LawDroid

AI Centrality

How much of the product is actually AI. Whether the machine learning is the mechanism the buyer is paying for or a feature layered onto conventional software, and whether the vendor is specific about which is which.

Two products with genuinely different centrality, and the record grades the composite honestly rather than the stronger half. Builder, the intake and automation product this category is about, is fundamentally a no code workflow and conditional logic engine: independent review states plainly that Builder draws on the workflows, templates, questions, logic and content the firm builds into it, and that it is not positioned as a broad legal research database. Chatbots built on decision trees and document assembly on conditional logic are deterministic automation, and natural language question answering sits on top as an addition. Copilot is the opposite and is genuinely model native, being an AI assistant with no function absent the model. LawDroid predates generative AI by years, having operated since 2016 when chatbot automation was rule based, and Copilot is described in independent review as a new product incorporating large language model generative AI into an existing repertoire. Graded C on the composite: the platform this category buys is an automation builder that now has AI in it, not a model that automation was built around.

Citation Accuracy and Hallucination Disclosure

Whether the vendor publishes measured accuracy on citations and assertions, grounds output to primary sources, and says plainly what its system does when it does not know. Legal has a documented public record of fabricated citations reaching filed briefs, so an untested claim of accuracy is not evidence.

A research capability with no disclosed sources and an explicit third party warning. Copilot is stated to assist with case law research, and independent review states that the specific sources from which it retrieves case law are not explicitly detailed, that outputs should be treated as work product requiring attorney review, and that attorneys should verify results before relying on them. Unsourced case law retrieval is the highest risk configuration in legal AI and it is exactly the shape that produces fabricated citations. Nothing published by the vendor addresses grounding, citation to source, accuracy, hallucination rate, evaluation or benchmark. The Builder side compounds it differently: chatbots answer frequently asked questions using natural language over firm supplied content, and nothing states what happens when a question falls outside that content. Checked the Copilot product page, the home page, the Builder material and independent review on 29 Aug 2026.

Autonomy and Oversight Model

What the system decides on its own, what a lawyer must approve, and whether the vendor documents where the review point sits. A tool that drafts under review and a tool that files without one are different products and different risks.

One published feature that directly answers this axis and is uncommon on the index. The vendor states human agent takeover, describing it as allowing a person to jump in and take over an automated conversation, with the framing that automation is good and automation plus the human touch is better. That is an explicit, named mechanism for a human to interrupt an autonomous process mid execution rather than review it afterward, which is a different and stronger thing than post hoc review, and it is the right control for a chatbot conversing with a member of the public in real time. Independent review reinforces the posture, noting output should be treated as work product needing attorney review. Held at B because nothing is bounded: no statement of what triggers a handoff, whether escalation can be automatic on a detected condition, what the bot does while waiting for a human, or what it may say unattended in the Enterprise tier where LawDroid builds and manages the chatbot on the firm's behalf.

Operational and Outcome Evidence

Named, dated evidence that the product works in production at real firms or legal departments. Case studies with figures and identified customers count. Unattributed testimonials and launch announcements do not.

Independent review coverage is the substance, and customer evidence is thin. The strongest element is sustained assessment by a named legal technology publication that reviewed both products separately, verified pricing and integrations, and recorded specific limitations including that no secondary source library was verified for Builder and that Copilot's case law sources are not detailed. Independent review that documents what could not be verified is better evidence than a testimonial. The vendor operates since 2016, which is unusual longevity in this market, and holds a named affinity partnership with that publication. Against that: one customer quote is attributed to a named individual as CEO and Founder of LexBlog, and no law firm, legal aid organisation, court or government agency customer is named anywhere despite all four being stated buyer types, no usage figure, no case study and no outcome claim with a figure were located, and a major review platform profile shows zero reviews. Held at C on that split.

Privilege and Confidentiality Posture

How client confidences are handled: attorney client privilege and work product treatment, segregation of one client matter from another, whether client data trains any model, and what the vendor commits to in writing rather than in marketing.

