DigitalOwl vs Tavrn: how they compare in 2026

The models are the entire product and there is nothing underneath them. DigitalOwl does one thing, converting unstructured medical records into structured data, and every part of that is model work: extraction from conventional and electronic health records, chronology construction, impairment and body part classification, identification of soft medical terms, condition status change detection, and a chat interface over the record set. The vendor states the AI is proprietary, built by in house experts and trained specifically for these use cases rather than adapted from a general tool. Connect exposes the model output as an API for other systems to consume, which is the clearest possible statement that the extraction is the asset. Remove the models and there is no product, only a file upload. Fifth consecutive A on this axis in this category. Entered under brief ruling 2 as an acquired product still sold under its own name, with Datavant named as owner.

Model first from founding, and the most operationally aggressive use of models in this category. The retrieval layer is the distinguishing case: AI voice agents contact medical providers the same day a request is placed and keep calling until records arrive, across thousands of providers in all fifty states, handling fax, email and portal channels. That is a model acting in the world on the firm's behalf rather than analysing documents it was handed, and it is the part of the product a firm cannot replicate without the models. Downstream, page classification and indexing categorise document types, chronology construction is generated, intake scoring extracts facts and classifies injury severity, and demand letters are generated from the structured output of the pipeline. Remove the models and what remains is a fax inbox. Fourth consecutive A on this axis in this category, which confirms the pattern noted from the first record: plaintiff side vendors were built model first because the underlying work was never automatable any other way.

THE FIRST PUBLISHED ACCURACY FIGURE IN THIS CATEGORY, and it is short of the evidence that would make it checkable. The vendor states the platform is third party tested at over 98 percent accuracy. In 45 records almost nothing on this index publishes a number here at all, and this vendor publishes one and attributes it to independent testing, which is materially more than its four category peers. What is missing is everything that would let a reader test it: no evaluator is named, no methodology, no sample size, no test corpus, no date, and critically no definition of what accuracy means for this task. Extraction accuracy could mean correctly transcribed values, correctly classified pages, correctly ordered events or correctly identified providers, and those are very different measures producing very different numbers. Held at B on that basis rather than A. The number is credited as a real disclosure and the note records that a figure without a definition or an evaluator cannot be verified or compared. Checked the self serve page, the SSP product page, the home page, the View and Connect product pages and the blog material on 29 Aug 2026.

Grounding is page level, structurally reinforced by owning the pipeline, and unmeasured. Chronology entries are hyperlinked with navigation back to the specific source page, output is stated to link directly to its source, and Bates stamping gives every page a stable identifier a reader can cite. The architectural argument the vendor makes is a real one and no competitor in this category can make it: because the same platform retrieved the records, classified the pages and built the timeline, the structured data carries into the demand without manual re keying, which removes a documented failure mode where a model drafts from unstructured files and misstates what the records show. Held at B because nothing is quantified: no extraction accuracy, no page classification error rate, no recall on treatment events, no measure of whether retrieval actually obtained the complete file, and no evaluation. One published phrase is noted and not credited, that where doubt is expected proof is embedded, which reads as an uncertainty commitment and specifies nothing. Checked the home page, the MedChrons page, the medical retrieval page and the published blog material on 29 Aug 2026.

A clear positioning statement exists and no operational boundary does. The vendor states plainly in published material that the platform is a tool to support legal professionals rather than replace them and that it cannot substitute for legal professionals' judgment, decision making or expertise. That is an unusually direct disclaimer of substitution and is more than most of this index offers. It is a statement about role rather than a description of the oversight model. Nothing published states whether any extraction or classification is applied without human review, whether a low confidence result is surfaced differently, what happens when the chat interface answers a question the records do not support, or what review the Self-Serve pipeline receives before output is delivered inside 24 hours. Checked the self serve page, the SSP product page, the View and Connect pages and the blog material on 29 Aug 2026.

Visibility is documented and boundaries are not, which are different things and are treated as such. The vendor publishes real oversight mechanics for the retrieval workflow: complete oversight to create and track requests, real time status tracking on every request, all fax and email inbounds in one place, and real time editing control over generated demand letters. A firm can see what the system is doing at any moment. What is absent is any statement of what it may do unattended. Autonomous voice agents contacting medical providers and persisting until records arrive is outbound action on a firm's behalf to third parties holding a claimant's health information, and nothing published describes what an agent is authorised to say, whether it identifies itself as automated, what happens when a provider disputes a request, or where a human must intervene. Tracking an action after the fact is not a bound on it. Checked the medical retrieval page, the home page, the MedChrons page and the blog material on 29 Aug 2026.

