Frequently asked questions

We deliver enterprise-grade AI with human oversight built in. Our products include YAPA (an AI personal assistant), PALS (an agentic AI workforce), Pocket CRM & HR Admin, and the AI Delivery Pod (a no-code platform for building production software). We also offer a software development service.

In short:

• YAPA, an AI personal assistant for tasks, CRM actions, leave, and expenses.
• PALS, an agentic AI workforce for building and deploying AI agent teams.
• Pocket CRM & HR Admin, a lightweight CRM and HR admin tool.
• AI Delivery Pod, a no-code platform that builds production websites, apps, and API services.
• Software development, delivered as a service.

Individuals can sign up directly to the relevant product. Enterprises typically start with a short scoping conversation and an Order Form. To begin, email enquiries@tangent9.com or use the contact option on the product page.

Everything lives in our resource centre, including our Privacy Policy, Terms of Service, and AI Governance Framework.

Singapore law governs all current deployments, and our data handling follows the Singapore Personal Data Protection Act 2012 (PDPA). See our Terms of Service and Privacy Policy for detail.

Email support@tangent9.com and a member of the team will help. For privacy matters you can also reach our Data Protection Officer at dpo@tangent9.com.

Our platforms use passwordless sign-in: a Magic Link, a One-Time Password (OTP), or WebAuthn biometric verification. You may be asked to re-verify periodically. If you think someone else has accessed your account, email support@tangent9.com immediately.

YAPA is a conversational assistant for everyday work: managing tasks, taking CRM actions, submitting leave and expenses, scanning business cards and receipts, and researching organisations. Extracted details are flagged for you to verify before they are saved.

PALS lets your organisation assemble AI agents from skills and tools, connect them to authorised systems, schedule them to run unattended, and chain them into multi-step workflows. Consequential steps require human approval, and unattended agents should be reviewed regularly.

No. AI outputs are probabilistic and can contain errors, omissions, or hallucinations, and the same input can produce different results. Treat them as drafts and recommendations, and review and verify anything before you act on it. See our AI Governance Framework.

You stay in control. Every action is classified by risk tier. Low-risk actions can proceed automatically, Medium actions need your explicit confirmation, and High or Critical actions (sending messages, changing or deleting records, moving money, changing permissions) require human approval. These controls are not designed to be bypassed.

They use OCR to extract contact or expense details from a photo. Extracted fields are labelled "AI extracted, please verify" and you must confirm them before saving a contact or submitting a claim.

The platform is a Progressive Web App (PWA) you can install on a phone or desktop. It needs an active internet connection: data lives on our servers and is not stored offline. Uninstalling the app does not cancel your subscription or delete your data.

Through the connector catalogue, agents can be granted access to authorised systems such as Gmail, Outlook, Slack, Microsoft Teams, Salesforce, HubSpot, Jira, Zendesk, GitHub, Notion, and Stripe, subject to your configuration and authorisation. Connector credentials are held in a trusted backend and are not exposed to the AI layer.

Pocket CRM & HR Admin is a lightweight tool that brings core CRM and HR admin, such as contact records, approvals for leave and expenses, into one place, with the same human-approval safeguards as our other products.

Our AI features run on third-party large language models from Anthropic (Claude) and OpenAI (GPT), as configured. Provider model updates can change AI behaviour; we monitor for material changes but cannot guarantee identical behaviour across versions.

No. We do not use personal data to train external AI models without explicit consent, and we do not sell or rent personal data. See how we use personal data.

Personal data detected in inputs, such as NRIC numbers, payment card numbers, or API keys, is designed to be masked before anything is sent to an LLM provider. We also use role-based access control, credential isolation, log masking, and encryption in transit and at rest where appropriate. Full detail is in our Privacy Policy.

Data may be processed or stored outside Singapore, for example through LLM provider APIs and cloud infrastructure. Where data is transferred abroad, we take steps to ensure a comparable standard of protection consistent with the PDPA. See cross-border data transfers.

In most deployments the customer organisation is the data controller and Tangent9 acts as data processor. Requests about data held in a customer deployment should go to that organisation first. We are the controller for data we hold independently, such as account and contact details.

If your organisation is the controller, direct the request to them. Where Tangent9 is the controller, email dpo@tangent9.com. We respond within 30 days of receipt. See data subject rights.

Yes. Inputs, outputs, actions, and approvals are recorded in an audit log for security, debugging, governance, and incident response, with sensitive fields masked. This is what makes actions traceable and reviewable.

Yes, our products are designed in line with the PDPA and Singapore's IMDA AI governance frameworks (the Model AI Governance Framework and its Generative AI and Agentic AI editions). Our AI Governance Framework explains how our orchestration layer maps to them.

The PA Platform (YAPA) is offered under Starter, Professional, and Trial tiers. The features in each plan are shown on the pricing page and may change over time. Enterprise use of PALS is arranged through an Order Form.

A 14-day free trial may be offered to new subscribers. At the end of the trial your account either converts to a paid plan or is suspended, based on your choice at sign-up, and we notify you 3 days before it ends.

Subscriptions renew automatically at the end of each billing period (monthly or annual). You authorise us to charge your nominated payment method at the start of each renewal period.

Cancel any time in your account settings. Cancellation takes effect at the end of the current billing period, and you keep access until then. Contact admin@tangent9.com if you wish to cancel your subscription.

Fees are generally non-refundable, except where required by Singapore consumer protection law (for example, under the Consumer Protection (Fair Trading) Act) or as stated in your plan. See the Terms of Service.

We may change pricing with at least 30 days' notice. Continuing to use the service after a change takes effect means you accept the new pricing.

Email partners@tangent9.com with a short note about your business and the markets you serve, and we will share the partner onboarding details. Programme terms are set out in our Partner & reseller programme and your partner agreement.

Our platforms support tenant-branded instances, so the app icon and branding your customers see can reflect your configuration. The exact white-label scope is defined in your partner agreement.

This is agreed in your partner agreement. Typically resellers provide first-line support to their own customers, with Tangent9 backing you up on platform and infrastructure issues. We can support a co-branded or escalation model.

Reseller pricing, discounts, and margin are set in the partner agreement and any Order Form, and depend on tier and volume. Contact partners@tangent9.com to discuss commercials.

Resellers are expected to ensure their customers accept our terms, configure deployments safely (including least-privilege agent and connector setup), maintain an acceptable use policy consistent with ours, and pass on relevant product and safety guidance. These flow from our Terms of Service.

Yes. Partners receive onboarding and enablement materials so your team understands the products, their AI features, and the human-oversight controls before going to market. Details are arranged during onboarding.

The AI Delivery Pod is a no-code platform that builds production websites, apps, and API services from plain-language descriptions, with real engineering rigour: your code lives in your GitHub repository and runs in your own cloud across Dev, UAT, and Production environments.

Yes. Alongside the AI Delivery Pod, we provide a software development service for cloud and AI-native applications. The right approach depends on your needs; we can advise on which fits.

You retain ownership of the content you submit, and delivery work is built into your own repository and cloud. Intellectual property in our own platform, software, and models remains ours. See intellectual property for specifics.

In your environment: your GitHub repository and your cloud account. The model is designed to avoid lock-in, so you keep control of what is delivered.

Engagements are scoped through an Order Form or statement of work. Service levels, where they apply, are set out in the Order Form or a separate Service Level Agreement.