No-Code AI for Legal Work in 2026: The Ethics Gate Comes Before the Price Tag

Why the Ethics Gate Comes Before the Price Tag

The gap between adoption and governance is now measurable. A compilation of industry surveys published by the NC Bar Association reports that, per the Clio 2026 survey of U.S. legal professionals, AI adoption stands at “71% of solo practitioners and 75% of small firms” — while in the same survey “57% of solos and 55% of small firms” have no AI policy (NC Lawyer). A separate 8am “2026 Legal Industry Report” of more than 1,300 legal professionals found that “43% said their firm has no formal AI policy and no plans to create one,” and “Only 9% said their firm has a written policy that is actually enforced” (NC Lawyer).

Clear the ethics gate first. Then do the price math.

Gate One: ABA Formal Opinion 512

On July 29, 2024, the ABA Standing Committee on Ethics and Professional Responsibility released Formal Opinion 512, the ABA’s first formal opinion on generative AI — a “15-page opinion” (ABA news release). Its core move is not to invent new duties. It is to say that the existing model rules on competency, informed consent, confidentiality and fees principally apply (ABA news release).

Competence and confidentiality

The opinion enumerates four provisions: “Model Rule 1.1 (Competence),” “Model Rule 1.6 (Confidentiality of Information),” “Model Rule 1.4 (Communications)” — including “Model Rule 1.4(a)(2)” — and “Model Rule 1.5 (Fees)” (ABA news release).

For a buyer, Rule 1.1 is the pre-purchase duty: you cannot competently supervise a tool you do not understand, so understand its failure modes before a client’s matter depends on it. Rule 1.6 is the disclosure duty: confidential information should not reach a third-party system without the informed consent Rule 1.6 ordinarily requires.

Communication and fees

Rule 1.4(a)(2) requires the lawyer to “reasonably consult” with the client about the means of pursuing the representation — which is why “the client never mentioned it” is not a durable position (ABA news release).

Rule 1.5 gets the opinion’s most quoted example. If a lawyer uses a generative AI tool to draft a pleading and expends “15 minutes to input the relevant information into the program,” the lawyer may charge that time plus review time — but “in most circumstances, the lawyer cannot charge a client for learning how to work a GAI tool” (ABA news release).

Important framing: ABA 512 interprets the Model Rules. It is not binding law in any state, and your jurisdiction’s rules control.

Gate Two: Florida Bar Opinion 24-1, the Vendor Checklist

Florida Bar Ethics Opinion 24-1 is dated January 19, 2024 — six months before the ABA opinion — and it opens by stating plainly: “Advisory ethics opinions are not binding” (Florida Bar Opinion 24-1). It is guidance, not law.

The pre-purchase vendor questionnaire

Under Rule 4-1.6 and 4-1.6(e), Florida says lawyers must research “the program’s policies on data retention, data sharing, and self-learning” before disclosing client information (Florida Bar Opinion 24-1). A “self-learning” program, the opinion warns, raises the possibility “that a client’s information may be stored within the program and revealed in response to future inquiries by third parties” (Florida Bar Opinion 24-1).

The opinion then adopts the Florida Op. 12-3 cloud-computing checklist, which becomes a ready-made questionnaire you can send before a demo call:

Question to ask the vendor	What Florida 24-1 says to look for
Is your confidentiality obligation to us contractually enforceable?	Ensure the provider’s confidentiality obligation is enforceable
Will you notify us of a breach or of legal process served on you?	Provider “will notify the lawyer in the event of a breach or service of process”
What are your security measures, reputation, and liability limits?	“investigate the provider’s reputation, security measures, and policies, including any limitations on the provider’s liability”
Do you retain our data, and do you claim rights to it?	“determine whether the provider retains information…before and after the discontinuation of services or asserts proprietary rights to the information”
Is our data used for training or shared with sub-processors?	Research “data retention, data sharing, and self-learning”

All rows paraphrase or quote the checklist in Florida Bar Opinion 24-1.

