Eve vs EvenUp: how they compare in 2026

The product was built on AI from its founding, and says so in its design. The company was founded in 2020 for this purpose, with no earlier product to attach models to. The January 2026 relaunch as an AI Workforce is built around three agent roles, Agents, Auditor and Analyst. The June 2026 EveOS launch positions it as an operating layer for the firm rather than a tool inside one. Every deliverable is generated: intake evaluation, medical overviews, demand letters, complaints and discovery responses in both directions. The system also adapts to a firm's writing style with use, which is model behavior, not configuration. Remove the models and nothing remains.

The models are the entire product, and the company was built around a proprietary dataset to feed them. Piai is named as EvenUp's own AI engine, and the vendor's central technical claim is that it is trained on a large dataset specific to personal injury rather than adapted from a general model. Every deliverable is model output. Medical chronologies are generated from raw records with treatment timelines and ICD extraction, demand letters are generated end to end, case valuations come from a settlement repository, and analytics sit on top of the extracted data. Remove the models and nothing remains but a document store the firm already had. Independent review material reaches the same conclusion, describing it as a vertical drafting engine trained on injury cases rather than a general legal assistant.

Eve builds grounding and checking into the product, but publishes no measurement. It says all sources for its answers are linked back to the firm's own documents, that the system checks its own responses, and that users can verify a response in one click. Self checking goes beyond showing sources, and the Auditor agent introduced in January 2026 suggests checking is a named part of the system. None of it is measured or specified. There is no accuracy figure, no extraction recall on medical records, no hallucination rate and no evaluation. Nothing says what the self check tests, what happens when it fails, or how the system behaves when records are incomplete or contradictory. A named auditor agent with no published output describes design, not proven accuracy.

Grounding is built into the deliverable, and the vendor publishes no measured figure. Medical chronologies carry citations back to source documents, so a reviewer can check any entry against the underlying record, which suits output that is a factual account of someone's treatment. Demand packages cite comparable verdicts from the settlement repository, so the valuation argument traces to named prior outcomes, and a Thomson Reuters and Westlaw integration supports legal citation. There are two gaps. Two accuracy figures circulate in third party material and neither is published in vendor material with a method attached: that chronologies capture over 90 percent of relevant medical information on first pass, and that demands are 69 percent more likely to reach policy limits. Neither comes with a sample, baseline or definition. And nothing says what the system does when a record is illegible, contradictory or missing, which in medical record work is the common case.

A strong autonomy claim comes with checking built in as a role, but no published limits. The AI Workforce is three coordinated agents that move cases forward without waiting for attorney prompts, a plain autonomy claim, and EveOS extends that to the firm's operations. The Auditor role is a named checking function within the workforce, not a caveat. That is the right design for agents that act on their own. But the limits are not described. Nothing says what an agent may do without attorney approval, whether anything is sent or filed unattended, what triggers Auditor review, or what standard it applies. Nothing says what happens when Agent and Auditor disagree, or how a firm sets the limits. Moving live cases forward on its own is a high stakes claim, and it comes with no published limits.

Oversight is sold as a product tier, which makes it visible rather than aspirational. Demands are offered in tiers from instant AI generation through Express Demands to an expert reviewed tier, where an in house team checks the document before delivery. Independent material notes that the human quality step adds turnaround time, a trade off made visible in the pricing structure. Output is consistently positioned as a draft that requires attorney review. The mechanics are not published: nothing says what the in house reviewers check or against what standard, or what distinguishes the tiers beyond speed. No confidence indication is given on generated content, and nothing describes the instant tier, where no human sits between generation and the attorney's inbox.

Named, dated, checkable evidence backs Eve's claims. Its dated September 2025 funding announcement reports more than 450 firms, more than 350 added in eight months, more than 200,000 cases a year, and firms recovering more than $3.5 billion in settlements and judgments in total. Named customers report results with attributed comment. Frontier Law Center's managing partner, Manny Starr, quoted by name in a named publication in March 2026, reports intake conversion rising from 10 percent to 35 percent, intake time falling by 50 minutes per call, and average case value rising 90 percent. James Scott Farrin attorneys complete medical chronologies in minutes that used to take weeks. A separate verified user reports case capacity rising up to 2.5 times without new staff. A reader can name the firm and partner, find the publication and check the date without contacting Eve. The figures have no method or baseline, and the $3.5 billion is a total of customer outcomes, not proof of Eve's contribution.

