Sonar Legal vs Spellbook: how they compare in 2026
Both products put AI contract drafting and review inside Word, but the adjacent workflows make the choice more specific. Sonar Legal combines Partner and playbooks with definition navigation, drafting checks and a formatter that applies a house style across the agreement after paragraph review. It also brings drafting and visible revisions into Outlook. Spellbook pairs its Word suite with Associate, which plans and executes transactional work across multiple files, and Market, which compares terms with similar agreements by industry, jurisdiction and deal type. Sonar's Tables extracts chosen information across contract sets, while its knowledge hubs can use internal material and connected Microsoft storage. Spellbook is the closer fit when a transaction requires coordinated revisions across related files or market comparisons; Sonar brings contract editing, house styles and correspondence together in everyday office work.
At a glance
All 15 axes, side by side
The same grid applied to every vendor in the index, graded from public sources. Hover a grade to see what the letter means on that axis.
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.
Sonar Legal combines AI drafting and review with everyday Word tools for definitions, error detection and formatting. Partner analyzes the open agreement, drafts clauses, prepares memoranda from attached research and responds to counterparty comments. Playbooks turn precedents or instructions into review rules, and Tables extracts selected fields across contract sets. The drafting workflow can compare a response enriched by a clause knowledge base with one generated without it. Definition navigation and parts of the document interface remain useful independently of generative output. Formatting also uses AI to suggest styles before a user applies them, so automation extends beyond writing new contract language.
The models are the product. Review, Draft, Ask, Compare and the Associate agent are all generative capabilities delivered through a Word add in that exists to carry them. The company pivoted its whole business to generative AI in 2022, rebranding from Rally to Spellbook. Remove the models and there is no product, only an empty add in.
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.
Partner answers from the open agreement and attaches citations to its analysis. Playbook results show a pass or fail for each rule, a supporting contract passage and proposed replacement text. Tables links each extracted result to the exact clause or section and lets a reviewer correct it. For clause drafting, Jina embeddings identify relevant knowledge base clauses that accompany the request to the language model. The web application also returns citations for legal research across more than 100 jurisdictions, but the underlying law collections and update schedules are unspecified. There is no reproducible accuracy test covering those research answers or the contract extraction workflow.
Ask can use selected iManage files and return citations to the team's own contracts. The legal data connection retrieves material from named public sources such as CanLII, CourtListener, Cornell Law, EUR Lex and regional legal information institutes, with links to the underlying authority. Those sources can inform questions, drafting and review inside the same workflow. Their currency follows the source operator's updates. A reproducible accuracy evaluation for the current product, an automatic subsequent treatment check and a stated abstention threshold are not specified. A citation makes an answer inspectable; it does not establish that a quotation or legal proposition has been verified by the reviewing lawyer.
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.
Users initiate AI requests and decide when to insert or replace text in Word. Suggestions can become tracked changes or ordinary edits, with a chosen author and optional comment. Playbook creation pauses for clarification when instructions are ambiguous, and reviewers can exclude rules before running a check. Results have severity labels and individual revisions. Formatting presents a paragraph by paragraph review table before Apply changes the file. Outlook separates reading a thread from composing a reply, and insertion requires Compose Mode. These are tangible review controls, but the wider agentic task feature lacks a complete account of escalation, failure handling and actions requiring approval.
Review proposes changes in Word for the lawyer to inspect and approve, either individually or in a group. Playbooks encode the team's instructions, while Associate takes on work across several files with user oversight. Proposed edits can carry the reviewer's identity when accepted, and citation links let the reviewer inspect supporting material. The terms place responsibility for reliance and verification with the customer. The existing suite therefore has recognizable human review surfaces, but no universal threshold is stated at which Associate must stop or escalate an unsupported conclusion. The separately announced Autonomous Contract Management offering remains in early access.
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.
TWL Law Group managing partner Bob Tseng describes using Sonar Legal for cross border transactions and credits its practical workflow and hands on coaching with improved efficiency. The testimonial identifies both the firm and the legal buyer, but gives no dated before and after measurement. Sonar also reports that 44 percent of users use the product more than five times daily, without a population, observation period or measurement method. Its claimed reduction of up to 80 percent in model token consumption concerns operating efficiency, not a measured improvement in legal outcomes. Other deployment examples in its startup application leave customer identities undisclosed.
