"Private" means two things at once. At the application layer, every client sits inside their own ringfenced vault — one engagement's context can never bleed into another's. At the data layer, that isolation runs all the way down: client data is stored and encrypted separately, not just logically separated in a shared table. Combined with AES-256 encryption at rest, TLS in transit, and a policy that client data is never used to train any model, that's what "private" means in PAL.

No — and that's deliberate. PAL isn't a model, it is a private infrastructure that gives you access to every leading AI model in one place: Claude, GPT, Gemini and Mistral, as well as Perplexity, Grok, Llama, DeepSeek and Qwen. You get frontier intelligence without betting your firm's IP on any single vendor's roadmap.

Claude, ChatGPT and Gemini are single-vendor tools — you're locked into one company's models, roadmap and pricing. PAL routes every task to whichever model is best suited to it, and if one provider has an outage or falls behind, PAL switches seamlessly to another — you'd barely notice. On top of that, PAL adds what none of them offer: ringfenced client vaults, so your client data stays private and compartmentalised by engagement, not pooled into one shared account history.

No. All model calls are made through stateless APIs. Client data is never used to train any model. PAL already has Zero Data Retention (ZDR) confirmed with Mistral, with a ZDR agreement for Gemini in progress.

A client vault is a ringfenced, isolated workspace created for each of your clients. All documents, context and outputs generated for that client live inside their own vault — completely walled off from every other client, so nothing can cross-contaminate between engagements even when you're working across dozens of accounts. Access to each project in each client vault is controlled by the your firm admin - so no consultant can scroll through a project information unless authorised

Absolutely! PAL offers UK GDPR / EU GDPR compliance pack, per-client 'allowed-models' and Data Processing Agreement, out-of-the-box. PAL is built with the EU AI Act's Article 50 transparency obligations in mind. Your data resides in EU. Now you can walk into every client pitch with a compliant AI posture from Day 0.

Yes. PAL offers UK GDPR / EU GDPR compliance pack, per-client 'allowed-models' and Data Processing Agreement, out-of-the-box. Now you can walk into every client pitch with a compliant AI posture from Day 0.

Yes — that's exactly what it's for. The compliance pack (DPA, sub-processor register, architecture diagram, plain-English disclosure) is written to be handed straight to your client's IT or legal team for review, so you can get sign-off to use AI on their engagement without building anything from scratch yourself.

No. Unlike self-hosted private AI infrastructure, PAL is a client-compliance-ready product you launch and use — no architecture to build, no servers to maintain, no compliance documents to draft yourself.

Absolutely! We understand that a modern consultant can work in a boutique or work across multiple teams independently. So we built PAL so you can form teams with members or your firm as well as with external advisors who join your project temporarily under their own email ids. The AI usage allowance is pooled. PAL has role-based permissions, so consultants, project leads and firm admins can collaborate within the same client vaults without losing the ringfencing that keeps client data isolated.

Your work stays with you — nothing is auto-deleted, so the institutional knowledge you built stays usable for future engagements. If a client requires full erasure, you can request deletion of that vault, and PAL will issue a formal deletion certificate you can pass on to them.