Independent recognition and a corporate outcome are both dated and checkable, and no customer is named. Independent and verifiable: the Self-Serve legal platform was named an Innovative Product Winner in the 2026 BIG Innovation Awards, a named programme with a dated announcement, and the company was acquired by Datavant in 2025, which is an arm's length transaction by a major health data company and is the strongest single validation event on this record. Published operational figures are specific: up to 72 percent of medical record review time saved, review time reduced by up to 50 percent, page count reduced by 90 percent, and a thousand page record reviewed in minutes. Held at B rather than A because no law firm, carrier or customer is named anywhere in located material, no case study with methodology exists, and every efficiency figure carries an up to qualifier with no baseline, sample or period. Same treatment as Exterro, which reached B on independent analyst placement with no named customer.

Operational figures are specific and no customer is identifiable. Published: 50 to 70 percent reductions in record review time, average retrieval turnaround cut by 12 days against standard practice, 70 percent faster turnaround than standard retrieval methods, chronologies delivered in under 24 hours, roughly 1,000 retrieval requests handled daily across all fifty states, and intake conversion improved by 35 percent. Throughput claims of that specificity are useful and the daily request volume is the kind of number a vendor cannot easily inflate without being caught by its own customers. Against that: no law firm is named anywhere in located material, and the one return on investment claim, a reported 6x return on platform cost within a single quarter, is published with the parenthetical that the testimonial is on file, which is a claim whose evidence the vendor states exists and declines to show. Held at C on that basis. Same placement as Supio and for the same reason: the operation is described, the customers are not.

A strong instrument set, and a structural conflict question this vendor raises more sharply than anything else on the index. Published: SOC 2 Type II, HIPAA and GDPR compliance, a business associate agreement offered to customers and covered entities, secure storage of uploaded records, and customer controlled manual deletion at any time. That is the second BAA on the index and a genuine set of protections for the claimant health information at stake. THE UNADDRESSED QUESTION: this vendor sells to both sides. Its legal offering serves plaintiff personal injury, medical malpractice and mass tort firms, and its insurance offering explicitly includes streamlining review of inbound settlement demand packages and bodily injury claims for carriers. Those are the same disputes viewed from opposite ends, and nothing published describes what separates a plaintiff firm's uploaded records and case insights from the carrier side of the business. No treatment of attorney client privilege or work product was located either. Held at B on the strength of the instruments, with the conflict question named rather than buried.

The right instrument for this category is offered, which no prior record on the index has done, and privilege itself is still unaddressed. The vendor states a business associate agreement for HIPAA covered workflows. A BAA is the actual contractual mechanism HIPAA requires between a covered entity and a business associate handling protected health information, and offering one is materially different from stating HIPAA compliance as a posture: it is the document a firm's compliance review asks for and can execute. Alongside it, a SOC 2 Type II attestation, AES 256 bit encryption at rest and TLS 1.2 in transit, which is the first record in this category to name the transit protocol as well as the storage cipher. What is absent is the legal dimension: no treatment of attorney client privilege or work product, and nothing about the confidentiality of the demand strategy and case valuation the platform generates. Graded B alongside EvenUp and Supio, and the note records that this record earns its B on a different and arguably more practical basis than either.

The claim is made and no framework sits behind it. The vendor publishes an explicit statement that the platform is a tool to support legal professionals and not replace them, and that it cannot substitute for their judgment, decision making or expertise, made in the context of addressing attorney resistance to adoption. That is a direct and relevant position and it is the reason this grades above the D that most of this category receives. What is absent: no statement that output is not legal advice, no positioning on the supervising attorney's duty over machine extracted medical facts that will support a demand, no guidance on verification before reliance, and no engagement with any bar guidance. A sentence in a blog post addressing adoption anxiety is a claim rather than a professional responsibility framework. Checked the blog material, the self serve page, the product pages and the site navigation on 29 Aug 2026.

Not located. Two aspects of this product engage the question and neither is addressed. The platform scores incoming cases on a numeric scale with configurable thresholds, which is a machine assessment feeding a decision about whether to accept a client, and nothing published addresses the attorney's role in that determination. Separately, generated demand letters go out under the firm's name to adverse insurers. No statement that output is not legal advice, no positioning on the supervising attorney's duty, and no engagement with any bar guidance was located. Checked the home page, the products pages, the blog material and the site navigation on 29 Aug 2026.