Spellbook has named customer accounts involving Dropbox, Panasonic, KMSC Law, Alturas Capital Partners and Elevare Law. Panasonic's account concerns saving three weeks while building an RFP process, while other testimonials describe drafting capacity and keeping work inside the legal team. The current pricing page states use by more than 5,000 legal teams and includes those stories alongside separate firm and in house offerings. These examples establish identifiable use rather than only a logo wall. A consistent dated measurement protocol, comparison baseline and workload definition do not accompany the range of savings claims, so the figures are specific customer accounts rather than a reproducible general result.
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.
The AI processing policy limits submitted text to the requested response and rules out training, fine tuning and future improvement by Sonar or its model providers. Requests are isolated from other users' requests. Hosted knowledge bases have row level security, with customer hosted and dedicated enterprise alternatives. The Marketplace license adds confidentiality duties for customer data and limits disclosures to representatives with a need to know. Stored playbooks, knowledge and usage information sit outside the simple transient request description. Privilege and work product receive no specific treatment, and the permission model does not explain matter walls or how connected repository restrictions carry into retrieval.
Spellbook's agreement expressly protects privileged information and work product and requires confidentiality for staff and service providers handling customer content. Its iManage integration preserves existing file permissions and ethical walls. Third party models cannot train on customer content and operate under zero retention commitments. Spellbook itself may use deidentified content for customer specific personalization and retain inputs and outputs temporarily for quality assurance, while the terms restrict those uses and access. These permissions are narrower than unrestricted cross customer training, but they are not an across the board ban on learning from client content. Application history also has no single stated in use retention deadline.
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. Where the advice line is not the duty a product raises, the axis is read through the nearest professional duty it does raise: judicial conduct rules and the reviewing duty for products sold only to courts, and the duty to bill for time actually spent for products that draft time entries.
Sonar Legal targets legal professionals, including law firms and corporate counsel, and its interface keeps drafting and revision work inside familiar office applications. Partner interprets clauses, identifies risks, prepares legal memoranda and generates replies to counterparty comments. Those functions support legal judgment, while tracked edits and source links give the practitioner a way to inspect the result. An explicit boundary between software output and legal advice is unspecified in the product policy and linked Marketplace license. The claim of support across more than 100 jurisdictions does not identify limits tied to professional admission, competence duties or supervision of nonlawyer users.
Spellbook's terms identify output as information rather than formal legal advice and place verification and appropriate lawyer review with the customer. The product is directed to law firms and legal departments, with review controls inside Word and oversight of Associate's work across files. The agreement also warns that non English use can produce unsatisfactory results. That is a concrete language limitation, but not a tested jurisdictional coverage matrix or a map of which regulated services the software can provide. Human acceptance of a redline remains part of contract work, and an answer from a cited public source is not a substitute for checking its legal application.
AI Governance and Bias Disclosure
Published governance over model behavior: who owns it inside the vendor, what is tested before release, and what is disclosed about disparate output across matter types, parties, or populations.
Playbook authors can express preferred positions, acceptable fallbacks and unacceptable terms, then revise the proposed rules before sharing them. The guidance explains how vague instructions can cause the model to weaken an already favorable clause. This is useful control over an individual review, rather than an organization wide account of model governance. Sonar describes secure development practices based on Microsoft SDL and OWASP, but those practices concern software security. A person accountable for model behavior, release evaluation procedure, incident review process for incorrect legal answers and results across languages or legal systems are unspecified. The broad jurisdiction and language coverage claims therefore lack corresponding quality comparisons.
Spellbook's terms describe internal AI risk and prohibited use policies, staff education and regular reassessment of ethical and legal requirements. Its security material also cites a CMS Law opinion classifying the product under the EU AI Act and identifying applicable controls. These are concrete governance arrangements beyond a general promise to act responsibly. The product supports verification through citations and user approval of edits, but those controls are separate from evaluating a model before release. A named internal owner, a reproducible release test set and disclosed results about uneven behavior across populations or matter types are not specified in the available framework.
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.
AI requests pass through Sonar to selected model providers, with a promise that prompts and responses are discarded when the response is returned. Google Cloud hosts account services in Singapore, and Jina provides knowledge embeddings from Germany. Hosted knowledge bases, playbooks, templates and usage records have different persistence from those transient requests; application backups run hourly. The privacy policy retains information as needed without a fixed deletion schedule. Access protections include role based controls, account lockout, optional two factor authentication and remotely revocable sessions. A signed DPA is available on request. Specific backup erasure timing and a product incident notification timetable remain unspecified.