Outcome claims are specific and named firms appear, but the best available source flags the whole class as vendor supplied. Lundy Law is reported to have increased output from about 30 to about 110 demand packages a month after adopting AI tooling without adding staff, and J. Chrisp Law to have reclaimed 80 hours of paralegal time per case. Those are named firms with quantified operational change. The vendor claim that demands are 69 percent more likely to reach policy limits is the most consequential figure, because it speaks to case outcome rather than throughput, and it carries no method, comparison group or sample. The independent guide reporting these figures says most such benchmarks come from vendor marketing and that results depend on case complexity and record volume. The adoption story is credible and consistently reported, but none of it is published as a vendor case study with a method.

Eve addresses attorney client privilege directly. It says the platform is built with attorney client privilege as a first principle, that case data is never used to train shared models, and that data is isolated to the firm at organization, user and workflow levels. Naming three levels of isolation is a specification, not just an assurance, and isolation by workflow addresses one matter's content surfacing in another. The credentials fit this market: SOC 2 Type 2 certification, HIPAA compliance for the claimant health information the product takes in, and AES-256 end to end encryption. Two limits: the isolation design is asserted rather than documented, and there is no attestation specific to privilege.

Confidentiality evidence here rests on a health specific credential. The product takes in complete medical records, billing statements and treatment histories for injured claimants, so the confidential material is not only the client's legal matter but a third party's protected health information. Against that sit an independently assessed HIPAA attestation validating safeguards for PHI and a SOC 2 Type 2 examination whose scope includes confidentiality and privacy as well as security and availability. The vendor also says it supports customers handling sensitive information through contractual, technical, organizational and compliance measures, which acknowledges the business associate relationship a plaintiff firm needs. Nothing addresses legal professional privilege or attorney work product. A demand package is work product and the case strategy in a valuation is privileged, but the published posture covers health data protection and general security without reaching the legal side at all.

Not located, and the gap matters because the autonomy claim is large. Agents are stated to advance cases proactively without waiting for attorney prompts, and the product drafts complaints and discovery responses, which are filed documents bearing an attorney's signature and certification. Nothing published addresses the supervising attorney's duty over proactively advanced casework, what an attorney is certifying when signing a machine drafted pleading, the professional responsibility dimension of generated discovery responses, or any bar guidance. The Auditor agent is a quality mechanism inside the product and not a statement about the lawyer's own obligations.

Not located in vendor material. The product generates the demand letter a firm sends to an insurer, which is an advocacy document making legal and valuation arguments on a client's behalf, and it generates the case valuation that shapes settlement advice. Independent material states consistently that output is a draft requiring attorney review, and that framing is not published as a vendor published position. Nothing addresses the supervising attorney's duty over machine drafted advocacy, the professional responsibility of relying on a machine generated valuation when advising a client whether to settle, or any bar guidance. The gap matters more here than on a research tool because the output goes out under the firm's name to an adverse party.

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 or result, no accuracy monitoring, no drift statement, no named governance body and no ISO 42001. The risk is sharp here: this product performs case evaluation and identifies missed case value, so it makes judgments about what a claim is worth, and it does so inside an agent architecture that advances cases proactively. A systematic tendency in valuation would propagate through 200,000 cases annually without an attorney necessarily prompting the judgment that produced it. The Auditor agent checks output within the system and is not an independent evaluation of the models.

Nothing published about how the models are governed or evaluated. No AI policy, no model card, no bias or fairness testing, no evaluation methodology or result, no accuracy monitoring, no drift statement, no named governance body and no ISO 42001. The untested risk on this product is specific and serious: case valuation is generated from a settlement repository of past outcomes, and historical personal injury settlements carry the demographic and geographic patterns of who was compensated well and who was not. A valuation engine trained on that history can reproduce those patterns for a new claimant while presenting the result as a data backed figure, and nothing published indicates whether that has ever been examined.

