C

CasePeer

Artificial intelligence is present, named and real, and it is peripheral to what the product is, which is the C band. The AI is 8am IQ and it resolves to exactly two named features on the pricing matrix, a Writing Assistant and a Translation Assistant, described on the vendor's own AI page as living inside notes, tasks and text messages to correct grammar, simplify language, shift tone between professional, empathetic and assertive registers, and translate between English, Spanish and Arabic.

That is a communications aid layered on a case management platform whose value plainly stands without it. Two facts settle the grade rather than the marketing. The Basic tier ships with no AI at all, so a buyer can and does receive the product with the models removed; 8am IQ appears only on Pro and Advanced. And the platform's own account of what it is for, on the page it writes for machine readers, leads with case grading, medical treatment tracking, lien resolution, settlement distribution and KPI reporting, with 8am IQ appearing as one line in a fourteen-item capability list.

Recorded and expressly not credited: the AI-heavy work in this product class reaches the customer through third parties, the integration catalogue naming EvenUp, Novo and Practice AI for demand letters and case summaries and Arctrieval for records retrieval. Those are other vendors' models and the ground rules put them on the other side of the arrow. Verified 12 September 2026.

Accuracy is asserted without measurement and the grounding claim is made without any described method, which is the C band. R15 governs how heavily this reads, and the note says so first: the product generates no legal authority. 8am IQ rewrites the user's own text and translates it, so the limbs about primary authority, openable citations and citator status do not bite on this product class, and it is not being penalised for lacking them.

What does bite is the accuracy claim, and there is one. The vendor's AI page answers the question of trust with a grounding assertion, that every response points back to its source so the user can check it before anything goes out, and nothing published describes what a source means for a tone rewrite, how it is surfaced, or how a user would follow it. Against that, the governing agreement is markedly less confident than the marketing: Terms of Service Part III states that 8am IQ may produce errors, omissions or inaccuracies, that 8am does not warrant or guarantee the accuracy, completeness or reliability of any Output, and that the customer is solely responsible for evaluating accuracy before relying on it.

R37 rule 1 gives that to the agreement, and the gap between the two is recorded here rather than resolved. R40 keeps the floor off D: the accuracy content is not a bare no-hallucination claim, because a contractual limitation and an express review requirement sit behind it. No accuracy figure, test set, error rate or evaluation is published. Verified 12 September 2026.

A written commitment that the models work alongside a supervising professional, with real review surfaces and a real firm-level control, short of the full structure. The commitment is contractual rather than promotional, which is worth more: Terms of Service Part III provides that 8am IQ does not provide legal advice, that the customer may not rely on any Output as a substitute for professional judgment or human review, and that the customer is solely responsible for evaluating Output before relying on it.

The IQ Usage Policy repeats it in operative terms, that all outputs should be reviewed and resources verified. The review surface is inherent and adequate: the assistant proposes text inside a note, task or message that the user edits and sends, so nothing reaches a client without a person pressing send. The genuine control, and the strongest thing on this row, is that availability of 8am IQ is governed by a firm-wide setting managed by the Account Owner, which the agreement expressly directs the firm to configure in accordance with its own policies and any applicable professional or ethical obligations.

What holds it off A is that the control is binary and firm-wide. Nothing lets a firm permit the assistant to lawyers and withhold it from intake staff or case managers, both of whom are named users of this platform and both of whom draft client communications in it. No threshold, confidence signal or class-of-work boundary is published, and what happens after the system is wrong is answered only by a disclaimer. Verified 12 September 2026.

Named customers with full attribution and no figures attached to any of them, which is the B band in its own terms. Three testimonials on the pricing page carry a named individual, that person's role and their firm: the founder of Josh Jones Law, P.A., a co-founder of Sam and Ash, LLP, and the founder of Oscar A. Garza Law. Under R122(2) that is named customer evidence rather than outcome evidence, and it is counted here as attribution.

What is absent is the other half. Not one figure is attached to a named firm on the surfaces read, no deployment is dated, and no result is quoted with a basis, so a reader learns who uses the product and nothing about what changed for them. Two things are recorded and expressly not credited, because they are a different kind of thing. Review-platform aggregates are published in quantity, including G2 at 4.7 across 188 reviews and Capterra at 4.8 across 216, and satisfaction scores are not deployment evidence.

And the statistics on the AI page, that 69 per cent of legal professionals use AI for work and 38 per cent save one to five hours a week, are industry survey findings about the profession rather than measurements of this product, which the page does not claim they are. The case studies library at the testimonials page is published and was not opened; under R25 it corroborates rather than carries a grade resting on the attributed testimonials already read, and it is the artifact that would move this row if its studies carry dates and a stated basis. Verified 12 September 2026.