The vendor tells buyers to demand exactly the disclosure it does not make, which is the third instance of this pattern in the pull after LinkSquares and FinregE. Its own published buyer guidance advises firms evaluating medical chronology software to ask whether the platform has undergone independent bias and accuracy testing, and states that reputable providers will be transparent about their validation processes and able to demonstrate their AI produces reliable, unbiased results across different case types and medical conditions. DigitalOwl publishes an accuracy figure and nothing on bias. No bias or fairness testing, no validation methodology, no AI policy, no model card, no evaluation output, no accuracy monitoring, no named governance body and no ISO 42001 were located. The untested risk is concrete for a product extracting clinical facts: documentation quality and diagnostic language vary systematically across providers and patient populations, and an extraction model's performance differences across those populations would shape which injuries appear in a demand. Fifth consecutive D on this axis in this category.

Nothing published about how the models are governed, evaluated or monitored. No AI policy, no model card, no bias or fairness testing, no evaluation methodology, no accuracy monitoring, no drift statement, no named governance body and no ISO 42001. The category risk recurs in its sharpest form on this record: intake scoring assigns a numeric case score with configurable thresholds based on extracted facts and injury severity classification, which is a model deciding which injured people a firm is likely to take on. Any systematic tendency in that scoring falls on prospective claimants who are never told a score existed, and nothing published indicates it has been examined. Fourth consecutive D on this axis in this category. Checked the home page, the products pages, the blog material and the site navigation on 29 Aug 2026.

The best data control disclosure in this category, and the training question remains unanswered. Published and specific: records uploaded to the Self-Serve portal are securely stored and the customer can manually delete their data at any time, which is a user exercisable control rather than a policy assurance and is the only such control located in this category. Alongside it a business associate agreement, SOC 2 Type II, HIPAA and GDPR compliance, and a published trust page. GDPR compliance carries data subject rights machinery that matters here because the individual whose records these are is a third party to the customer relationship. Held at B rather than A because nothing states whether uploaded medical records, generated chronologies or extracted structured data are used to train or improve the proprietary models, no retention period is published for data the customer does not delete, and nothing describes what happens to derived structured output when the source records are deleted. A deletion control is a strong answer to how long, and not an answer to what for.

Protection is specified with unusual technical precision and use is not addressed at all. Published: AES 256 bit encryption for stored data, TLS 1.2 for network communications in transit, a secure upload portal, a SOC 2 Type II attestation and a BAA for HIPAA covered workflows. Naming the transit protocol version alongside the storage cipher is more specific than any other record in this category manages. What was not located is any statement on whether medical records, chronologies, demand drafts or intake data are used to train or improve models, how long content is retained, or whether a firm can require deletion at matter close. The gap is wider here than for peers because this vendor holds records it retrieved itself directly from providers, so it is the original recipient of claimant health information rather than a processor of files a firm already had. Checked the home page, the products pages, the blog material and the site navigation on 29 Aug 2026.

No published position located beyond the BAA, which allocates HIPAA obligations between the parties rather than product liability. Nothing was found on liability for AI output, warranty, service levels or remedy. The exposure is asymmetric in a way particular to this product: a missed diagnosis, an omitted provider or a misdated treatment in an extracted chronology understates an injury, and on the plaintiff side that flows into a demand that settles low while on the carrier side it supports a denial, so the same extraction error harms a claimant in both directions. The published 98 percent accuracy figure implies a residual error rate by its own arithmetic and no published position addresses who bears it. Checked the self serve page, the product pages, the trust page references and the site navigation on 29 Aug 2026.

No published position located beyond the BAA, which allocates HIPAA obligations rather than product liability. Nothing was found on liability for AI output, warranty, service levels or remedy. The exposure has a shape unique to this record: the retrieval layer's failure mode is an incomplete file rather than a wrong sentence, and a chronology built on records that were never obtained will look complete and consistent while omitting an entire course of treatment. Nothing published addresses what happens when retrieval misses a provider, and no service level commitment was located for a platform whose value proposition is turnaround time. Checked the home page, the medical retrieval page, the products pages and the site navigation on 29 Aug 2026.

