L

Law Practice AI

Demand AI reads uploaded case files and drafts the demand letter, with sections for facts and liability, diagnosis and treatment, past and future medical expenses and an exhibit packet with Bates numbering. The whole price list is built on it. Case Summary AI condenses police reports, medical records and depositions and adds an assistant that answers questions about them. The intake agent answers and places calls, texts and emails, screens leads and routes urgent ones.

Document Collector is different in kind: its requests to clients and providers, reminders, portal and tracking of received and missing files are workflow. According to the FAQ, it does not generate legal documents but only collects and verifies them. Without the models, the collection tool remains while the demand drafting, summaries and intake agent disappear.

Law Practice AI's security page says every AI generated document is checked before attorney review against the firm's uploaded case documents, ICD code descriptions from the National Library of Medicine, NHTSA complaint and recall data and US Department of Labor wage data. That ties diagnosis language, vehicle defect claims and lost wage figures to named public sources, and exhibits are Bates numbered and linked to their documents.

No measurement is published. Its home page credits the product with a 90 percent reduction in documentation errors, and its FAQ says it reduces missed details. Neither gives a test set, period or definition of an error, and a case study speaks of significantly fewer missed issues without a number. Section 3 of the Terms puts responsibility for the accuracy of AI output on the user.

The security page calls the approach human in the loop and verification first, says every draft requires attorney review before use, and says no document leaves the platform without explicit sign off. Section 3 of the Terms says the AI system is designed to generate outputs automatically and without human intervention, and makes users solely responsible for reviewing and verifying them. The intake agent works on its own in practice: it answers calls, screens leads, detects invalid contact details and routes high priority leads.

Firms set follow up sequences, contact windows and a setting the FAQ calls agent outreach aggressiveness. Review surfaces exist, including call recordings, conversation summaries, agent reports and live dashboards. What the intake agent may decide or say before a person sees the lead, and when it must hand off, is not described, and the gate on documents is not in the agreement.

The home page quotes attorneys by name and firm, including David Rowland of Lemon My Vehicle, Chris Cunningham of McWhirter, Bellinger & Associates and Jordan Ariel of Ariel Law Group, and says more than 300 firms use the platform. The case studies give the figures. A solo personal injury attorney increased signed cases by 38 percent and cut missed calls by 89 percent with the intake agent. Another cut case summary time to 15 minutes a file and reclaimed more than 40 hours a month.

Others report a demand drafted in 30 minutes with $3,000 in monthly savings, deposition preparation cut to 20 minutes and review at 1,400 pages an hour. A 12 attorney firm reports 57 signed cases a month and $1.2 million in annual uplift. None of the case studies names the firm or carries a date, and none says how the before and after figures were measured.

The security page says the firm's client data is used only to generate output for its cases and is never reused or stored inside AI models, that AI sessions are reset after every request with no health information retained, and that a business associate agreement is signed with every firm before any protected health information enters the platform. The Privacy Policy, which governs use of the AI products, says inputs, outputs and personal information are shared with and processed by AI service providers including Microsoft Azure AI.

It says unredacted documents and health information may be used for system development, and that adjuster information from insurance claims may train its machine learning models. The only mention of privilege is an FAQ answer saying uploads are processed under strict confidentiality standards to protect attorney client privilege. Nothing addresses separation between firms or between matters. Nothing says which document binds, and the Privacy Policy, the one that governs use of the AI products, is the less protective.

Section 3 of the Terms says users are solely responsible for the accuracy and outcomes of AI generated documents, must review and verify them before use and must comply with applicable law, and that the company is not liable for legal outcomes. The intake terms make the firm answerable for the legality of its intake messaging and have it indemnify the company for violations. The FAQ answers the question "Will callers know they're speaking with AI?" by saying most will not notice, and that the firm can choose whether the agent introduces itself or blends in with the team.

The agent speaks to injured members of the public, and no disclosure commitment is made. No statement says the product does not give legal advice, nothing connects the demand drafting to a lawyer's duties of competence and supervision, and no jurisdiction limits are named beyond California law governing the agreement.

The company describes a verification step that checks each draft against named datasets, which is a control on individual outputs rather than a governance program. No one is named as accountable for how the models behave, nothing describes what is tested before a change to the drafting or intake models reaches firms, and nothing addresses whether drafts value injuries differently across claimants, injury types or languages, or whether the intake agent screens out some callers more often than others. The product proposes settlement figures and decides which injured callers get routed as high priority.

