Legal services
AI agents built for the evidence your ethics counsel will ask for.
ABA Formal Opinion 512 requires lawyers to validate AI tools for ethics compliance. GoodHelp gives small and mid-size law firms agents whose every action is hash-chained, approval-gated, and exportable as signed evidence — the documentation your ethics counsel needs to evaluate the tool.
Bar counsel needs evidence. Most tools can’t produce it.
- ABA Op. 512 and state bar opinions require you to validate any AI tool. Most can't produce the evidence.
- Privilege and confidentiality apply to every matter — generic AI tools can't show per-matter scoping.
- Trust account (IOLTA) rules apply if AI touches client funds workflows.
- Malpractice exposure compounds if you can't show what an AI tool did.
What your bar counsel will ask to see.
Bar-counsel-ready evidence
Per-matter action ledger filtering. Signed evidence export — a JSON manifest plus body (CSV/TSV/JSON), with a KMS-rooted signing key.
Approval gates on anything material
Junior-associate workflow with partner-approval gates configurable per matter. Enforced fail-CLOSED (SOD_REQUIRE_AUDIT_AVAILABLE=true): if the SoD audit context is unavailable the approval is blocked rather than allowed.
Per-matter confidentiality controls
Application-layer per-matter scoping. Cross-matter access requires an explicit matter context at the routing layer. Privilege determinations remain with the supervising attorney; we provide the access controls and audit trail that support your analysis.
ABA Op. 512 alignment
Documentation supporting your duty of competence with AI tools. The duty is on the attorney; we provide the evidence trail and approval-gate controls that help you meet it.
Model Rule 5.3 supervision evidence
Override any agent action at the SoD approval gate before it commits — every override, every approval, every cancellation lands on the hash-chained per-org action ledger. The supervisory record your malpractice carrier asks for. (Per-firm agent enable/disable today; per-matter scoping is at the routing layer; per-matter agent disable on the roadmap.)
The control surface, in audit-ready language.
Action-ledger filtering for bar counsel review
Evidence pack — KMS-rooted signing key
Supporting your duty of competence
Where small and mid-size firms put us to work.
Document review triage
Classifies and flags for partner review with full action trail. Every classification decision is hash-chained and exportable.
Time-entry assembly
Drafts billable-hour entries from calendar and email signals, requires associate approval before posting to the matter.
Conflict-check data assembly
Pulls and normalizes party data from your conflict database. Surfaces records matching your configurable criteria. The conflict determination remains with the attorney; the agent does not clear conflicts.
Administrative intake support
Appointment scheduling, document collection, intake-form routing — with full disclosure of AI involvement to the prospective client. The agent does not evaluate the merits of the matter or render any legal opinion.
Matter-specific knowledge management
Per-matter document retrieval with application-layer scoping. The router enforces per-matter org_id before any retrieval call lands.
What law-firm partners ask first.
Is GoodHelp ABA Op. 512 compliant?
ABA Op. 512 places the duty on you, the attorney. GoodHelp produces the evidence and approval-gate controls that help you meet that duty. There is no ABA certification for vendors — we do not represent ourselves as ABA-certified or Op. 512-certified.
What about state-specific opinions?
Major state bar opinions (CA, NY, FL, TX) generally extend ABA Op. 512's themes — competence, confidentiality, supervision, communication. We track them; your bar counsel should make the per-jurisdiction call. We can route you to relevant claim files for diligence.
Can you sign a confidentiality / NDA?
Standard mutual NDA available on request via sales.
Does the AI access client matter data?
Only within the scope you configure per matter. Cross-matter access requires an explicit org_id at the routing layer — the platform enforces per-org Firestore sub-collection isolation at the application layer. We do not enforce isolation via Firestore Rules.
Pricing?
Per-agent monthly fee plus LLM passthrough margin. Per-run cost cap. See the pricing page for current numbers.
Under the hood
The solutions and platform behind it
Every industry runs on the same agents and the same governed platform. Jump to the solution that matches your work — or see how the platform keeps every action approved, attributed, and audit-ready.
Agent platform
Hire, configure, and route a team of production agents across departments.
Back-office automation
Finance, vendor, and bookkeeping work — proposed by agents, approved by you.
Compliance evidence
Per-action ledger, approval history, and signed evidence exports for auditors.
Content operations
Research → draft → score → distribute, priced by the output and refundable.
Show your bar counsel the evidence trail.
Thirty minutes — bring the questions your firm’s ethics committee would ask. We’ll walk through the per-matter ledger, the approval gates, and the signed evidence pack.