A real integration product exists and it is aimed at the other half of the customer base. Connect is a medical data API allowing clients to retrieve and integrate structured output into rules engines, workbenches and custom workflows, and it is sold as a named product rather than mentioned as a capability, which is genuine integration substance. The language and the named destinations are insurance systems: rules engines and workbenches are carrier claims and underwriting infrastructure, not law firm systems. No legal case management system is named anywhere in located material, no document management connector, and no integration with the platforms a personal injury firm runs on. A plaintiff firm gets a self serve upload portal and an API it would need to build against itself. Compare Tavrn at B, naming Filevine, Litify, Clio and Smokeball alongside API access. Checked the Connect product page, the self serve page, the SSP product page and the site navigation on 29 Aug 2026.

The strongest integration footprint in this category. Four case management systems named, being Filevine, Litify, Clio and Smokeball, which is more than any plaintiff side peer, and the Smokeball relationship is corroborated by a listing in that vendor's own partner marketplace rather than asserted only by Tavrn. API access is separately offered for connection to systems without a native connector, which is the element independent review identified as missing from a competitor in an adjacent category. Records are stated to be delivered pre organised, named and tagged directly into the case file, which describes the integration doing useful work rather than merely existing. Held at B rather than A because no API documentation was located, nothing states what moves in which direction with any named system, and no authentication, permission or field mapping detail is published. Top of the B band. Checked the products pages, the blog material, the partner marketplace listing and the site navigation on 29 Aug 2026.

Nothing located. No hosting provider is named, no region or data residency commitment is published, and no single tenant or dedicated instance option is described. Stated GDPR compliance implies the vendor handles European personal data and says nothing about where it is processed or stored, and a company holding United States claimant medical records under HIPAA alongside European obligations is exactly the case where residency would be stated if it were designed for. Compare Supio at B, which names data centres in three countries. Checked the self serve page, the SSP product page, the trust page references, the product pages and the site navigation on 29 Aug 2026.

Nothing located. No hosting provider is named, no region or data residency commitment is published, and no single tenant or dedicated instance option is described. A passing reference to the broader cloud environment appears in security material without identifying it. For a platform that retrieves claimant medical records directly from providers across all fifty states and holds them, the location and tenancy of that data is a question a firm's HIPAA diligence asks alongside the BAA, and the BAA governs obligations rather than geography. Checked the home page, the products pages, the security material and the site navigation on 29 Aug 2026. Compare Supio at B, which names data centres in three countries.

Three regimes, a trust page, and the second BAA on the index, held below A on currency alone. Published: SOC 2 Type II, HIPAA and GDPR compliance, a business associate agreement offered to customers and covered entities, and a dedicated trust page published as Trust at DigitalOwl and linked from product pages. Offering a BAA is the instrument HIPAA actually requires rather than a claim about compliance, and it puts this record alongside Tavrn as one of only two on the index to do so. Held at B rather than A on the same absences that separate the tiers: no auditing firm is named, and no examination period, scope or certificate date is published for the SOC 2, so currency cannot be established at all. That is the specific gap EvenUp closes with a dated April 2026 recertification to reach A, and it is the only material difference between the two records on this axis. Correction candidate: the trust page itself was not read in this pass and is the surface most likely to carry dates and scope.

A named attestation, the right HIPAA instrument, specific cryptography, and an inconsistency in the vendor's own language that is recorded rather than smoothed over. Published: a SOC 2 Type II attestation, a BAA for HIPAA covered workflows, AES 256 bit encryption at rest and TLS 1.2 in transit. THE INCONSISTENCY: vendor material elsewhere describes the same posture as a security program aligned with SOC 2 criteria and HIPAA compliance standards. Aligned with criteria and holding a Type II attestation are materially different claims, the first being a self assessment and the second an independent examination over a period, and a reader encountering only the weaker formulation would conclude no attestation exists. Under the check the noun rule this is exactly the kind of slippage that matters. The stronger claim is credited because the vendor makes it in its own material, and the inconsistency is named here so a later reader can resolve it against the actual report. Held at B on the standard absences: no auditing firm named, no examination period or date published, and no trust centre or self serve documentation request route located. Correction candidate in both directions.

Nothing located. The AI is described as proprietary and built by in house experts, which is an ownership statement rather than a supply chain disclosure, and no foundation model provider, model family or version is named. No subprocessor list was found. The acquisition adds an unaddressed dimension rather than resolving one: Datavant is a health data platform company and nothing published describes what data or processing relationship now exists between the two, which for a business associate handling protected health information is a question a covered entity is entitled to have answered. Checked the product pages, the self serve page, the blog material including the acquisition post and the site navigation on 29 Aug 2026.

