EvenUp

EvenUp is an AI platform built exclusively for plaintiff personal injury law firms. Founded in late 2019 by Rami Karabibar (CEO), Raymond Mieszaniec (COO), and Saam Mashhad (CPO, a former litigation attorney), the company trains its proprietary AI model, Piai, on hundreds of thousands of PI cases, millions of medical records, and a structured database of 250,000+ jury verdicts and private settlements. The result is a system that can produce a full settlement demand package in a fraction of the time it takes a paralegal working from scratch. Over 2,000 personal injury firms use EvenUp today, including 20% of the Top 100 U.S. personal injury firms, and the platform has helped resolve more than 200,000 cases and recover over $10 billion in damages for injury victims.

EvenUp's flagship product, Demands, generates comprehensive demand packages with six auto-populated sections: facts and liability narrative, medical chronology with ICD codes, past and future damages calculations, pain and suffering analysis, verdict comparisons from the PI database, and court-ready exhibits extracted from raw medical records. Beyond demand drafting, the platform has expanded into Medical Management (real-time treatment tracking to prevent case-value erosion from treatment gaps), Communication Agents (AI voice and outreach tools for client follow-up), AI Drafts Suite (complaints, discovery responses, and motion drafting), Smart Workflows (CRM-integrated automation for case intake and routing), and Executive Analytics (firm-wide performance dashboards). The platform integrates with major case management systems including SmartAdvocate, Litify, CasePeer, and Clio, pulling case data automatically to reduce manual uploads.

What EvenUp actually is in May 2026

EvenUp is not a chatbot, a general legal research tool, or a contract assistant. It is a vertical AI platform for one specific practice area: plaintiff personal injury law in the United States. Every piece of training data, every feature, and every workflow assumption is built around the PI case lifecycle, from initial client intake through pre-litigation demand submission to settlement negotiation and, in newer products, trial preparation.

The underlying model, Piai (Personal Injury AI), is trained on data that general-purpose legal AI tools simply do not have access to: structured records from hundreds of thousands of PI cases, verdict and settlement data spanning the full range of injury types and jurisdictions, and clinical records covering the treatment trajectories that insurance adjusters scrutinize. EvenUp supplements the model with a team of over 100 nurses, paralegals, adjusters, case managers, and attorneys who review AI output before it reaches the firm. The company calls this "AI + Professional Review" and claims 99% accuracy as a result, though former employees have publicly disputed how much of that accuracy depends on human correction rather than the AI itself (more on that below).

In October 2024, EvenUp raised a $135 million Series D round led by Bain Capital Ventures, with participation from Premji Invest, Lightspeed Venture Partners, Bessemer Venture Partners, SignalFire, and B Capital Group. The round pushed the company's valuation past $1 billion and its total capital raised to $235 million. CEO Rami Karabibar said at the time: "We empower personal injury firms to deliver higher standards of representation, with the goal of ultimately helping the 20 million injury victims in the U.S." The company followed with a $150 million Series E in October 2025 at a $2 billion+ valuation, led by Bessemer Venture Partners.

Where EvenUp sits versus Eve Legal and Harvey AI

The two most relevant comparisons are Eve Legal on the plaintiff-AI side and Harvey AI on the enterprise-legal-AI side. They illustrate EvenUp's positioning better than any self-description.

Eve Legal is EvenUp's most direct competitor. Eve (see our Eve Legal listing) raised $47 million in a Series A led by Andreessen Horowitz in January 2025 and reached a $1 billion valuation by the end of 2025. Like EvenUp, Eve targets plaintiff personal injury firms exclusively. The architectural difference is significant: Eve is designed as a primary AI-native case workspace, meaning firms run their entire case operation inside Eve's interface rather than connecting a specialized tool to their existing CMS. Eve's "Auditor" feature scans all active caseloads nightly to surface missed opportunities, such as potential traumatic brain injury patterns or mass tort eligibility flags, and produces medical chronologies in roughly 30 minutes and demand letters in seconds. EvenUp, by contrast, is built as a specialized layer that integrates with existing tools like Litify, SmartAdvocate, and CasePeer rather than replacing them. Whether a firm wants a CMS-integrated overlay (EvenUp) or a full workspace replacement (Eve) is the central fork in the road when evaluating the two.

Harvey AI (see our Harvey AI listing) operates in an entirely different market segment. Harvey is a general-purpose legal AI used by more than 200 law firms including 20%+ of the AmLaw 100. It is built on a customized large language model fine-tuned on legal text broadly, and it handles M&A due diligence, regulatory compliance, litigation strategy, and contract analysis across practice areas. Harvey's custom case law model achieves 94.8% accuracy on document Q&A tasks and a 0.2% hallucination rate on citation tasks. Harvey costs $1,000+ per lawyer per month with a 20-seat minimum. It has no PI-specific training data, no verdict comparison database, and no pre-built medical chronology tooling. Harvey is where BigLaw firms go for cross-practice AI; EvenUp is where plaintiff PI firms go for case-specific automation. Firms choosing between them are not typically choosing between equivalent products. They serve fundamentally different workflows.

