Legal AI architecture built around the work you need to complete
External models: Available. Legal LLM/RAG: Research Programme — internal engineering; no customer use or published evaluation.
Research Programme Tools: Available. BI/Matter & Assurance: Early Access — synthetic-data walkthroughs only. LLM/RAG: Research Programme — internal engineering; no customer use or published evaluation.
After signing in, open Support → New support request and choose General / Query. Quote "Architecture briefing".
Three core layers with distinct jobs
The language model layer
Lightweight legal LLM: Research Programme — internal engineering; no customer use or published evaluation. Current products use external models.
The legal RAG layer
Legal RAG: Research Programme — internal engineering; no customer use or published evaluation. Retrieval cannot guarantee accuracy.
The proprietary legal layer
Existing playbooks, review rules and document structures: Available inside supported products.
Explore the workflow step by step
-
Confirm the agreement type, governing law, review purpose and user role
Set the context for the review.
-
Select a supported review
Choose commercial risk, clause comparison or another supported workflow.
-
Retrieve permitted document passages
Load approved legal context relevant to the task.
-
Apply the relevant playbook
Prepare a structured issues list based on the legal layer.
-
Show clause location and source support
Present questions and suggested next actions with evidence.
-
Route to professional review
Send work to the appropriate user or professional before external use.
How the layers connect
| Layer | Purpose and customer benefit | Delivery position |
|---|---|---|
| User workspace | Start with the work at hand | Available — supported features and plan limits apply. |
| Intake and scope | Reduce assumptions about the task | Supported features: Available. Extensions: Research Programme — architecture briefing only; customer use excluded. |
| Model access and orchestration | Support defined legal tasks and prepare structured responses | External models: Available. LLM/RAG/routing: Research Programme — internal engineering; no customer use or published evaluation. |
| Legal retrieval | Make evidence easier to inspect | Research Programme — internal engineering; no customer use or published evaluation. |
| Proprietary legal controls | Make repeatable work more consistent | Supported features: Available. Extensions: Research Programme — architecture briefing only; customer use excluded. |
| Validation and evaluation | Surface issues before reliance | Release checks vary; legal LLM/RAG: Research Programme — internal engineering; no customer use or published evaluation. |
| Professional review | Keep professional decisions visible | Early Access — synthetic-data walkthroughs only; customer rollout planned. |
| Connected business services | Carry work forward in the platform | Supported tools: Available. Assurance: Early Access — synthetic-data walkthroughs only; customer rollout planned. |
A legal and business platform with an existing foundation
A legal and business platform with an existing foundation.
Dogetlawyer's existing platform includes document and contract tools, CRM, HR, billing, accounting, project management and CLM-related functionality. Early Access walkthroughs (synthetic data only; not yet in customer use) illustrate legal-assurance and BI connections, preserving the records held by each module.
That creates practical expansion paths: a contract issue can become a review task; an approved contractual obligation can support renewal tracking; an invoice record can provide context for an authorised follow-up. Each new connection must be validated for its actual workflow before being described as available.
For investors and ecosystem partners
Building reusable legal intelligence around an existing software platform.
Research Programme — internal engineering; no customer use or published evaluation. Request an architecture briefing through Support.
Architecture built for products people buy
Architecture attracts technical interest; repeatable outcomes create commercial viability. Each technical layer connects to a paid workflow that is available or described as research.
| Buyer | Immediate problem | Product wedge |
|---|---|---|
| UK startups and SMEs | Routine contracts are slow, fragmented and expensive | Guided drafting, contract review and CLM-lite |
| Small legal practices | Knowledge and matter workflows are spread across tools | Legal CRM, drafting and authorised-data assistance |
| Employers and HR advisers | Employment documents and obligations need consistent handling | Employment document automation and review playbooks |
| Professional-service firms | Repetitive client intake, review and reporting | White-labelled workflows and API access (Research Programme — architecture briefing only; customer use excluded) |
| Larger enterprises | Governance, permissions and auditable contract operations | Private knowledge: Research Programme — architecture briefing only; customer use excluded. |
Questions, answered
What is the legal LLM Research Programme?
Does RAG guarantee accurate legal answers?
What makes the legal layer proprietary?
How can I access each feature?
Which models can I use?
Can a firm connect its private knowledge?
Is Dogetlawyer approved by the SRA or BSB?
How is this useful to a customer?
Explore your next workflow
Dogetlawyer is a technology platform, not a law firm. "AI Lawyer" is a product name for AI software. Independent professionals provide services under separate engagement terms. This is general legal information, not legal advice.