OpenAI and Anthropic process Spellbook requests under zero retention and no training commitments. Spellbook's own terms permit quality assurance copies of inputs and outputs for up to 90 days, restricted to authorized staff, and self service abuse monitoring for up to 30 days unless an investigation or law requires longer. Microsoft identity integration lets organizations enforce their authentication policies, and iManage access preserves source permissions. The agreement covers confidentiality, security incidents and deletion or return of customer material, while the processor register identifies downstream services. The ordinary lifetime of all application history is not stated as a single fixed period.
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.
The License Agreement link on Sonar's Microsoft Marketplace listing leads to Microsoft's standard publisher contract. That contract includes a warranty of substantial conformity with the product instructions and defense against intellectual property and applicable law claims, with exclusions for customer data, other products and unauthorized modifications. Subscription liability is capped at the greater of the preceding 12 months' fees or $500,000; certain customer data breaches have a higher cap, and defense obligations have exceptions to the limits. These are Marketplace acquisition terms, whose applicability to a separately purchased subscription is unclear. A specific warranty for correct legal conclusions or generated contract language is absent.
Spellbook gives an indemnity for specified intellectual property claims against its software and provides service warranties, while disclaiming the accuracy of generated output and responsibility for customer reliance on it. The general limit is twelve months of fees, with exceptions including confidentiality, security, data incidents and indemnification duties. The agreement includes mutual protection for specified data incident claims and gross misconduct. Beta Features have separate warranty exclusions and a CAD 10,000 cap. These are concrete protections for platform and data risks, but they do not provide an accuracy warranty compensating a client when an accepted clause or legal assertion is wrong.
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.
Word installation works through Microsoft Marketplace or administrator deployment to selected users and groups. Partner reads the active agreement and writes accepted clauses or replacements back into it, including optional tracked changes and author names. Outlook reads a thread, prepares responses and inserts text into a compose window, with a separate feature for visible revisions. Agentic email tasks require additional tenant administrator consent, which must be revoked separately when the add in is removed. OneDrive and SharePoint supply knowledge material to the web application. Email code authentication cannot enable features requiring Microsoft 365 consent, and older volume licensed Office editions lack the required browser component.
The iManage connection starts from Add from iManage in Ask or an Associate Project and authenticates the user's account once. The same connection then works in Word and Associate. Users select files through the attachment control or project file list; selected content becomes available as chat context or project references without a manual download. Answers cite the imported contracts, and precedents can guide review of incoming paper. Existing iManage permissions and ethical walls govern file access. The connection describes movement from iManage into Spellbook's working context; it does not promise automatic return of revised files to the source repository.
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.
Most Word processing takes place on the user's device, while explicitly initiated AI requests travel through Sonar's backend to model services. Google Cloud in Singapore hosts the manifest and authentication services, and the processor list places Jina embeddings in Germany. Hosted knowledge bases can use Google Cloud with row level security, a dedicated enterprise server or a customer's own environment. Dedicated application hosting is also an enterprise option. The locations used by Azure OpenAI, OpenAI, Anthropic, Vertex AI and optional OpenRouter routing are unspecified. The privacy policy permits transfers to Thailand and other jurisdictions, so local document editing does not mean every AI request remains local.
Spellbook uses AWS as its primary cloud provider and states that customer content is stored and processed in Canada and the United States. Third party model requests are handled under zero retention agreements with OpenAI and Anthropic, while Spellbook itself has separate retention rights for quality and abuse monitoring. The Word add in and Associate connect to shared cloud services, including optional sources such as iManage. The public description identifies storage and processing countries but does not offer a per tier residency selector or a dedicated customer deployment. A source system's location therefore does not confine every later processing step to that same location.
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.
Sonar describes encrypted transport, secure WebSocket connections, role based access, device management and regular security testing. Its internal access practices include hardware security keys, while customer sessions can use two factor authentication and are revocable. The architecture follows Microsoft SDL and OWASP guidance, which are development practices rather than independent company certifications. The processor page refers to SOC 2 Type II certified infrastructure for Jina embeddings, but does not establish an attestation covering Sonar's own application and operations. A named auditor, certificate or report period for Sonar is unspecified. Technical security explanations and the DPA request route are available without a sales demonstration.