For completeness, Spellbook and Robin AI are frequently mentioned in the "legal AI" category but focus on transactional contract review and drafting inside Microsoft Word. Neither has PI-specific capabilities, and neither competes with EvenUp in practice. Firms seeking help with written legal output more broadly, outside the PI context, may also look at CustomWritings for document generation needs.

How AI actually works inside EvenUp

Piai processes incoming case files through four stages. Entity Extraction parses raw documents, including handwritten notes, scanned images, and digital records, into structured data. Data Cleanse reconciles discrepancies across providers, eliminates duplication, and categorizes information by relevance to the demand. Output Generation produces firm-specific language with line-level citations, so attorneys can click through to the source document for any claim in the demand. Quality Control runs human reviewers from EvenUp's internal team of medical and legal specialists across every deliverable.

The Demands product offers two service tiers: "AI + You," where the firm finalizes the draft itself after EvenUp produces the skeleton, and "AI + Professional Review," where EvenUp's internal reviewers work the draft before return, with a 1-5 business day turnaround. The platform claims demands produced with its tools are 69% more likely to achieve policy limit settlements than non-EvenUp demands, based on analysis of its case outcome dataset.

"We went from an initial $50,000 offer to a $1.75 million settlement in mediation by using EvenUp's Case Companion to dismantle the defense's arguments in real time. Their pace of innovation is unmatched: every month, they ship something new that meaningfully improves our workflow." - Kyle Wright, Managing Attorney, Wisehart Wright (Ohio), EvenUp case study, 2025

"We scaled past $500M in annual results with 70% year-over-year growth, all without increasing headcount. EvenUp has been a key partner in that journey." - Steve Mehr, Founding Partner, Sweet James (California), EvenUp case study, 2025

The Case Companion feature, introduced alongside the demand suite, allows attorneys to query thousands of pages of case records in real time, surface relevant precedents, and prepare for mediation or depositions without manually searching through file stacks. It is the feature attorneys most frequently cite in testimonials.

The reliability and transparency concerns users keep raising

In December 2024, Business Insider published an investigation based on interviews with multiple former EvenUp employees. The report documented a significant gap between EvenUp's marketing claims and what employees said happened in practice. Specifically: the AI regularly missed injuries in medical records, fabricated medical conditions that were not present in source documents, and misrecorded doctor visit details. Supervisors, according to former staff, at times instructed workers to skip the AI entirely and complete tasks manually. Employees described working until 3 a.m. on tasks that had been presented to clients as automated, and noted that inadequate human oversight at any point could have resulted in clients receiving reduced settlement payouts.

EvenUp's public response emphasized its hybrid model: human review is a design choice, not a deficiency, and the claimed 99% accuracy is the result of combining AI drafting with professional editorial oversight. The company also noted that AI generates 72% of demand letter content, reducing drafting time by 20%. The incident was logged in the AIAAIC (AI, Algorithmic, and Automation Incident and Controversy) repository under "Legal tech company EvenUp accused of systemic misleading marketing."

Beyond the Business Insider episode, users and former employees on Glassdoor describe additional friction. Pricing opacity is a consistent complaint: firms cannot get a rate sheet without engaging in a sales process, and per-case or per-demand costs with add-ons have surprised some buyers. Output format rigidity is another concern. EvenUp demand letters have a recognizable structure, and some attorneys report that insurance adjusters discount them as templated, reducing negotiating leverage. Turnaround time under the professional review tier can slip beyond the promised 1-5 days during high-volume periods. And because EvenUp integrates with rather than replaces a firm's CMS, it is always an additional tool purchase, which strains firms already paying for case management software.

The honest picture is that EvenUp is a powerful tool that meaningfully reduces the paralegal hours required to produce a demand package, and its outcome data is compelling. The gap is between what the AI autonomously achieves and what the final product looks like after human correction. Firms that use EvenUp as a first-draft accelerator, not a hands-off autopilot, consistently report the best results.

Who EvenUp is for

EvenUp performs best for mid-size to large plaintiff personal injury law firms running 50+ active cases monthly who want to cut paralegal hours on demand preparation, access verdict comparison data for settlement positioning, and standardize output quality across junior and senior staff. Firms that have already adopted a compatible CMS (Litify, SmartAdvocate, CasePeer, or Clio) get the most out of the integration layer.

EvenUp is not the right fit for solo practitioners who cannot absorb custom enterprise pricing, for defense firms (the platform is plaintiff-side only by design), for general litigation or transactional practices, or for law firms outside U.S. personal injury practice. It is also not a CMS replacement: firms that need comprehensive practice management software alongside AI demand generation will pay for two platforms, not one.

At the 4.3 score, EvenUp is a genuinely strong tool for its specific vertical. The $10 billion in resolved claims, 2,000+ firm client base, and outcome data on policy limit settlement rates are real. The December 2024 controversy is real too, and it introduces a legitimate question about how much attorney oversight the tool actually requires in practice. For firms willing to treat it as an accelerator rather than a replacement for attorney review, EvenUp delivers measurable results. For firms expecting a fully autonomous system, the reality is more complicated.