Nothing located. No foundation model provider, model family or version is named, no subprocessor list was found, and nothing states whether third party models sit underneath any component. The gap is broader than for peers because this product runs more distinct model types: voice agents conducting telephone calls with providers, page classification, chronology generation, intake scoring and document drafting. Voice in particular almost certainly involves a speech provider and none is identified. For a platform operating under a BAA, every subprocessor touching protected health information is a party the covered entity is entitled to know about, and none can be identified from public material. Checked the home page, the products pages, the blog material and the site navigation on 29 Aug 2026.

No pricing published by the vendor at any level. No price, no range, no unit of charge, and no indication of whether the Self-Serve portal prices per case, per page, per record set or by subscription, which matters because a self serve product is precisely the configuration where a firm expects to see a price before engaging. An independent review platform reports a starting figure of $7,500 per month, which is third party reconstruction rather than disclosure and is recorded as context rather than credited, and if approximately right it places the product well outside small firm reach. The contingency fee point recorded across this category applies: case costs are advanced against a claimant's recovery, so undisclosed cost is undisclosed cost to an injured person. Checked the self serve page, the SSP product page, the pricing navigation and independent review material on 29 Aug 2026.

No price published at any level, with one genuine differentiator that does not reach the bar. A free trial is offered, which is the only self serve entry point in this category and is materially better than the universal demand a demo pattern of its peers, since a firm can evaluate the product without a sales conversation. But a trial discloses the product, not the price. No figure, range, tier structure or unit of charge is published, and nothing states whether pricing runs per record page, per retrieval request, per chronology, per demand or per seat, which for a retrieval driven product are wildly different economics. Independent material describes the target as contingency fee practices requiring predictable costs, which implies a predictable model without disclosing it. The contingency fee point applies as it does across this category: undisclosed per case cost is undisclosed cost to an injured claimant. Checked the home page, the products pages, the pricing navigation and independent review material on 29 Aug 2026.

The broadest coverage in this category by a wide margin, spanning both sides of the market. Legal coverage names personal injury, medical malpractice and mass tort. Insurance coverage names underwriting, claims review, post issue audits, long term care application review for material misrepresentation, workers compensation, property and casualty claims, bodily injury, and review of inbound settlement demand packages. Record coverage extends to both conventional and electronic health records, which is a meaningful technical distinction since EHR exports and scanned paper files present different extraction problems. Held at B rather than A because breadth is asserted rather than characterised: no statement of which record formats, provider systems or document types the extraction handles reliably, no jurisdictional scope, and no indication of whether the 98 percent accuracy figure holds evenly across record types or was measured on a favourable subset. Breadth without a characterised boundary is the same gap as everywhere else on this axis.

The broadest practice coverage in this category and the most concrete operational reach. Practice areas named: personal injury, medical malpractice, workers compensation, SSDI and general liability, which is wider than the personal injury and mass tort focus of the peers built so far and reaches into disability and comp work with different record profiles. Operational coverage is stated numerically: thousands of providers nationwide, all fifty states, roughly 1,000 retrieval requests handled daily, with fax, email and portal channels supported, which for a retrieval product is the coverage question that matters since a provider the platform cannot reach is a record the firm chases by hand. Independent material identifies the firm size target as roughly 10 to 75 professionals. Held at B rather than A because the reach is asserted rather than characterised: no provider network is enumerated or described, no success rate on retrieval requests is published, and nothing indicates whether coverage is even across states or concentrated where the largest provider systems sit.

Silent. The quoted commitment is the strongest data control statement in this category and it governs deletion rather than use: a customer can remove uploaded records at will, and nothing states whether those records were used to train or improve the proprietary models before deletion or whether anything derived from them persists afterward. No statement in either direction was located. The question is pointed for this vendor because the AI is described as proprietary and trained specifically for these use cases, and the obvious training corpus for a medical extraction model is medical records, which this platform receives at volume from both law firms and insurance carriers. Recorded as silent, not as a negative commitment. Correction candidate: the trust page at Trust at DigitalOwl was not read in this pass. Checked the self serve page, the SSP product page, the product pages and the blog material on 29 Aug 2026.