Spellbook states that it is SOC 2 Type II compliant and provides a trust portal for reports, processor information and compliance resources. Its security program uses encryption, Microsoft identity controls and agreements governing model providers. The public security explanation also describes HIPAA support through business associate agreements where appropriate and a legal opinion on EU AI Act classification. These are different assurances with different scopes. The accessible product explanation does not establish the current SOC audit period, auditor and full service scope. The named SOC standard therefore supports a specific certification claim without treating every resource or regulatory statement as an equivalent independent security audit.
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.
The AI provider list names Microsoft Azure OpenAI, OpenAI, Anthropic and Google Vertex AI. Customers can choose or exclude providers, and optional OpenRouter routing supplies an alternative when the primary provider is unavailable. Sonar says it disables providers that permit training and opts out of OpenRouter prompt and completion logging. Jina creates embeddings for clause retrieval before selected knowledge accompanies the drafting request. The guide describes direct and routed workflows, but does not identify model versions or map each feature to an exact model and processing region. A commitment to notify customers of every model substitution is also unspecified.
Spellbook names OpenAI and Anthropic as language model suppliers and uses zero retention agreements that keep request and response content in memory for processing rather than persistent storage at those providers. AWS is the primary cloud provider, and storage and processing are described in Canada and the United States. The trust portal lists processors and records updates when suppliers change. Exact production model versions and a routing map for each feature are not specified. A public processor update history is not an explicit commitment to notify every customer whenever a model version or deployment location changes within the same supplier.
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.
A permanent free plan includes unlimited definitions, ten monthly Error requests, ten Partner requests and five uses of Sonar playbooks. Value costs $61 per user monthly or $49 per month with annual billing, covering 100 Error and Partner requests and 50 playbook and formatting requests monthly. Premium costs $99 monthly or $79 per month with annual billing and adds unlimited Word usage plus Outlook. Those plans restrict playbooks and formatting to Sonar's supplied options. Enterprise uses a quote and adds the web application, custom playbooks, house styles and a prompt library. Implementation charges and a standard cancellation arrangement for direct subscriptions are unspecified.
Spellbook prices the suite according to the number of team members on a license. Law firm and in house packages both include the Word add in and Associate, training and online support, with dedicated support for teams above ten. The in house offering includes a playbook build service. A seven day trial is available, with extended trials possible for larger organizations. Subscription amounts remain custom rather than a public per user rate. The terms describe annual renewal by default and included maintenance and support, but the price of a particular deployment and any separately agreed services remains part of its order.
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.
The product serves solo practitioners, small and midsize firms and corporate legal departments, with Word and Outlook as the main working environments. Its practice examples include corporate transactions, acquisitions, commercial contracts, financing, litigation and tax. Tables handles bulk agreement review, while knowledge hubs reuse internal precedents and memoranda. Sonar claims legal research across more than 100 jurisdictions and formatting templates for over 150 languages, including bilingual layouts and Asian scripts. Specific legal source coverage and comparative performance by jurisdiction are unspecified. The installation guidance provides concrete Office version limits, and custom knowledge, playbooks and styles belong to the enterprise offering.
Spellbook separates law firm and in house packages and organizes its principal tools around transactional drafting, review, negotiation and questions about agreements. Associate extends work across several files, and iManage access allows firms to use existing contract precedents. Industry material covers energy, healthcare, finance, technology, manufacturing and consumer businesses. The legal data connection adds government and court sources across many regions, but that source breadth does not establish tested competence in every practice. The terms warn about non English results, while Autonomous Contract Management remains in early access. A complete account of unsupported jurisdictions or specialized legal tasks is not specified.
The 12 legal signals, side by side
Recorded rather than graded. These are the questions a practitioner has to answer before a tool touches a client matter, and the answers are taken from public material only.
Client Data in Training
Can material a lawyer puts into this product be used to train a model?
Sonar's AI policy says submitted text is used only to generate the requested response, with no reuse for training, fine tuning or future improvement by Sonar or its model providers. Optional OpenRouter routing excludes providers that allow training, and prompt logging is disabled. The Marketplace license protects customer confidentiality but does not expressly prohibit model training. The separate privacy policy allows service improvement using account, usage and uploaded template information, without overriding the more specific AI request promise.