Encryption is TLS 1.2 or later in transit, 4096 bit authentication tokens and encryption at rest. Access is role based, with separate levels for admins, staff and specialists, and incident response is documented, with prompt notice aligned to the HIPAA Breach Notification Rule. Data is kept only as long as workflows need it, then deleted or anonymized, with a 30 day export window after termination followed by permanent deletion.

For the AI itself the security page promises that model memory is wiped after each use and no health information is retained. The Privacy Policy says otherwise on the same subject: uploaded documents and health information may be used for system development, adjuster information may train models, and Microsoft Azure AI receives inputs and outputs. No subprocessor list is published and no retention period is stated for anything during the term.

Section 3 of the Terms says the company is not liable for any legal outcomes or consequences from AI generated documents or information, and that users are solely responsible for their accuracy and outcomes. Section 6.1 disclaims any warranty about the accuracy, reliability or completeness of content and any liability for errors, mistakes or omissions in it. The intake terms have the firm indemnify the company for any breach of telemarketing, messaging or data handling law in its intake outreach.

No indemnity, insurance, service credit or output warranty runs to the firm, and nothing addresses a demand that understates damages or misstates treatment, though the output goes to an insurer as a settlement demand. The allocation is published.

The integrations page says Law Practice AI imports matter, client and document details from CASEpeer, MyCase and SmartAdvocate to start a demand, and exports finished or verified documents back to the matter. The FAQ adds that the intake product syncs with HubSpot, DocuSign, Microsoft Teams and Zapier and schedules callbacks through Google Calendar and Outlook. Collected documents can sync to Google Drive, OneDrive or Dropbox, and Demands AI connects to existing legal software in one click.

The platform is also offered white labeled to other companies' clients. Filevine and Litify are not listed, and no setup documentation is published: no field mapping, no statement of the permissions the connection holds, and no account of what happens when a matter changes on the other side.

The platform is sold as software as a service with sign up, a dashboard and a white label option, and the security page says every feature runs under the same controls regardless of firm size. The Terms say the site is hosted in the United States and the Privacy Policy says personal information is transferred to and processed in the United States. The AI processing runs through outside providers including Microsoft Azure AI, and nothing says which region those services run in, whether documents and health information stay in the US while they are processed, or where the platform's own storage sits.

No tenancy model, private deployment or regional choice is described, including for the enterprise tier, whose page describes scale rather than deployment.

The security page says Law Practice AI is SOC 2 certified and independently verified by a third party audit, fully HIPAA compliant, and holds ISO 27001 and HITRUST. The FAQ and product pages say SOC 2 Type II, and the security page commits to a signed business associate agreement with every firm before health information enters the platform. No auditor, audit period, certificate number, scope statement or trust center is published, and no route is offered to request a report without a sales conversation. No certificate or registry entry is shown for ISO 27001 or HITRUST.

The Privacy Policy says the AI products are provided through third party AI service providers including Microsoft Azure AI, and that a user's inputs, outputs and personal information are shared with and processed by those providers. So the cloud AI service that reads clients' medical records is named in a document the firm is bound by. The list is introduced with the word including, so it does not claim to be complete, and no model family, version, region or change notification is published.

The security page's promise that sessions are reset after every request describes how the models are used, not which ones.

The pricing page offers one plan, Essentials, at $97 a month, including one AI generated demand letter and Document Collector, with each additional demand at $97, unlimited revisions, real time usage tracking and no long term contract. The home page offers a seven day free trial. Firms producing a high volume of demands get custom pricing, and nothing is published for the intake agent, Case Summary AI, litigation support or the white label arrangement, although the case studies credit the intake agent with some of the largest results. Implementation, onboarding and integration costs are not addressed.

Law Practice AI says it is built for plaintiff law firms and has pages for personal injury and lemon law, where the demand templates show it: personal injury demands with diagnosis, treatment and expense sections checked against medical code data, and lemon law demands checked against vehicle complaint and recall data. Document Collector also names workers' compensation and employment documents. There are pages for solo, small, midsized and enterprise practices, with case studies from solo attorneys up to a 12 attorney firm and a statement that more than 300 firms use it.

The edge is not stated: whether demand drafting supports medical malpractice, mass tort, premises or employment claims, which states' damages rules the templates follow, or what the enterprise tier offers beyond volume pricing.