Substantive published commitments readable before signing, failing the limb R33 makes decisive. What is met, and met contractually in the parent agreement that this product's own footer adopts as its terms: no training on customer content, expressly and in unusually complete words, covering models developed by the vendor and by third parties alike; a stated position on what the third-party model provider may retain, which is the limb most records in this corpus never reach, the agreement providing that the LLMs used to deliver 8am IQ may retain IQ Content for up to thirty days and may not train on it; and an access model documented at the level a firm configures, with user permission groups, granular user permissions, two-factor authentication and role-based access control.

The limb that fails is privilege and work product, and it fails completely. Neither the agreement, the AI page, the security material nor the machine-readable company page addresses attorney-client privilege or work product at all, on a platform that holds the entire case file of a contingency practice including medical records, treatment histories and settlement negotiations. R33 makes that limb sufficient on its own to hold the record at B. Two further gaps are recorded: nothing describes segregation between matters or between users inside one firm beyond the permission model, and no retention period is stated for case content, the agreement deferring to a Data Retention policy whose period is not published. Verified 12 September 2026.

A real published position that reaches further than most records at this grade, including into supervision, and stops short of A on one limb that matters on this product. What is published, all of it readable before signing and all of it in instruments this product's own footer adopts: the platform is stated to be for legal, accounting and other client-based professionals located in the United States and Canada, which answers both who may subscribe and the jurisdictional boundary; the vendor states that it does not provide legal advice and is not engaged in the practice of law; the customer is required to use the platform in compliance with any rules of professional conduct applicable to its profession; and unauthorized practice is named as a prohibited activity twice, in the Terms of Service and again in the IQ Usage Policy, which bars engaging in the unauthorized practice of law or offering tailored legal advice without a qualified person reviewing the information, and adds that the models are not fine-tuned to provide legal advice and should not be a sole source of it.

That is the supervision dimension addressed rather than omitted, which is what usually separates B from C here. The limb that fails is who may operate the tool. Availability of 8am IQ is a single firm-wide switch held by the Account Owner, with no per-role or per-user control, on a platform whose named users include intake specialists and case managers and whose assistant drafts messages that go to clients. A firm cannot give the assistant to its lawyers and withhold it from its non-lawyer staff. Nothing on CasePeer's own marketing estate addresses professional responsibility at all. Verified 12 September 2026.

A published governance framework with real substance, short of testing results and a named owner, which is the B band. The framework is the 8am IQ Usage Policy, updated 19 August 2025, a standalone dated instrument that names CasePeer IQ expressly, is incorporated into the Agreement by the Terms of Service, and carries a stated enforcement ladder from a request to change through suspension or termination of the account.

Its content is a real acceptable-use regime rather than a principles page: it prohibits use of 8am IQ for illegal activity, for hateful, harassing or violent content, for malware, for activity carrying high risk of physical or economic harm including automated determinations of eligibility for credit, employment or public assistance, for fraud and deception, and for high-risk government decision-making in law enforcement, criminal justice, migration and asylum.

Three of its prohibitions are squarely bias and rights provisions: no facial recognition of private individuals, no classifying individuals based on protected characteristics, and no biometric identification or assessment. Alongside it the Tool Terms function as governance in their own right, carrying the no-training covenant, the thirty-day model retention limit, the human review requirement and the firm-wide availability control.

What A asks for is absent. Nobody inside the vendor is named as accountable for AI, nothing is published about what is evaluated before an IQ feature ships, and no testing results or findings are disclosed. The honest limit on the whole row is recorded: this framework governs what the customer may do with the tool, not what the vendor does when building it, and nothing addresses whether the assistant's output is even across the populations a personal injury practice serves, which matters on a product whose translation feature is aimed at clients who do not read English. Verified 12 September 2026.

Substantive published policy covering most of the ground, missing the subprocessor list and a vendor-side incident commitment, which is the B band and its two named examples exactly. What is published is specific and, on the AI side, better than this corpus usually sees: prompts and outputs are addressed directly rather than left to a general privacy policy, the agreement stating that the LLMs used to provide 8am IQ may retain IQ Content for up to thirty days, that it will not be used to train any model, and that inputs and outputs may be reviewed by aggregated human and automated means to monitor accuracy, safety and performance.

Access control is documented control by control: encrypted connections with multi-factor authentication for production access, network segmentation, unique authentication for production datastores, role-based access, two-factor authentication for all users, a documented password policy, centralized log management with automated alerting, annual third-party penetration testing with SLA-bound remediation, and background checks on all employees.