Silent. The quoted phrase is the most specific data statement the vendor publishes and it describes protection at rest, not purpose. No statement was located in either direction on whether retrieved medical records, generated chronologies, demand drafts or intake data are used to train or improve any model. The gap is wider on this record than on its peers for a structural reason: this vendor retrieves records directly from providers rather than receiving files a firm already holds, so it is the original recipient of a claimant's protected health information and accumulates it across every firm it serves. A BAA is offered for HIPAA covered workflows, and a BAA governs permitted uses and disclosures under the agreement rather than being a public statement about model training. Recorded as silent, not as a negative commitment. Checked the home page, the products pages, the security material and the blog on 29 Aug 2026.

Customer configurable, and the only record in this category to reach this value. The vendor states that medical records uploaded to the Self-Serve portal are securely stored and that the customer can manually delete their data at any time, which is an exercisable control in the customer's own hands rather than a retention policy they must trust. That is materially better than the not addressed recorded on all four category peers. Held at configurable rather than the zero retention value because deletion is manual and elective rather than automatic or configurable to a default: nothing states a retention period for data a customer does not delete, nothing describes automated deletion at matter close, and nothing states whether generated chronologies, extracted structured data and API delivered output are removed when the source records are, which for a platform whose product is derived data is the load bearing question.

Not addressed. No retention period is published for retrieved medical records, generated chronologies, demand drafts or intake scoring data, and nothing indicates whether a firm can require deletion at matter close or settlement. The question is sharper here than for peers because the vendor holds records it obtained itself from providers, so deletion by the firm from its own case management system does not reach the vendor's copy. Checked the home page, the medical retrieval page, the MedChrons page and the site navigation on 29 Aug 2026.

Not addressed, and this record raises the sharpest version of the question on the index. No permission model, matter level restriction or tenant segregation description was located. The structural issue is not internal to a firm but across the vendor's own market: DigitalOwl sells its legal product to plaintiff personal injury, medical malpractice and mass tort firms, and sells its insurance product to carriers for bodily injury claims, claims review and specifically for streamlining review of inbound settlement demand packages. Those are the same disputes from opposite sides. Nothing published describes what separates a plaintiff firm's uploaded records, extracted chronologies and case insights from the carrier side of the business, whether the two operate on shared infrastructure, or what governs staff access across them. A firm's own compliance review would ask this before uploading a client's records, and public material does not answer it. Checked the home page, the self serve page, the View and Connect product pages and the site navigation on 29 Aug 2026.

Not addressed. No permission model, matter level access restriction or tenant segregation description was located. The platform integrates with four case management systems that carry their own permission models and nothing states whether Tavrn inherits them or applies its own. A specific unanswered case: the retrieval operation runs roughly 1,000 requests daily across many firms through shared fax and email inbound channels, and nothing published describes how inbound records are attributed to the correct firm and matter or what prevents misrouting. Checked the medical retrieval page, the products pages and the site navigation on 29 Aug 2026.

Not addressed. No government or law enforcement request clause, no commitment to notify a customer before producing their data, and no transparency report were located. A business associate agreement would govern permitted uses and disclosures between the parties and is not a public notice commitment. The vendor holds claimant medical records from both plaintiff firms and insurance carriers, and following the Datavant acquisition sits inside a larger health data organisation, which broadens rather than narrows the range of parties a request could reach. Checked the self serve page, the trust page references, the product pages and the site navigation on 29 Aug 2026.

Not addressed. No government or law enforcement request clause, no commitment to notify a customer before producing their data, and no transparency report were located. A BAA would govern permitted disclosures between the parties and is not a public notice commitment. The vendor holds claimant medical records it retrieved directly from providers across all fifty states, which makes it a single point of access to health information for many claimants at once. Checked the home page, the products pages, the security material and the site navigation on 29 Aug 2026.

Not addressed, and inapplicable in the usual sense with a live residue. The platform operates on the customer's own uploaded medical records rather than a published law corpus, so there is no external legal source to name, license or date. The residue is what the proprietary extraction models were built on. Recognising soft medical terms, classifying impairments by body part, detecting condition status changes and distinguishing conventional records from electronic health record exports all imply training on very large volumes of clinical documentation, and nothing published states whether that corpus was licensed, synthetic, de-identified, publicly sourced or accumulated from customer uploads. The Datavant acquisition makes the question larger rather than smaller and nothing published addresses it. Checked the product pages, the self serve page and the blog material on 29 Aug 2026.

