C

Case Status

The product was a client portal before it was an AI product, and the portal still stands on its own. Real time case status, the branded app without passwords, secure messaging, document upload, stage videos and review prompts triggered by NPS scores are all conventional software. The sync with the case management system is an integration rather than a model. On top of that sits a named set of model features the vendor sells individually: Agentic AI Translation, Agentic AI Case Summary, Agentic AI Response and Prioritization, an AI Triage Agent, AI Automation Agents and Client Sentiment Tracking. The models power a core capability, and the product would still work without them as a communication and status system.

Nothing is published on accuracy or grounding for a system that drafts and translates messages sent to represented clients. No accuracy figure, evaluation, test set or error rate is published, and nothing describes how a drafted reply is grounded in the case record. The exposure is specific. Agentic AI Translation renders a firm's message into another language for a client who cannot check it against the original, and Agentic AI Case Summary condenses a matter for a lay reader.

Accuracy is not even asserted, and grounding is not claimed. The individual AI feature pages were not opened and may say more.

A review point is published and the rest of the control structure is not. The home page says Case Status uses case context to suggest responses that staff can review and send in seconds, which puts a person between the model and the client on the drafting path. The Agentic AI Response and Prioritization feature is described as ranking messages by urgency for staff rather than acting on them. Against that, the product also markets AI Automation Agents and claims that up to 80% of routine client communication is handled automatically.

That implies some messages reach clients without review of each one, and nothing published reconciles the two or says which categories run unattended. No threshold, escalation route or statement of what happens after a wrong message is sent is published.

One named firm with a dated figure, and everything else anonymized. Pond Lehocky is named on the home page with close to 15,000 active clients using the app across its workers compensation and social security disability departments. It gives a before and after review figure with a stated window: 4.6 with 417 reviews before, and 4.8 with more than 3,000 reviews within 24 months. Shawn Lehocky, Chief Executive Officer at Pond Lehocky Giordano, appears on record.

The three growth case studies are unattributed and identified only by practice and city: an SSD firm in Charleston WV that doubled caseload in a year, a PI firm in Chicago at 40% growth year over year, and a large PI firm in Kansas City with a 35% increase in case velocity. The firm count is inconsistent: 600 plus in the page header and more than 500 in the body. G2 ratings, badges and review quotes were set aside as directory material.

Confidentiality is asserted in general terms, and the agreement that would carry it is missing. The security commitment page describes multiple layers of encryption, least privilege development, mandatory two factor authentication, background checks and ongoing security training. The GDPR section states that Case Status acts as a data processor for law firms. But no law firm customer agreement is published; the only agreement on the site is the End User Terms of Use binding the firm's clients.

Nothing addresses training on client content, matter or tenant segregation, or privilege and work product, which matters because the platform carries attorney client communications by its own description. The privacy notice also grants the app access to contacts, microphone, calendar, reminders, SMS messages and continuous location. That is wider than a case status app needs, and nothing published explains why.

Nothing is published on the advice line for a product that generates and translates messages to represented clients. That matters because the reader is a claimant who cannot judge whether what arrives is legal advice. No disclaimer of any kind, no statement that AI drafted or translated content is not legal advice, no requirement that a firm tell its client about AI involvement, and no reference to ABA Formal Opinion 512 or any state bar guidance is published.

Pages checked were the home page, security commitment page, privacy notice, Terms of Use and its Limitation on Liability subpage. The End User Terms do address the reader as a client, and use that space to disclaim liability rather than to explain what the technology does. The individual AI feature pages were not opened and may say more.

No governance position is published. Across every page read there is no responsible AI page, no named owner accountable for the AI, no testing regime before release and no ISO 42001 or equivalent. The security commitment page names a Chief Technology Officer, Charles Lane, and the privacy notice names a Data Protection Officer, Andy Seavers, but both roles are framed around information security and privacy rather than model governance.

Nothing addresses whether sentiment scoring, urgency ranking or translation quality behaves evenly across languages or client groups. That is the live question for a product whose clients include immigration and workers compensation claimants writing in languages the firm may not read.