Deletion is addressed, customer data being deleted or de-identified after the applicable retention period on departure, with export and deletion rights preserved. Two gaps hold it at B. No subprocessor list exists for the platform or for the AI: the only third parties named anywhere in the agreement are payment processors, and the LLM providers behind 8am IQ are never identified. And there is no commitment by the vendor to notify the customer of a breach of the vendor's systems; the agreement's data incident regime runs the other way, obliging the customer to notify within 24 hours and reserving to 8am the sole right to decide whether any notice is given. No retention period is stated for case content itself. Verified 12 September 2026.

Liability is addressed only through a limitation regime that disclaims the exposure this product creates, which is the C band, and the drafting is unusually explicit about doing so. What is published and precise: total cumulative liability is capped at the net fees earned by the vendor during the three-month period immediately preceding the event giving rise to the claim; indirect, incidental, special, consequential, exemplary and punitive damages and loss of profits, revenue, goodwill, use and data are excluded; claims must be commenced within one year; and the platform is supplied as is and as available with all warranties disclaimed, including any warranty that information or documentation provided through it will be accurate, complete or compliant with applicable law.

The AI-specific provision is the sharpest and it points the same way: the agreement states that 8am will not be liable for any damages arising out of or related to errors or inaccuracies in Output, the customer's use of or reliance on Output, or any interruption or failure of 8am IQ. There is no indemnity running to the customer at all. The only indemnity in the agreement runs the other way, the customer indemnifying the vendor, and it extends to compensating the vendor for personnel time if the vendor must respond to a subpoena relating to the customer's account.

Recorded and not credited, because it is not a published position with any scope a buyer could rely on: a certificate of insurance is listed among the resources at the trust centre, available on request. No service level, no warranty of output, no indemnity and no insurance terms were located. Verified 12 September 2026.

Real integrations, named individually and grouped by what they do, short of the depth an implementer could work from. The pricing matrix names twenty-one connectors by name as a tier-differentiated feature list, and the machine-readable company page groups them by function with a line on each: LawPay for payments; QuickBooks Online, Milestones and Lead Docket for practice operations; EvenUp, Novo and Practice AI for AI-generated demand letters and case summaries; Arctrieval for medical and billing record requests; Case Status, Hona, Kenect and RingCentral for client communication, calling and fax; Dropbox, Google Drive, Microsoft Word and OneDrive for documents; Google and Outlook calendars; and Zapier.

Direction is described for the few where it matters most, Novo syncing completed documents back to case files and Arctrieval syncing records and provider details directly to cases, and the LawPay relationship is deeper than an integration, being a sister brand embedded for IOLTA-compliant trust handling. API Access is published as a named Advanced-tier feature and flagged as new. What holds it off A is documentation.

No public API reference for CasePeer was located; the developer portal on the parent's estate documents the payments and merchant referral APIs and is not this product's interface, and crediting it here would be the one-fact-two-axes error. For most connectors nothing states what objects move, in which direction, or what a firm must configure, and the vendor describes its API library as growing rather than as a specified surface. Verified 12 September 2026.

Cloud delivery is stated outright and neither the tenancy model nor the region is published, which is the C band. What is published is the host and the backup practice: the pricing matrix carries securely hosted with AWS and automated rolling backups as named security features, and the pricing FAQ repeats that data is hosted and backed up with Amazon Web Services in a certified environment. Naming the cloud provider is infrastructure, and under the ground rules infrastructure is not tenancy and not residency, so it lifts this record off D by making delivery explicit rather than implied and does no more than that.

Everything the band asks for above that is absent from every surface read. No region or availability zone is named, no residency option is offered or refused, and nothing addresses whether a firm's data can be held in a particular country, which is a live question for a platform sold in both the United States and Canada under an agreement that names both. No tenancy model is described, so a buyer cannot learn whether the deployment is multi-tenant or isolated, and nothing distinguishes where data is stored from where it is processed.

The processing question has a specific edge here that nothing answers: the agreement discloses that the LLMs behind 8am IQ may hold prompts and outputs for up to thirty days, and never says where those models run. One retrieval fact belongs on the record and does not carry the grade: the trust centre at trust.8am.com is Vanta-hosted and returned page metadata with no body to this index's fetcher on the date shown, a documented persistent limit rather than a fault of the site. Verified 12 September 2026.

Certification is real, named and current, and the evidence behind it is not reachable without a request, which is the B band. Two attestations are claimed in the vendor's own voice and named with specificity rather than displayed as badges: SOC 2 Type 2, stated to be currently attested with the report available through the trust centre, and PCI DSS v4.0.1 compliance through the embedded LawPay payment integration, alongside IOLTA trust accounting compliance by the same route.

Around them sits a control inventory published in unusual detail for a record at this grade, covering infrastructure, application access, operational security including annual third-party penetration testing with SLA-bound remediation and annual control self-assessments, and personnel and governance controls including background checks, confidentiality agreements, an annual board briefing on cybersecurity and privacy risk, and a whistleblower channel.