The terms prohibit third party models from retaining or training on customer content. Spellbook itself may process deidentified customer material to improve that customer's features and calibrate personalized systems, without using that personalization for other customers. It can also evaluate prompt and system improvements through limited quality monitoring. These own service improvement permissions are distinct from the provider prohibition and do not require a separate opt in in the standard agreement.
Prompt and Output Retention
How long does the product keep what a lawyer typed, and can that be set to zero?
For AI requests, Sonar promises to discard the input and generated response immediately after returning the result, without keeping a prompt archive or cache. That fixed zero retention statement concerns the processing path. Playbooks, house styles, hosted knowledge bases and usage records can persist separately, with information kept as needed under the privacy policy and hourly application backups. Their deletion periods are unspecified. The general website claim of no metadata retention is broader than the usage logging described in the privacy policy.
Model providers handle request and response content in memory under zero retention agreements. Spellbook may keep quality assurance inputs and outputs for up to 90 days, and self service abuse monitoring data for up to 30 days, with investigation and legal exceptions. Those are specific secondary uses rather than a stated ordinary lifetime for all application history. Deletion and return provisions do not establish a configurable zero history setting.
Ethical Walls and Matter Segregation
Does retrieval respect the firm’s ethical walls, or can the model read across them?
A playbook can remain private or be shared with every member of an organization; only its owner can edit a shared playbook. Standard hosted knowledge bases use row level security, and enterprise customers can have dedicated storage or host knowledge themselves. AI requests are processed separately and are not combined with another user's request. The Microsoft connection requires authorized users and tenant consent, but repository permission inheritance at retrieval time and formal matter conflict walls are unspecified.
The iManage integration applies existing iManage permissions and ethical walls to file access inside Spellbook. A user authenticates the source account once, then selects permitted files for Ask in Word, Ask in Associate or an Associate Project. Selected material becomes context for answers and review. This protection is specifically described for the iManage connection; it does not establish the same permission inheritance for every optional storage connection.
Third Party Request and Subpoena Notice
If someone subpoenas the vendor for a firm’s data, does the firm hear about it first?
The publisher contract linked from Sonar's Microsoft Marketplace listing requires notice before legally compelled disclosure, when notice is lawful, so the customer can seek a protective order. The same contract requires prompt notice of unauthorized use or disclosure of confidential information. Its scope is Marketplace orders. Sonar's own privacy policy separately permits disclosure required by law, regulation or court order without a corresponding notice sentence. A general request transparency report and equivalent direct subscription terms are unspecified.
Spellbook's terms require prompt notice of a legal or government demand for customer data and reasonable cooperation with efforts to obtain a protective order or contest disclosure. The confidentiality provisions also address compelled disclosure and protection of privileged material. These contractual steps give the customer an opportunity to respond to a demand. A published report enumerating government requests and their outcomes is not specified alongside that commitment.
Primary Law Corpus Provenance
Where does the law in this product come from, and does the vendor have the right to use it?
The web research service claims coverage of more than 100 jurisdictions in their native languages, but does not identify the underlying court, legislation or publisher collections. Its drafting knowledge base contains more than 50,000 selected clauses, and customers can add their own agreements and memoranda. That clause library is not an identified primary law collection. Source citations help the user inspect an individual answer, while acquisition rights, refresh schedules and jurisdiction by jurisdiction completeness remain unspecified.
The legal data connection names public sources including CanLII, Cornell Law, CourtListener, EUR Lex, BAILII and national government registers. Spellbook retrieves material directly from each source of record and follows its update cadence. The same workflow can use the customer's own contracts and iManage precedents. The named source list makes jurisdiction coverage inspectable, but does not state an individual license or public domain basis for every external collection or its separate market benchmark data.
Good Law Verification
Does the product tell you when the authority it just cited has been overruled?
Legal research answers include citations, and the litigation offering refers to statutes, precedents and case law. Neither that description nor the research workflow identifies a commercial citator, a subsequent history calculation or an indicator that an opinion has been overruled. Contract passage links in Partner and Tables address a different question: where the answer or extracted field came from. They do not establish the continuing authority of a court decision used in a memorandum.
Spellbook's legal data feature returns citations to public case law, statutes and registers, with currency tied to the operator of each source. Ask, Draft and Review can use that authority within the contract workflow. The feature does not describe a licensed citator or a method for checking subsequent judicial treatment of a case. A current link to the underlying decision is therefore separate from a signal that its legal proposition remains reliable.