A data stewardship position is stated, along with a detail Eve openly disclosed without having to. Case data is encrypted with AES-256 end to end, never used to train shared models, and isolated to the firm at organization, user and workflow levels. Eve also says the system learns a firm's tone, language and style and adapts with use. Together these describe adaptation within each firm with no pooling across customers, and stating the second point openly is more candid than staying silent about it. Missing: no retention period for case content or output, no deletion right at matter close, and nothing on whether firm level adaptation can be turned off. Nothing says what happens to an adapted model when a firm leaves. HIPAA compliance covers protection of claimant health data, not how it is used.

Data handling is credentialed and the training question is unanswered. The HIPAA attestation and the SOC 2 Type 2 privacy and confidentiality scope show that safeguards over protected health information have been independently assessed, which is real stewardship evidence. No statement is published on whether medical records, demand drafts, case files or settlement outcomes submitted by firms are used to train or improve Piai, how long client content is kept, or whether a firm can require deletion. The question is central for this vendor: its main technical claim is a proprietary model trained on a large personal injury dataset, and nothing published says where that dataset came from or whether customer matters continue to feed it. A credential covering how data is protected is not a statement about what it is used for.

No published position located. Nothing was found on liability for AI output, warranty, service levels or remedy. The exposure is large because the autonomy is: agents that advance cases proactively can act on a matter without an attorney initiating the step, and the product drafts pleadings and discovery responses subject to filing deadlines and certification requirements. A missed deadline, a defective discovery response or an understated case valuation all carry consequences that fall on the firm and ultimately on the injured claimant, and no published vendor position addresses any of them. Enterprise agreements govern this and are not public.

No published position is published on liability for AI output, warranty, service levels or remedy. The exposure is concrete here. A missed injury, a misread treatment date or an omitted provider in a generated chronology flows straight into a demand letter and can understate a claim, and an understated demand that settles is a loss the claimant never learns about. A generated valuation that anchors a firm low has the same effect. No published vendor position addresses any of it, so the risk sits with the firm and ultimately the injured claimant. Enterprise agreements govern this and are not public.

No integration was found on the pages checked. There was no case management, document management, records retrieval or research integration, and no API documentation. The pages checked were the home page, the working with Eve page, the product material, press coverage of the AI Workforce and EveOS launches, and the site navigation. EveOS is described as an operating platform for the firm, which suggests Eve means to be the operating layer rather than a component that plugs into one. That is an inference, not an answer. This is an absence on the pages checked.

One substantive integration is named, and the connections this buyer needs most are not. A Thomson Reuters and Westlaw integration supports legal citation inside generated demand packages, a real integration for a plaintiff drafting tool that gives the citation layer an established source. No case management system integration is published, and for a personal injury firm that is the connection that matters, since the practice runs on a case management platform holding intake, treatment tracking and deadlines. No API or export documentation is published either, and nothing describes how records get in or how a finished demand returns to the matter file. Independent comparisons in this market treat case management integration as a primary criterion, which makes its absence notable.

Nothing located beyond a generic phrase. Secure cloud infrastructure is stated and names no provider, no region and no residency commitment, and no single tenant or dedicated instance option is described. The firm level isolation the vendor does describe is a logical separation claim rather than a deployment or residency statement, and the two should not be conflated: isolated to your firm says nothing about which jurisdiction the data sits in. For a platform holding claimant protected health information at the scale of 200,000 cases annually, residency is a question a firm's own compliance review asks directly.

Nothing is published. No hosting provider is named, no region or data residency commitment is published, and no single tenant or dedicated instance option is described. For a platform holding protected health information for injured claimants, where that data sits and how it is separated is a question a firm's compliance review would ask directly. The HIPAA attestation shows that safeguards were assessed without saying where the data sits. The Trust Center's request access route was not used and may hold residency detail.

Eve names its certifications and an annual audit cycle, but offers no route to anything behind them. It claims SOC 2 Type 2 certification, HIPAA compliance, AES-256 end to end encryption, secure cloud infrastructure, and policies reviewed annually by independent audits, which speaks to currency. Missing: no auditing firm is named, no examination period, scope or certificate date is published, and there is no trust center, security page or way to request documents. Eve also calls itself the first legal AI to be SOC 2 Type 2 certified. That claim cannot be checked from any published source and says nothing about the current scope or strength of the examination.