Nothing substantive located. The only statements found are that the platform maintains confidentiality and safeguards sensitive information, which is unfalsifiable phrasing naming no mechanism, and an independent review checkbox indicating the product claims to keep data secure. Nothing addresses attorney client privilege or work product. The gap has a specific shape in this category that no prior record has raised: an intake chatbot converses with a prospective client before any engagement exists, and the information collected sits in the uncertain zone before privilege attaches, which is both a professional responsibility question and a conflicts question, since a firm that learns a prospective client's confidences may be conflicted out of representing the other side. Nothing published engages any of it. Checked the Copilot page, the home page, the Builder material and independent review on 29 Aug 2026.

UPL and Professional Responsibility Posture

Whether the vendor is clear that it supplies a tool rather than legal advice, who its audience is, and how it addresses unauthorized practice of law, competence and supervision duties, and jurisdiction limits. ABA Formal Opinion 512 is the reference point.

Not located, and this is the sharpest instance of the question on the index so far. The vendor markets Builder to firms as a way to scale legal expertise and charge for services such as self serve legal documents, issue spotting and legal guidance, delivered while the lawyer sleeps. Issue spotting and legal guidance delivered autonomously to a member of the public, for a fee, without a lawyer present is the unauthorised practice scenario in its most direct form, and the vendor names it as the product's purpose rather than as an edge case. Buyer types include legal aid organisations and courts, where the end user is a self represented member of the public. Nothing published states that output is not legal advice, addresses where the line sits between legal information and legal advice, or engages any bar authority or access to justice guidance on unbundled or automated delivery. Checked the Copilot page, the Builder material, the home page and independent review on 29 Aug 2026.

AI Governance and Bias Disclosure

Published governance over model behaviour: who owns it inside the vendor, what is tested before release, and what is disclosed about disparate output across matter types, parties, or populations.

Nothing published about how the models are governed, evaluated or monitored. No AI policy, no model card, no bias or fairness testing, no evaluation methodology, no accuracy monitoring, no drift statement, no named governance body, no ISO 42001 and no EU AI Act positioning were located. The untested risk is specific to intake rather than generic: an intake chatbot performs triage, deciding which enquiries convert to leads and which are turned away, and any systematic tendency in that triage falls on prospective clients who are never told a machine assessed them and have no route to appeal. In the legal aid and court configurations the vendor names, the people being triaged are self represented and often have no alternative. Checked the home page, the Copilot page, the Builder material and independent review on 29 Aug 2026.

AI Safety and Data Stewardship

Retention, deletion, access control, and what happens to prompts and documents after they are processed. Whether the vendor states its subprocessors and its incident practice, or leaves the buyer to assume.

No stewardship position located. Nothing states whether intake conversations, uploaded documents, firm built workflow content or Copilot prompts are used to train or improve models, no retention period is published, and no deletion right is described. The content at stake includes conversations with prospective clients about their legal problems, which is sensitive personal information collected from members of the public who are not the customer and have no relationship with the vendor at all. The only located statements are generic assurances of confidentiality and privacy. Checked the Copilot page, the home page, the Builder material and independent review on 29 Aug 2026.

AI Liability and Recourse

What the vendor stands behind contractually when its output is wrong. Indemnities, caps, carve outs, insurance, and whether any of it is published or only reachable through a negotiated agreement.

No published position located. Nothing was found on liability for AI output, warranty, service levels or remedy. Three exposures are distinctive here. An unsourced case law citation from Copilot reaching a filing is the fabricated citation risk in its classic form. A chatbot delivering issue spotting or legal guidance that is wrong reaches a member of the public directly with no lawyer in between, which is the vendor's stated use case rather than a misuse. And in the Enterprise tier LawDroid builds and manages the chatbot on the firm's behalf, so the vendor authored the conversation flow that spoke to the public, and nothing addresses who carries that. Checked the Copilot page, the Builder material, the home page and the site navigation on 29 Aug 2026.

Practice Systems Integration Depth

How deeply the product reaches into the systems legal work already lives in: document management such as iManage and NetDocuments, Word and Outlook, contract lifecycle management, matter management, e-billing, and court filing systems.