C does not fire: these are explained claims answering direct questions, not unsupported marks on a page. R16 is satisfied in the crediting direction and by both connectors, the parent's agreement naming 8am CasePeer expressly as a covered Practice Management Solution, and this product's own footer linking the parent's trust centre as its Security surface. What is missing for A is the evidentiary apparatus and a reachable route to it.

No auditor is named, no report period or observation window is given, and no certificate or report number is published. The trust centre is Vanta-hosted and returned metadata with no body to this index's fetcher, so the access flow could not be observed; the vendor's own description is that the SOC 2 report, the PCI attestations, a security questionnaire and a certificate of insurance are available upon request. R5 governs where the portal cannot be read: describe what was seen and grade the lower tier. Verified 12 September 2026.

The vendor confirms that large language models sit underneath and identifies none of them, which is the C band. The agreement is candid that models are involved and vague about whose: the Tool Terms state that 8am IQ is enabled via large language models and that the LLMs used to provide it may retain IQ Content for up to thirty days. No model is named, no version is given, no provider is identified, nothing states where inference runs, and no commitment to notify customers when any of it changes was located.

There is no subprocessor list anywhere on either estate that would answer the question by another route; the only third parties named in the agreement are payment processors. The sharper point on this row is a contradiction between the marketing and the agreement, and R37 rule 1 resolves it against the marketing. The AI page tells a buyer there is no third-party AI. The agreement tells the same buyer that the LLMs the vendor uses may retain their content for thirty days, and separately prohibits the customer from using IQ Content to improve any AI system whether developed by the vendor or a third party.

Read together the agreement contemplates third-party models, and a buyer relying on the page would be wrong about who processes their clients' case notes. One adjacent disclosure is recorded and expressly not credited to this vendor's own supply chain, because it answers a different question: the Tool Terms name Claude as an example of a third-party service that is not an 8am AI feature and is governed by the customer's own agreement with that provider. Verified 12 September 2026.

A buyer can learn what this costs without entering a sales process, which is the A band, and the tier contents are published at a level of detail few records match. Rates: Basic at 79 dollars, Pro at 119 dollars and Advanced at 149 dollars, each per user per month, stated on the pricing page and repeated as a direct answer in the pricing FAQ. The unit is named and its consequences are explained, the vendor stating that there is no limit on how many attorneys and staff may use the platform and that the rate depends on the number of active users on the account.

The structure is set out in a comparison matrix running to roughly seventy rows across nine groups covering case management, automation, intake and CRM, communication, security, administration, services, integrations and AI, with items marked as included, optional or add-on tier by tier, so a buyer can see precisely what a tier upgrade buys. What implementation adds is answered rather than left open: no long-term contract and no setup fee, tier changes in either direction at any time and prorated, cancellation at any time, guided implementation and live training included, with e-signature and data migration named as optional add-ons.

The AI is priced honestly within that structure, 8am IQ being included with Pro and Advanced at no additional cost and absent from Basic, which is stated on the pricing page, in the FAQ and on the AI page alike. Two limits are recorded and neither displaces the grade. Add-on prices are not published, so the cost of e-signature and data migration is not knowable in advance. And the agreement reserves additional fees where usage exceeds thresholds identified in the pricing pages or an order form, without publishing the thresholds. Verified 12 September 2026.

Who this product serves is documented precisely, with the boundary stated rather than implied, which is the A band. The segment is named and named exclusively: plaintiff-side personal injury and mass tort law firms, from solo practices to multi-location operations, a statement the vendor repeats as the organising claim of the product rather than as a marketing line. Roles are enumerated individually and the platform is priced and packaged around them, the audience being managing attorneys and firm owners, case managers and paralegals, legal intake specialists, and operations and firm administrators, each of which also appears as a persona the tiers are described against.

Practice coverage is not asserted broadly but demonstrated through the data model, the published feature set being specific to the personal injury case lifecycle: medical treatment tracking and provider management, injury tracking, lien and letter-of-protection tracking, a settlement calculator with distribution and expense tracking, case grading, custom mass tort tracking and statute-of-limitations alerts. The limits are stated directly, which is the limb most records leave open.

The vendor defines the product against generalist platforms in terms, saying that where generic legal platforms try to serve every practice area this one is purpose-built for plaintiff-side personal injury, and that generalist tools require firms to retrofit workflows. The agreement adds a jurisdictional boundary, the platform being for professionals located in the United States and Canada. R15 applies to the band's reference to in-house and government use: neither is applicable to a plaintiff contingency product and neither is claimed, and the record is not credited or penalised for them. Verified 12 September 2026.