A generic privacy notice and a narrative security page, with the operative details missing or blank. What is published: multiple layers of encryption, least privilege, mandatory two factor authentication, password and lockout policies, employee background checks, ongoing security training, AWS hosting, and continuous compliance monitoring through Secureframe. What is not: any subprocessor list, any incident or breach notification practice, or any retention period.

The retention gap is specific. Under the heading on how long information is kept, the privacy notice of 17 July 2025 says no purpose in the notice will require keeping personal information for longer than, and the sentence simply ends. The number is missing from the published document. Deletion is promised once no legitimate business need remains, with an exception for backup archives.

A published allocation that is total, uncapped, and pointed at the wrong party. The Limitation on Liability subpage excludes liability for damages of any kind under any legal theory arising from use or inability to use the service, direct or indirect, whether in negligence or contract and even if foreseeable. It expressly names personal injury, pain and suffering, emotional distress, loss of data and loss of goodwill among the excluded categories.

There is no cap, because there is no fee: these are End User Terms binding the law firm's client, not the firm. The excluded heads of damage include personal injury, pain and suffering and emotional distress, in a product sold into personal injury, workers compensation, disability and mass tort practice. No agreement with the firm, indemnity, warranty on output or insurance position is published. A specific and readable allocation exists, and nothing in it runs toward either the firm or its client.

The integration set is named in full. Thirteen case management platforms are listed with individual integration pages: Clio, Litify, MyCase, Neos by Assembly, Filevine, CASEpeer, SmartAdvocate, SmokeBall, PracticePanther, CoCounselor, Meruscase, Docketwise and Salesforce. That places the product inside the plaintiff and small firm stack rather than beside it. Case Status describes itself as the communication layer living inside the case management platform, with staff working where they always have and everything syncing automatically.

One published customer account describes it running inside Litify in a frame. What syncs, in which direction, and what a firm must configure is not described on any page read. The individual integration pages were not opened and may say more.

Cloud delivery is named, and neither the tenancy model nor a region choice is stated. The security commitment page says Case Status runs on Amazon Web Services and explains the choice as scale and redundancy, not as a tenancy design. The privacy notice says servers are in the United States and that information from users outside the US will be transferred to and processed there. That fixes a default location but is not a residency offering.

No tenancy model is published, no region option is offered, and processing location is not separated from storage. The product's stated practice areas include immigration, so its users include clients outside the United States, which makes the lack of any region choice worth noting.

Detailed attestation particulars are published, with no way to obtain the report. The security commitment page names the auditor, Modern Assurance, as an independent third party accounting and auditing firm. It states the scope as all five trust services criteria by name (security, confidentiality, availability, privacy and processing integrity) and gives completion dates: SOC 2 Type I in September 2023 and SOC 2 Type II in July 2024.

HIPAA compliance is claimed on the basis that legal practices serve clients who are also patients, and Secureframe is named for continuous monitoring. Two gaps remain. There is no trust center or portal and no stated route for a buyer to request a report, so the evidence is described rather than accessible. And the most recent attestation stated is now over two years old, with ongoing audits expressed as an intention rather than a published later report.

Nothing is published about the model supply chain a customer inherits. The product markets six separately named AI capabilities and a whole Data and Agentic AI solution page. Yet no model, model provider, inference location or commitment to notify changes appears on any page read. There is no subprocessor list of any kind on the site, which is the usual place for this disclosure. AWS is named on the security commitment page, but hosting says where the software runs, not whose model reads client messages. No reference to an underlying model layer is published at all.

No pricing information at any level. There is no pricing page. The navigation, footer and resource menus were checked across the home page, the security commitment page, the terms index, the Limitation on Liability subpage and the privacy notice, and none carries a pricing entry. No tier names, packaging structure, unit of charge or figure is published, and every commercial route is a demo booking, a tour or a calendar link. Neither the shape of the pricing nor a number is visible.

Practice coverage is specific and firm segmentation is not. Eight practice areas each have their own page describing the work: personal injury, workers compensation, immigration, employment, disability, mass tort, estate planning and criminal. That is a real statement of where the product is aimed, and it hangs together: all eight are high volume matters where a client waits a long time with little news. What is missing is any segmentation by firm size, any in house or government position, and any statement of where the product stops. The customer evidence points to plaintiff and claimant firms of varying size, but the vendor never says so directly.