Refusal and Uncertainty Behavior
What does the product do when the answer is not in the corpus?
When building a playbook from instructions or precedents, Sonar pauses and asks clarification questions if the instructions are ambiguous. The rule guidance also explains how an imprecise payment or confidentiality requirement can produce unwanted edits and shows how to preserve acceptable terms. These are concrete uncertainty controls for playbook construction. A comparable no answer path for unsupported legal research questions is unspecified.
High, medium and low labels in review results describe issue severity rather than calibrated confidence in the model's answer.
Source citations and selected precedent files provide context for Spellbook's answers and proposed redlines. A lawyer can inspect that material and accept or reject suggested changes. The product does not describe an automatic no answer response, an exposed confidence threshold or a required escalation when available material cannot support a conclusion. A contractual warning about unreliable output assigns verification to the user rather than defining an abstention mechanism.
Fabricated Citation Record
Does a public court record exist addressing fabricated or hallucinated legal citations in output from this product?
No court record naming Sonar Legal or its former AutoDox name was identified in the AI Hallucination Cases database as of October 16, 2026. The product generates contract language and also offers research and memorandum assistance, so citations can enter legal work through several features. This result concerns identified court incidents, not a test of citation accuracy. Contract source links and the ability to inspect research citations do not establish that every answer is correct or current.
Youssian v Village of Kings Point, 2026 NY Slip Op 26150, Supreme Court of New York, Nassau County, September 9, 2026, addresses counsel's fabricated quotations and misrepresented holdings. The opinion recounts counsel accepting Spellbook AI changes without repeating a citation check. It sanctions counsel $5,000, awards $10,000 in fees and strikes both petitions. The findings concern the lawyer's verification and response to the errors; the product attribution appears in counsel's explanation.
Bar Guidance Alignment
Has the vendor engaged in public with the ethics opinions its buyers are bound by?
Sonar is built for legal professionals and provides review controls such as tracked changes, source citations and editable playbook rules. Its compliance materials discuss privacy, security practices and Microsoft deployment requirements. They do not map the workflow to named bar opinions or jurisdiction specific duties concerning competence, supervision, confidentiality or disclosure of AI use. Broad research coverage across many jurisdictions is a product coverage claim and does not establish alignment with each jurisdiction's professional conduct requirements.
Spellbook's educational material addresses competence, candor and checking AI output in general terms, and the company has a relationship with the Canadian Bar Association. Its terms also require appropriate legal review and distinguish output from formal advice. Those positions concern professional responsibilities, but the available account does not map the suite to a named ethics opinion or a set of jurisdiction specific bar requirements. A partnership alone does not supply that mapping.
Billing and Fee Posture
Does the vendor address what happens to the bill when the work takes an hour instead of six?
The law firm offering emphasizes faster drafting, formatting and review, with TWL Law Group describing efficiency in cross border work. Sonar's pricing and token efficiency claims address the software cost rather than how a lawyer charges a client. Usage analytics can count generated clauses, errors and active users, but no client invoice record separates machine effort from lawyer review. Guidance on passing subscription charges through, reducing time entries or describing AI assisted work on a legal bill is unspecified.
Spellbook presents the law firm suite as a way to improve profitability and the in house suite as a way to retain more work internally. Customer stories describe time savings, including Panasonic's RFP process. The product's per user subscription is separate from a lawyer's charges to clients. A matter level invoice record of AI assisted time or instructions for disclosing those savings on a client's bill are not described in the current pricing offer.
Outside Counsel Guideline Readiness
Can a firm get this vendor through a client’s AI clause without a bespoke negotiation?
The processor register names Google Cloud in Singapore and Jina in Germany, alongside the available language model providers and optional OpenRouter routing. Technical pages explain local processing, request transmission, authentication and knowledge hosting options. A signed data processing agreement is supplied on request. These materials can inform a firm's vendor assessment, but they do not form a client specific AI disclosure package. Exact model versions, all inference regions and a template for client approval remain unspecified.
Spellbook identifies OpenAI and Anthropic, provides a processor register and explains provider retention and training restrictions. Its security resources address SOC 2 Type II, HIPAA support and an EU AI Act opinion, while the agreement covers privileged information and legal demand notice. Those materials support a firm's client review. They do not form a single client consent pack, and the firm's account must distinguish provider restrictions from Spellbook's own personalization and monitoring permissions.