Not addressed, and inapplicable in the usual sense. The platform operates on the customer's own case records rather than a published law corpus, so there is no external legal source to name, license or date. The live residue is what the models were built on: page classification distinguishing lab reports from medication lists, injury severity classification, and chronology construction all imply training on substantial volumes of clinical documentation, and nothing published states whether that corpus was licensed, synthetic, publicly sourced or accumulated from records retrieved on behalf of customers. That last possibility is the one a claimant would care about and it is neither confirmed nor excluded. Checked the products pages, the MedChrons page and the blog material on 29 Aug 2026.

Not addressed, and close to inapplicable. The product extracts and structures medical facts rather than researching legal authority, and its outputs are chronologies, summaries, timelines and medical evidence supporting demand letters rather than legal argument, so there is little authority for a citator to check. No research provider or citator is named. Recorded as a scope fact rather than a disclosure failure, consistent with the treatment on Tavrn, with the same caveat: the platform supports demand letter production and demand letters commonly cite authority on liability, and nothing published states whether this product generates any or where it would source it. Checked the self serve page, the View product page and the blog material on 29 Aug 2026.

Not addressed, and closer to inapplicable than for the other records in this category. Tavrn's outputs are fact driven rather than authority driven: chronologies summarise treatment and demand letters are stated to draw on injuries, treatments and damages extracted from the records themselves. No research provider or citator is named and no legal authority appears in described output, so unlike EvenUp and Supio, which integrate Thomson Reuters research, there is little for a citator to check. Recorded as a scope fact rather than a disclosure failure, with the caveat that demand letters in practice often cite authority on liability and nothing published states whether this product does or where it would source it. Checked the home page, the products pages and the blog material on 29 Aug 2026.

Not addressed, and the published accuracy figure makes the omission more visible rather than less. A stated over 98 percent accuracy implies a residual error rate by its own arithmetic, and nothing published describes how the remaining cases behave: whether a low confidence extraction is flagged, whether an illegible page is reported rather than skipped, whether a gap in the treatment timeline is surfaced, or whether the chat interface will decline to answer a question the records do not support. Publishing a headline accuracy number while describing no behaviour at the margin tells a reader how often the system is right and nothing about how a user would recognise the times it is not. Checked the self serve page, the SSP product page, the View product page and the blog material on 29 Aug 2026.

Not addressed, with one published phrase that gestures at it and specifies nothing. The vendor states that where doubt is expected, proof is embedded, which reads as an uncertainty commitment and describes no behaviour: nothing states whether a low confidence extraction is flagged, whether a gap in the treatment timeline is surfaced, whether an illegible page is reported, or whether the system will decline. The retrieval layer raises the sharpest version of the question and it is unanswered: nothing describes what happens when a provider cannot be reached or returns an incomplete file, and whether the resulting chronology is marked as built on partial records. A chronology that looks complete over records that were never obtained is the dangerous output of this product. Checked the home page, the medical retrieval page, the MedChrons page and the blog material on 29 Aug 2026.

None located, with the instrument named. General web searches combining the vendor name with court, order, sanction, fabricated citation and medical record terms returned nothing on 29 Aug 2026, and no named docket database or court record tracker was searched. Recorded as a statement about what this search found, not as a clearance. The exposure shape is a misstated or unsupported medical fact rather than a fabricated legal citation, since the product generates no legal authority, and the adverse finding to look for would be a court addressing an extracted chronology that misrepresented the underlying records.

None located, with the instrument named. General web searches combining the vendor and product names with court, order, sanction, fabricated citation and demand letter terms returned nothing on 29 Aug 2026, and no named docket database or court record tracker was searched. Recorded as a statement about what this search found, not as a clearance. The exposure shape is narrower than for the litigation drafting products in this category since output goes to insurers rather than courts, and the analogous risk is a misstated treatment fact or a fabricated entry in a chronology attributed to a source page that does not support it.

Not addressed. No named ethics opinion, no ABA Formal Opinion 512, no state bar guidance and no engagement with professional conduct rules was located. The vendor does publish a statement that the platform cannot substitute for legal professionals' judgment, decision making or expertise, which is graded on the UPL axis and is a positioning statement rather than engagement with any professional guidance framework. Fifth consecutive record in this category with this value, and the pattern now holds across every plaintiff side vendor built. Checked the blog material, the self serve page, the product pages and the site navigation on 29 Aug 2026.

