PROVEAiBLE is designed to run your client intake on your organisation's own AI provider account — a direct relationship with Anthropic (Claude), not one filtered through a vendor. When you connect your own key, every AI call is made under your account and your provider's terms, your organisation is the data controller, and we never mark up a token.
Your organisation's AI account — for example Anthropic (Claude) — with your own API key. It stays yours; we never hold it or see your account.
Every AI call is billed by your provider, under your terms. Identifiers are replaced with role tags before the case-analysis model sees them — so even your own model never works from who the client is.
The intake is transient. The completed file is deleted on delivery, and — when you configure your own key — the processing happened in a relationship you control, not a vendor's shared account.
You bring the key. PROVEAiBLE connects to your account — it never stores the key or logs in for you.
Your provider account, your key.
Under your account and your terms.
Transient, deleted on delivery.
When intake runs on a vendor's shared AI account, your clients' words are pooled in someone else's relationship with the model. On your own key, the conversation stays inside the account you control — under your provider's terms, not a middleman's.
of US adults say the growing use of AI will make their personal information less secure.1
Pew Research Center, 2026of employees admit entering non-public corporate or personal data into public GenAI tools.2
Cisco Data Privacy Benchmark, 2025global average cost of a data breach — the exposure a shared AI account quietly adds to.3
IBM Cost of a Data Breach, 2025On your own key, the client's story is processed under your account — not held in a vendor's pool.
The difference BYOK makesControl over where data lives — and whose account it lives in — is now a primary trust factor in AI decisions. BYOK puts your clients' data on the account you already trust.2
Cisco Data Privacy Benchmark, 2025Stays in your relationship. The story is processed on your own account — not pooled in a vendor's shared one.
Redacted before the model reads it. Identifiers become role tags before the case-analysis model sees them.
Nothing kept afterward. The completed intake is transient and deleted on delivery; incomplete intakes are purged after 90 days.
A generic bot runs on a vendor's account and asks generic questions. PROVEAiBLE runs on yours — and it's built around the work you take, the way you qualify it, and the rules where you practise, layer on layer.
Runs on your own AI account and the model you choose. No vendor lock-in — change provider or model without re-platforming your intake.
The work you actually take — not a menu of everything. It asks the questions your practice asks, in your terms.
Tuned to how you prioritise, your auto-decline criteria and the rules where you practise. An organisation across several countries gets one intake that adapts to each.
Configured hands-on at onboarding — connected to your account and tuned to your practice before a single client sees it.
That's the difference between an AI that runs in your name and a chatbot that runs in someone else's.
The intake is yours, not welded to one vendor's account. Change provider or model and it keeps running — you're never trapped by someone else's roadmap or repricing.
No vendor holds your account hostage.AI usage is billed by your provider directly, at their published rates — we never mark up a token or sit in the middle of your AI spend. You see exactly what the AI costs, on your own bill.
Pay your provider, not a reseller.When you configure your own key you're the controller, on terms you already hold — a cleaner answer for your DPA, your regulator and your clients about where their data goes.³
Source: IBM Cost of a Data Breach, 2025 See the compliance detail →See exactly how bring-your-own-key fits your data-protection paperwork on our compliance page →
PROVEAiBLE's intake is designed to run on your organisation's own AI provider account — for example Anthropic (Claude). You connect your own API key, and every AI call is made under your account and your provider's terms. We never mark up AI usage: you pay your provider directly, and we never hold your key or bill you for a token.
The intake conversation is processed through your own AI account. Identifying details are replaced with role tags before the case-analysis model sees them, and PROVEAiBLE keeps no copy of the completed intake — it's transient and deleted on delivery, with incomplete intakes purged after 90 days.
When your organisation configures its own key, the AI processing happens under your provider relationship — so your organisation is the controller of that data, rather than a vendor pooling it in a single shared account.
No and no. We don't train models on your data and we never mark up a token. Leading providers such as Anthropic also state that, by default, they do not use inputs or outputs from their commercial products (including the API) to train their models.
Bring-your-own-key is how PROVEAiBLE's production intake is designed to run — it's supported in the platform, and we're rolling it out with organisations as we launch. It's the model the product is built around; request a demo and we'll walk you through connecting your own account.
I'll send you a private demo and walk you through connecting your own AI provider account. Or see the full product →
Free, no card, no call.
