Drawn from real conversations with UK firms in early access. If yours isn't here, email hak@vpnetworks.co.uk and we'll add it.
Entirely on your firm's hardware. Susan runs as a Docker stack on a server you own (or a private VPS you control). No client data is sent to any cloud unless you explicitly enable an external AI provider, and even then it's redacted of common UK PII before leaving your network.
Susan implements SRA Accounts Rules controls including Rule 4 reconciliation, Rule 7 client interest, Rule 12 accountant report exemption logic, AML gates on matter creation, conflict-of-interest checking, and a 7-year audit log retention policy. It is your responsibility to operate the system in accordance with SRA Standards and Regulations.
For a small firm (1–10 users) running the on-server AI: a workstation or small server with 32GB RAM, 1TB SSD, and a recent NVIDIA GPU (RTX 4060 or better). For larger firms, or for cloud-AI mode with Ollama disabled: any modern Linux server or Windows machine running Docker Desktop with 16GB RAM and 500GB SSD.
Yes. Susan supports four AI providers: on-server Ollama (default — fully private), Groq, Anthropic Claude, and OpenAI. External providers require explicit consent in settings. When external AI is enabled, prompts are scrubbed of UK-specific PII (NI numbers, postcodes, sort codes, phone numbers, emails) before being sent, and every external call is logged to the audit trail.
£35 per user per month, billed annually. No hidden fees, no per-feature unlocks, no AI usage charges (the on-server AI is included). Optional hardware costs £6,000–£22,000 one-time depending on firm size; the alternative is a UK-based VPS for £400–£2,500 per month.
We're running a Pilot Programme for the first cohort of customers — a 6-month engagement at 50% off the published price in exchange for structured feedback and a written reference.
From a clean machine with Docker installed, a fresh setup takes 15–45 minutes — most of it is the AI model download. The setup script asks three questions (firm name, admin email, admin password), generates all secrets, and brings everything online. You can be drafting your first matter the same day.
Updates are pulled from a private Docker registry. You receive a one-line update command (docker compose pull && docker compose up -d). Your data, settings, and uploaded documents are never affected by updates — only the application image is replaced.
Your data is yours. Susan provides full database export tools and an encrypted backup format you can restore on any machine. We have no commercial incentive to lock you in — the architecture is intentionally portable.
No. Susan is a paid commercial product. The Docker image is private. We chose this model so we can charge fairly and continue developing the product without diluting professional support.
Yes — Susan has a built-in branding studio with 8 colour themes, 5 document header styles, 4 typography pairings, custom firm logo on all generated documents, configurable footer (address, phone, email, SRA number, VAT number), and per-user email signatures. Point and click — no design skills required.
Yes. Susan connects to any IMAP mailbox (Outlook, Gmail, Microsoft 365, Fastmail, etc.) for incoming email, and any SMTP server for outgoing. Calendar events generated by Susan can be synced via standard ICS feeds. See all integrations →
Yes — with a Twilio account configured, Susan can answer incoming calls with a custom voicemail greeting (read by the AI voice), transcribe the caller's message, summarise it, detect intent, match it to an existing contact if possible, and create a follow-up task on the matter automatically. This is optional and requires a separate Twilio subscription.
No. Susan does not fine-tune any AI model on your data. The on-server AI uses a pre-trained model (Llama 3.2 by default) that runs locally and treats each request as stateless. Your data is never used to train any model — yours, ours, or anyone else's.
Email hak@vpnetworks.co.uk or book a 30-minute demo and we'll answer them in person.
Book a demo →