.
• Audit records of sign-ins, changes, actions, approvals and agent tool calls.

Organizational controls

• Least-privilege access for our own personnel, with no standing access to Customer Data.
• Written confidentiality obligations for everyone who works for us.
• Reviewed, tested changes and a single controlled release path.
• An incident-response process, described in section 14.

No system is perfectly secure. More detail is on our Security page, including how to report a vulnerability.

• If we become aware of a breach affecting personal data we process, we notify the affected Customer without undue delay — and, where feasible, within 72 hours.
• Our notice describes what happened, the data and people likely affected, what we have done, and what the Customer can do.
• Where the law requires it, we notify the supervisory authority and the people concerned, or help the Customer to do so.
• We keep a record of every incident and the steps taken.

Under Ethiopia’s Personal Data Protection Proclamation No. 1321/2024, and other data-protection laws where they apply to you, you may have the right to:

1. Email contact@pannoesis.com with the subject “Privacy request”, saying which right you want to use.
2. We may ask for information to confirm your identity, and use it only for that.
3. We respond within 30 days. If a request is complex, we tell you within those 30 days why we need longer and by when you will hear from us.
4. If we cannot do what you ask, we explain why and tell you how to challenge our answer.
5. If your request concerns Customer Data, we pass it to the Customer responsible and help it respond; it decides.
6. You may act through an authorized representative, who must show their authority.

• We disclose information to an authority only when legally compelled to.
• We check each request for legal validity and challenge requests that are unlawful or overbroad.
• Because we hold no standing access to Customer Data, we first direct the authority to the Customer.
• Where we must disclose, we disclose only the minimum required, and notify the Customer beforehand unless the law prohibits it.
• We keep a record of every request we receive and how we answered it.

• We send Customers’ administrators the service, security and legal notices their agreement requires.
• We send news or event invitations only to people who asked for them; every such message has a way to stop them.
• We never send marketing through a Customer’s connected accounts.

The website and the Services are not directed to anyone under 18, and we do not knowingly collect personal data from children. If you believe a child has given us personal data, contact us and we will delete it.

• When we change this policy, we update the version and effective date at the top.
• For a material change, we notify Customers’ administrators at least 30 days before it takes effect.
• A change never reduces the protection of information collected before it without your consent.
• Earlier versions are available on request.

Pannoesis Technology One-Member Private Limited Company
Summit, Lemi Kura Sub-City, Addis Ababa, Ethiopia
contact@pannoesis.com

If you are not satisfied with our answer, you can complain to the Ethiopian Communication Authority, the supervisory authority for personal data protection in Ethiopia, or to the authority where you live.