Scope, currency and a self serve route are all published. A Trust Center at trust.evenuplaw.com describes security, privacy and compliance practices and offers a request access route to documentation rather than a sales gate. The attestation is named with its scope: a SOC 2 Type 2 examination covering security, availability, confidentiality and privacy, four of the five Trust Services Criteria. Currency is dated, with recertification announced in April 2026. Alongside it sits an independently conducted HIPAA compliance assessment and attestation, the credential this product needs given that it processes protected health information. The vendor correctly explains that HIPAA has no formal certification and that what exists is a third party attestation. The auditing firm is not named for either the SOC 2 or the HIPAA assessment.

Nothing located. No foundation model provider, model family or version is named, nothing states whether third party models sit underneath the agent architecture, and no subprocessor list was found. The commitment that case data is never used to train shared models establishes that a distinction between shared and firm scoped models exists in the architecture without identifying what either is built on. For a platform processing protected health information across 200,000 cases annually, the identity of every processor in the chain is a question a firm's HIPAA business associate diligence asks directly and it cannot be answered from public material.

Nothing located. Piai is named and claimed as proprietary, trained on a personal injury specific dataset, and that is a statement about ownership rather than about supply chain. No foundation model provider is named, nothing states whether third party models sit underneath Piai or process any part of the pipeline, no subprocessor list is published, and the Thomson Reuters and Westlaw integration establishes at least one external data relationship without any accompanying disclosure of what flows to it. For a product handling protected health information, the identity of every party in the processing chain is a question a firm's HIPAA business associate review asks directly and it cannot be answered from public material.

No pricing published at any level. No price, no range, no unit of charge, and no indication of whether the model is per seat, per case, per matter or per firm, which matters because the product is positioned as an operational platform rather than a per document tool and the pricing unit would signal which. Independent material notes that the target segment is small to medium plaintiff firms without any published attorney count, revenue band or case volume threshold defining it, so a firm cannot self assess fit either. The contingency fee point recorded on EvenUp and Supio applies equally: case costs are advanced against a claimant's recovery, so undisclosed pricing is undisclosed cost to an injured person.

No pricing is published at any level. Independent review material describes the model as charged per case rather than per seat, so cost scales with caseload rather than headcount, and notes that the platform pays off only at sufficient injury volume. None of that comes from the vendor: there is no price, range, per case figure or tier structure, and no indication of what a demand package costs. Every route is a sales conversation and an annual contract. The absence weighs more for contingency fee firms, which carry case costs themselves, because a per case charge comes directly out of a claimant's eventual recovery.

The whole case lifecycle is covered and practice limits are stated honestly, but target firms are not defined. Coverage runs from intake to resolution: intake and case evaluation, medical overviews and chronologies, demand letters, complaints, and discovery in both directions, so it reaches into litigation rather than stopping at the demand. Personal injury is the main focus, with medical malpractice, employment and general litigation. Independent review found no sign of family, criminal, intellectual property or transactional work, so the scope claim holds. Missing: the target of small to medium firms has no attorney count, revenue band or case volume, as independent material notes. Nothing describes depth: which record types or formats are handled, which jurisdictions the drafting covers, or whether discovery covers both state and federal practice.

Depth in one practice area, stated plainly, with the boundary acknowledged. The vendor says it builds for plaintiff personal injury and nothing else, and independent review confirms it is not a general legal assistant. Within that scope coverage is deep. Inputs include police reports, medical records, billing statements, photographs and intake notes, and outputs include demand letters, chronologies, case valuations, negotiation preparation and caseload analytics, with a settlement repository supporting comparables by jurisdiction. Specializing in one practice is a design decision rather than a coverage failure. What is not described is the data behind the main differentiator. Nothing says how many verdicts or settlements the repository holds, which jurisdictions it covers, its date range, how often it updates, or whether it is deep enough in a given venue for a comparable to carry weight, which an attorney relying on that citation in negotiation needs to know.