Florida recommends that a lawyer “obtain the affected client’s informed consent prior to utilizing a third-party generative AI program if the utilization would involve the disclosure of any confidential information” — and states the converse: “If the use of a generative AI program does not involve the disclosure of confidential information to a third-party, a lawyer is not required to obtain a client’s informed consent pursuant to Rule 4-1.6” (Florida Bar Opinion 24-1). The opinion also notes that the confidentiality risk “may be mitigated by use of an inhouse generative AI rather than an outside generative AI,” and points to Rule 4-5.3(a) supervision standards, which extend not only to the firm’s own employees but “over entities outside the lawyer’s firm with whom the lawyer contracts” (Florida Bar Opinion 24-1).

What Happens When You Skip the Gate: Mata v. Avianca

Mata v. Avianca, Inc. was decided June 22, 2023 by the U.S. District Court for the Southern District of New York, Judge P. Kevin Castel presiding (Mata v. Avianca, Inc.). Plaintiffs’ lawyers were sanctioned with a “$5,000” fine after filing fake precedents generated by ChatGPT (Mata v. Avianca, Inc.).

The detail that matters for buyers is what the court did not say. The opinion acknowledges that “there is nothing inherently improper about using a reliable artificial intelligence tool for assistance” (Mata v. Avianca, Inc.). The sanction was for non-verification, not for using AI. The failure was procedural: nobody checked whether the citations were real.

Billing and Intake Rules to Settle Before You Buy

Your pricing model is part of ethics compliance, and most firms have not touched it: “86% of solo firms and 78% of small firms have not adjusted their pricing models to account for AI use” (NC Lawyer).

Before you sign, settle four rules:

Actual-cost billing requires notice. Charging the actual cost of generative AI requires informing the client “preferably in writing,” charges must be reasonable and not duplicative, and there is to be no improper billing “such as double-billing” (Florida Bar Opinion 24-1).
Unallocable costs are overhead. If a lawyer is “unable to determine the actual cost associated with a particular client’s matter, the lawyer may not ethically prorate the periodic charges of the generative AI and instead should account for those charges as overhead” — and may not charge for “time spent developing minimal competence in the use of generative AI” (Florida Bar Opinion 24-1). This aligns with ABA 512’s rule that you “cannot charge a client for learning how to work a GAI tool” (ABA news release).
Chatbots must disclose themselves. A generative AI chatbot talking to clients or third parties “must include a disclaimer indicating that the chatbot is an AI program and not a lawyer or employee of the law firm,” and the lawyer “will be ultimately responsible in the event the chatbot provides misleading information to prospective clients” (Florida Bar Opinion 24-1).
Advertising AI use has limits. Under Subchapter 4-7, Rule 4-7.13, lawyers may advertise AI use but “cannot claim their generative AI is superior to those used by other lawyers or law firms unless the lawyer’s claims are objectively verifiable” — Rule 4-7.13(b)(3) governs objectively verifiable comparisons and Rule 4-7.13(b)(5) governs voice/image disclaimers (Florida Bar Opinion 24-1).

What the Tools Actually Cost (Published Numbers Only)

Very little real pricing is public, and the table below quotes only figures that appear on vendor pages.

Comparison table

Vendor / product	Role	Published price, verbatim	AI included at that price?	Data-training statement?	Exit terms stated?
Clio	Practice management	Plans “start at $49 USD per user per month” (Starter)	No — “AI included on plans from Core up”; Core/Signature/Elite prices not published	Yes: “never used for AI training or any other external purpose”; “Zero-retention agreements with LLM providers”	Yes: open-format export, 90-day post-cancellation window
Gavel Document Automation	Document assembly	Lite “$83/mo”, Standard “$210/mo” (monthly-basis figures)	Not stated for these tiers	Not stated	Seat and template caps stated; export terms not stated
Spellbook	Contract review (Word add-in)	No published price — “custom pricing,” “Book a Demo”	Not applicable at a published price	Yes: “Zero Data Retention agreements, preventing data use for training”	Not stated in these sources
Lawmatics	Intake / CRM	No published price — “custom quote”; tiers Essential, Premium, Enterprise	AI features listed on Essential, Premium and Enterprise	Not stated	Contact caps stated (500 / 10,000); export terms not stated
Docassemble	Guided interviews / document assembly	“free, open-source”	Features include AI (LLMs/machine learning)	Not applicable (open-source, self-hosted)	Self-hosted; export is your own data store