Court Disclosure Support
If a judge’s standing order requires an AI disclosure, can the product produce one?
Word users can apply an AI revision as a tracked change with a chosen author and optional comment, or insert it without tracking. Playbook results include the supporting contract passage and the proposed replacement. Tables links extracted values back to their source, and Outlook has a separate feature for visible revisions. These features preserve parts of review work, but do not create an immutable record of prompts, model versions and lawyer approval. A court disclosure export or filing template is unspecified.
Word review retains proposed tracked edits, and answers based on contracts or public legal data include source citations. Those features preserve portions of the review process and allow a lawyer to check supporting material. An export combining exact model version, retrieved sources and human verification for each generated passage is not described. The available editing and citation records are therefore partial support for an account of AI use rather than a complete court certification package.
The questions both sides leave open
Derived from the records above rather than written, so it cannot favor either vendor. Take these into both conversations and ask each side the same question.
- Good Law Verification
Which one fits
Choose Sonar Legal if
- House style consistency matters alongside legal content. Sonar lets a reviewer adjust proposed paragraph styles before applying formatting to the whole agreement.
- The team wants legal drafting assistance in Outlook as well as Word. Sonar reads the active thread, prepares replies and supports visible revisions for review between colleagues.
- Review rules need careful authoring from a precedent or plain instructions. Sonar asks clarifying questions, supports fallback examples and lets the lawyer select which rules to run.
Choose Spellbook if
- A single instruction must lead to changes across a transaction's related files. Associate breaks a project into tasks and can revise party details, terms and other issues across the set.
- Negotiators need to compare a position with similar commercial agreements. Market filters comparisons by industry, jurisdiction and deal type and shows the prevalence of terms and alternatives.
- Existing files are distributed beyond Microsoft storage. Spellbook connects iManage and Microsoft repositories alongside Google Drive and Dropbox for retrieval and reuse in contract work.
In summary
Sonar Legal
Sonar Legal turns precedents or instructions into editable review rules and returns cited findings in Word. For lawyers preparing agreements alongside the surrounding correspondence, the AI Legal Index highlights its paragraph formatting review and separate Outlook revision tools. Enterprise access adds custom playbooks, house styles and the web application, while smaller plans use Sonar's supplied options.
Spellbook
Spellbook offers Review, Draft, Ask, playbooks and Market inside its contract suite. Associate handles projects spanning several files, including financing packages, disclosure schedules and employment materials. Connections include iManage, OneDrive, SharePoint, Google Drive and Dropbox. Autonomous Contract Management is an early access offering and should be distinguished from the current Word and Associate workflows.
Questions buyers ask
How does Sonar's bulk review differ from Spellbook Associate?
Sonar Tables extracts user selected data into a structured view with links to the relevant clauses and editable results. Associate plans and executes work across files, such as updating financing terms or preparing a disclosure schedule. An extraction table and a project that revises multiple files are different workflows. From the AI Legal Index, based on each vendor's own published materials as of October 16, 2026. No vendor pays for placement.
Can Spellbook Market automatically revise an unusual term?
Market currently compares terms and shows alternatives and percentage breakdowns. Its dedicated action for inserting a redline to correct a term that differs from market patterns is labeled coming soon. That planned action should be separated from the contract redlining already available through Spellbook's review tools. From the AI Legal Index, based on each vendor's own published materials as of October 16, 2026. No vendor pays for placement.
Which product has a permanent free plan?
Sonar has a free plan with unlimited definitions, ten monthly Partner requests, ten Error requests and five Sonar playbook uses. Spellbook offers a seven day trial. Sonar's enterprise features, including custom playbooks and the web application, are outside its free package. From the AI Legal Index, based on each vendor's own published materials as of October 16, 2026. No vendor pays for placement.
What work does Spellbook Associate coordinate across files?
It handles transactional projects such as financing materials, disclosure schedules and employment packages, carrying changes across related files. For transactional drafting tasks that span several files, the AI Legal Index identifies Associate's planning and execution workflow as distinct from a single agreement review. Lawyers still need to inspect the resulting work.
Neither vendor paid for inclusion, placement or a grade, and neither reviewed this page before it published. Everything on it comes from public material on the dates shown. How the index grades.
Sonar Legal is the current name of the AutoDox product. Spellbook's ACM and Market redline roadmap items are separated from available drafting functions here. Neither vendor reviewed this page.