Named practice management integration with the distinction between products stated precisely, which is unusually candid. Builder integrates natively with Clio Grow and Clio Manage, which are the intake and practice management systems the target buyer actually runs on, and native integration into the intake product specifically is the correct connection for this category since a captured lead must become a matter. Copilot connects only through Zapier, and independent review states this plainly, noting that any practice management system connecting to Zapier could potentially work with Copilot. Publishing that one product integrates natively and the other does not, rather than claiming integration platform wide, is the kind of precision this index credits. Lead capture into a firm's case management system is stated as a Builder function, and API workspace integrations are referenced. Held at B rather than A because only one practice management vendor is named, no API documentation was located, and the Zapier dependency is a real limitation for Copilot rather than an integration.

Deployment Model and Data Residency

Where the software runs and where the data sits. Multi tenant cloud, single tenant, private deployment, on premises, and whether region of residence is a published option or an enterprise conversation.

Nothing located. No hosting provider is named, no region or data residency commitment is published, and no deployment options are described beyond the products being web based and cloud hosted. Residency matters here because the vendor states buyers including courts and government agencies, which routinely carry procurement requirements about where public data is processed, and because intake conversations with members of the public are personal data subject to state and international regimes. Checked the Copilot page, the home page, the Builder material and independent review on 29 Aug 2026.

Security Certifications and Trust Center

Independent attestation a buyer can pull without a sales call: SOC 2, ISO 27001, penetration test summaries, a trust center with current reports and named scope rather than a badge image.

No certification, attestation, trust centre or security page was located. No SOC 2 of either type, no ISO 27001, no named auditor, no penetration testing partner and no encryption statement were found. The only located material is a generic claim that the product keeps data secure, recorded through an independent review checkbox whose own framing notes that most AI products are built in such a way that they are inherently insecure, which is the reviewer's caution rather than a finding about this vendor. Under the three tier test the artifact is absent rather than gated. The gap is material given the stated buyer set includes courts and government agencies, which ordinarily require a named attestation before procurement. Checked the home page, the Copilot page, the Builder material, the site navigation and independent review on 29 Aug 2026.

Model Supply Chain Disclosure

Which models sit underneath, whose they are, where they run, and whether the vendor commits to telling customers when that changes. A legal buyer inherits every dependency it cannot see.

Nothing located. No foundation model provider, model family or version is named, no distinction is drawn between proprietary and third party models, and no subprocessor list was found. Independent review describes Copilot as incorporating large language model generative AI, which confirms an external or licensed model layer exists without identifying it, and the absence of any named source is compounded by the separate finding that Copilot's case law retrieval sources are also undisclosed, so neither the model nor the corpus behind the research capability can be identified from public material. Checked the Copilot page, the home page, the Builder material and independent review on 29 Aug 2026.

Commercial Transparency

Whether a buyer can learn what this costs without entering a sales process: published rates, the unit being charged, what sits behind an enterprise tier, and what implementation adds.

Complete published pricing across a multi product line, including the awkward combination case most vendors hide. Published and independently verified: Copilot at $25 per user per month, Builder at $99 per user per month, both month to month with no annual contract required, a 7 day free trial on Copilot, and the two products together at $99 per user per month which requires an annual commitment. That last item is the notable one, because publishing that the bundle costs the same as Builder alone but locks the buyer into a year is a real commercial trade off disclosed rather than buried, and a buyer can weigh it without a sales conversation. An Enterprise tier exists for managed delivery and is unpriced, which is the only gap. Third A on this axis in 55 records, after Descrybe and Huski.ai, and the pattern across all three is the same: small vendors selling to solo and small firms publish their prices, and enterprise vendors selling to large firms do not.

Firm and Practice Coverage

Who the product is actually built for. AmLaw, midlaw, small firm and solo, in house departments, government and courts, and which practice areas are supported rather than merely claimed.

Buyer coverage is stated broadly and practice coverage is not stated at all. Named buyer types span law firms, legal aid organisations, courts and government agencies, which is a genuinely wide institutional range and the only record in this pull to name courts and legal aid as direct customers, and firm size targeting is clear from pricing and independent review as solo practitioners and small firms. What is absent is any characterisation of substantive coverage: no practice areas are named, no jurisdictions are stated, and because Builder operates on firm supplied content, its subject matter coverage is whatever the customer builds rather than anything the vendor provides, which the vendor never says. On the Copilot side independent review records that no secondary source library was verified and that case law sources are undetailed, so research coverage cannot be assessed either. Held at C: the buyer set is characterised, the substance is not.