OCG-compliant AI routing for law firms.
Restricted matters stay on your hardware. Everything else uses the best public model. One policy engine decides which, for every prompt.
Free for the first 20 Am Law 51-200 firms. In two weeks, the audit maps which client guidelines your current AI tools already violate, matter by matter.
Your clients have already decided where AI can go
Latham built its own AI stack: 900 technologists, a hundred of them on AI, a colo cage, and three years. After Latham’s build made the news, managing partners started asking their CIOs what their version is.
You don’t have a hundred ML engineers, and your clients’ outside-counsel guidelines apply to you anyway.
Some of those guidelines prohibit third-party AI on the client’s matters outright; a vendor DPA doesn’t satisfy them. Most firms can’t tell which matters are covered, and almost none can enforce the rule at the point of use.
A better or cheaper model won’t fix this, because compliance depends on where each matter’s data goes. Matterwall controls that routing and leaves you free to change AI vendors.
Why not just…
…use Harvey / ChatGPT Enterprise?
Keep using them on unrestricted matters. When a matter is restricted, Matterwall routes it to your private stack instead, so your lawyers can still use AI on it.
…buy two GPU boxes and hire an engineer?
You can. Then you own model selection, quarterly refreshes, eval harnesses, OCG mapping, DMS integration, and the on-call rotation. One engineer can't cover all of that, and Matterwall's team does.
…wait for the labs to offer on-prem?
When they do, their model becomes one more route in the router, because Matterwall works with any vendor. Your private stack gets cheaper, and the policy layer matters more with more routes to choose from.
One policy engine, not a rack of boxes
Matter-aware router
The policy engine checks each prompt's client, matter, and OCG terms. Restricted work goes to your private stack, and unrestricted work goes to whichever frontier API is best that week. Lawyers use one interface. Compliance gets one log that records each routing decision against the OCG clause behind it.
Private stack, refreshed quarterly
The latest open-weight frontier model on hardware dedicated to your firm: a colo cage, a dedicated cloud tenancy, or a DGX Station on-prem. We benchmark every release against the public frontier on legal task suites, so you know precisely what you're trading for privacy. Today that's about three months behind the best public models on most tasks. When the gap is wider, we'll tell you.
Matter walls at retrieval
Retrieval runs against iManage and NetDocuments and enforces matter walls as it searches, so the model can only see documents the lawyer asking could already open.
Firm-tuned models
Fine-tune models on your own work product. Latham argues that no AI company can replicate its lawyers, and that should be true of your firm too.
Security
- Firm-dedicated compute
- No shared tenancy
- No training on your data, ever
- Matter walls at retrieval
- Full audit log per prompt, per matter, per OCG clause
- Deployable inside your colo, your cloud account, or your building
Your hardware, our team
We never put GPUs on our balance sheet or bury their cost in your subscription. Your firm buys or leases the hardware directly, and we quote all three fees after the audit.
Stack live in weeks rather than quarters
Managed operations: refresh, eval, monitoring, on-call
Router seats across the firm
You keep the asset. We keep it current.
Partners
Matterwall is built with NVIDIA, integrates with iManage and NetDocuments, and deploys with major colocation providers.
What the audit finds
We read every client's outside-counsel guidelines, map each AI clause to the matters it governs, and check them against the AI tools your lawyers already use. You get one table your GC and CIO can both work from.
Find out which clients you're already out of compliance with.
The OCG audit is a $25K engagement, free for the first 20 Am Law 51-200 firms while we build the benchmark. In two weeks you get a matter-by-matter map of AI restrictions across your client base and, usually, the budget line to fix it.