Not addressed. No named ethics opinion, no ABA Formal Opinion 512, no state bar guidance and no engagement with professional conduct rules was located, despite the vendor publishing a substantial blog library comparing tools and advising firms on adopting AI in personal injury practice. That library engages the buying decision and not the professional duties attaching to it. Fourth consecutive record in this category with this value. Checked the home page, the products pages, the blog index and the site navigation on 29 Aug 2026.

Savings claims only. Published: up to 72 percent of medical record review time saved, review time reduced by up to 50 percent, page count reduced by 90 percent, and a thousand page record reviewed in minutes. Every figure describes the reviewer's own labour and carries an up to qualifier without baseline, sample or period. Nothing appears on the claimant's side of the equation: no position on whether platform cost is a case expense or firm overhead in contingency work, no disclosure guidance, and no record a firm could produce showing what portion of a demand rests on machine extracted findings. The dual market makes the omission notable in a second way, since the same efficiency argument is sold to carriers reviewing the demand packages plaintiff firms send. Checked the self serve page, the SSP product page, the home page and the product pages on 29 Aug 2026.

Savings claims only, including one whose evidence the vendor states it holds and does not show. Published: 50 to 70 percent reductions in record review time, retrieval turnaround cut by an average of 12 days, chronologies in under 24 hours, and a reported 6x return on platform cost within a single quarter carrying the parenthetical that the testimonial is on file. Publishing a figure while noting the supporting evidence exists but is not shown is a distinct move from publishing an unsourced figure and is recorded as such. Nothing appears on the claimant's side: no position on whether platform cost is a case expense or firm overhead in contingency work, no disclosure guidance, and no record showing what portion of a demand was machine generated. Checked the home page, the products pages and the blog material on 29 Aug 2026.

On request, and this record carries both routes to the value rather than one. A published trust page at Trust at DigitalOwl gives a firm a defined destination, and a business associate agreement is offered to customers and covered entities, which is the executable contract HIPAA requires rather than an assurance about it. EvenUp reaches this value through a portal and Tavrn through the contract; DigitalOwl offers both. Named regimes behind them are SOC 2 Type II, HIPAA and GDPR. Held at on request rather than higher because nothing is published open: no subprocessor list, no named model provider, no downloadable report summary, no auditor and no examination dates were located outside the trust page, which was not read in this pass. The Datavant relationship is also undisclosed as a processing matter, which is precisely the question an outside counsel guideline questionnaire would reach.

On request, and this record reaches the value on the strongest instrument in the category. A business associate agreement is offered for HIPAA covered workflows, which is not a claim about compliance but the executable contract HIPAA actually requires between a covered entity and a business associate handling protected health information. A firm can obtain and sign it, which is more than repeating a vendor's assurance. Alongside it a SOC 2 Type II attestation is claimed and specific cryptography is published. Held at on request rather than higher because nothing is open: no subprocessor list, no named model provider, no trust centre, no downloadable summary and no published request route for the SOC 2 report itself. Note the contrast within this category: EvenUp reaches this value through a Trust Center portal, Tavrn reaches it through the contract, and Supio and Eve reach it at all despite naming more regimes, because neither offers a route or an instrument.

Partial record. The structured output is inherently traceable in form, converting unstructured records into chronologies, timelines and summaries that surface provider and billing details and identify the medical evidence behind an assertion, so a party can show which providers and treatments a claim rests on. Published guidance on Bates numbering indicates the vendor understands the citation convention litigation uses, though Bates stamping is not described as a platform feature the way it is on Tavrn. The familiar two limbs are absent: nothing indicates that output records which model produced a given extraction, and no human verification record is captured. The published 98 percent accuracy figure sharpens the second gap rather than closing it, since a party asked in a deposition whether a chronology entry was machine extracted or human verified would find the product captures no evidence either way.

Partial record, and the only record in this category with Bates stamping, which matters more than it might appear. Chronology entries are hyperlinked to specific source pages with navigation back to the document, output is stated to link directly to its source, and Bates stamping assigns every page a stable identifier, which is the convention litigation actually uses to cite a document in a filing, a deposition or a production. A party can therefore point to a page in a way an opposing party or a court can independently locate. The familiar two limbs are missing: nothing indicates that output records which model generated it, and no human verification record is captured. The retrieval specific gap is worth naming as well: nothing in the described output evidences which providers were requested, which responded and which did not, so a party cannot show the completeness of the record set its chronology rests on.