The traps

The Clio AI-tier trap. Starter is “$49 USD per user per month,” but “AI included on plans from Core up,” and no Core, Signature or Elite price appears on the page — those tiers say “Get pricing” (Clio pricing). The true cost of AI-capable Clio is therefore more than $49 and not established by these sources.

The Gavel product split. The “$83/mo” and “$210/mo” tiers are Document Automation — 1 Builder Seat with 10 or 50 templates respectively (Gavel pricing). Gavel Exec, the separate AI contract review product with Draft, Review, Redline, Playbooks and Benchmarking, has no published price. Never attach $83 or $210 to contract review. The page has a “Billed monthly” / “Billed annually” toggle, but only monthly-basis figures are quoted here; no annual discount should be assumed.

Badges are not clearance. Spellbook’s security page displays “SOC 2 Type II Compliant,” “HIPAA Compliant,” “GDPR Compliant” and “EU AI Act Compliant” (Spellbook security); Clio states “SOC 2 Type II” certification (Clio pricing). These are security and privacy attestations. None of them constitutes compliance with ABA 512 or Florida 24-1, a Rule 1.6 determination, or any legal-ethics clearance.

The Exit Math: Lock-In Is an Ethics Question

Florida 24-1 asks you to determine whether a provider retains information “before and after the discontinuation of services” (Florida Bar Opinion 24-1). That makes export terms part of the purchase decision, not a footnote.

Clio is the most explicit: you can retrieve contacts, matters and financial records “in open, nonproprietary formats at any time while your account is active, or within 90 days of cancellation,” and cancellation “takes effect at the end of your current term” (Clio pricing).
Gavel caps you at “1 Builder Seat” and 10 or 50 templates (Gavel pricing) — templates are firm intellectual property, so confirm ownership and export before committing.
Lawmatics tiers are defined by contact ceilings: Essential includes “500 contacts,” Premium and Enterprise “10,000 contacts” (Lawmatics pricing). What happens to those contacts on cancellation is not stated in these sources.
Spellbook states no export or post-cancellation retention terms in these sources (Spellbook pricing).
Docassemble has low license lock-in — free and open-source, MIT-licensed code distributed via GitHub and PyPI — but high skills lock-in: it is “a free, open-source expert system for guided interviews and document assembly, based on Python, YAML, and Markdown,” and you run the server, including its security (Docassemble). It is not no-code for a non-technical operator.

Only Clio states a data-training policy in these sources. Silence from the others does not mean anything in particular — it means your Florida 24-1 research duty is still undischarged, and you must ask.

Adoption Reality Check (Read the Numbers Carefully)

The governance gap, not the technology, is the story. Beyond the figures above: time is the biggest hurdle for “27% of solos and 33% of small firms,” while “86% of mid-sized firms reported they have adopted AI” (NC Lawyer). Clio’s own 2026 Legal Trends Report is the origin of the Clio figures cited here (Clio 2026 Legal Trends Report).

Client pressure is minimal: “only 6% of respondents said clients are explicitly pushing for AI-linked cost reductions,” and “Fifty-four percent said their firm provided no training on responsible AI use and has no plans to do so” (NC Lawyer).

Three cautions on reading these numbers:

Do not merge them. The Clio and 8am figures come from different populations and different questions and are not directly comparable. Clio’s “57% / 55% no AI policy” and 8am’s “43% no policy, 9% enforced policy” are not two views of one statistic.
The base is law-firm-heavy. The NC Lawyer compilation reports no in-house-specific adoption or pricing data.
Tool categories matter. The NC Lawyer article notes many common tools are “generic consumer products rather than legal-specific systems, which can increase confidentiality risk and create workflow fragmentation when teams copy information between separate platforms.”

FAQ

Is ABA Formal Opinion 512 binding on me?

No. It interprets the ABA’s Model Rules; state rules control. Florida’s opinion is explicit that “Advisory ethics opinions are not binding” (Florida Bar Opinion 24-1), and the same caution applies to ABA guidance.

Can I bill clients for AI time?

Partly. Under ABA 512 you may charge the “15 minutes to input the relevant information into the program” plus review time, but “in most circumstances, the lawyer cannot charge a client for learning how to work a GAI tool” (ABA news release). Under Florida 24-1, charging actual AI cost requires informing the client “preferably in writing,” and unallocable periodic charges must be accounted for “as overhead” rather than prorated (Florida Bar Opinion 24-1).

Not under Florida’s formulation. Consent is recommended where the use “would involve the disclosure of any confidential information”; if no confidential information reaches a third party, “a lawyer is not required to obtain a client’s informed consent pursuant to Rule 4-1.6” (Florida Bar Opinion 24-1). Rule 1.4 consultation duties may still apply.

No. A SOC 2, HIPAA, GDPR or EU AI Act badge is a security-privacy attestation, not a determination under ABA 512 or Florida 24-1. Spellbook lists those badges on its security page; they answer a diligence question, not the ethics question.

Is Docassemble no-code?

No. Authoring is “based on Python, YAML, and Markdown,” and you run the server yourself (Docassemble). It is free and open-source, which is a different attribute from being no-code.

What does Spellbook cost?

It publishes no price — “custom pricing” with a demo request (Spellbook pricing). Lawmatics likewise offers no price, only a “custom quote” across Essential, Premium and Enterprise tiers (Lawmatics pricing).

The Bottom Line

Clear the ethics gate first. Send the Florida 24-1 questionnaire above before any demo. Get the answers in writing.
Put one page in place before the first invoice: which tools may touch confidential information, what consent language you use, how AI time is billed, and how chatbot-facing tools are labeled.
Read published prices as floors for the wrong things. “$49” is Clio Starter, without AI. “$83” and “$210” are Gavel Document Automation, not contract review. Neither number is the cost of what most buyers are shopping for.
Treat silence as a work item. Where a vendor publishes no price, no data-training statement, and no export terms, the absence itself is the answer to “what does it cost” — you cannot complete diligence on a blank page, and Florida’s research duty still sits with you.
Verify your own jurisdiction. ABA 512 interprets Model Rules, both opinions predate 2026, and updated guidance may exist. This guide states what these documents say; it is not advice about what your state permits.

How This Guide Was Built

This post was assembled from 12 primary and secondary sources, all fetched 2026-11-01: the ABA news release on Formal Opinion 512, Florida Bar Opinion 24-1, the NC Lawyer survey compilation, the Clio 2026 Legal Trends Report, Clio’s pricing and legal AI pages, Spellbook’s pricing and security pages, Gavel pricing, Lawmatics pricing, Docassemble, and the Mata v. Avianca entry.

What was not done: no product was tested. Nothing here measures accuracy, hallucination rates, review quality, or whether any tool “works.” No contract was run through any system, and no vendor was contacted.

What the sources cannot establish: prices for Spellbook, Lawmatics, Gavel Exec, or Clio Core/Signature/Elite; export and retention terms after cancellation for Spellbook, Gavel and Lawmatics; whether any tool is genuinely no-code for non-technical operators; any in-house-specific adoption or pricing data; and any claim that a vendor satisfies the consent, billing, or chatbot duties of ABA 512 or Florida 24-1. Both ethics opinions predate 2026, and neither establishes what any particular state has adopted since.

ToolBrain — tool reviews, LLM comparisons, and AI workflow guides
CodeIntel Log — code quality, debugging, and software engineering benchmarks

Cross-links automatically generated from NoCode Insider.