The Access to Justice Technology Gap: Measuring What Works
The United States faces a civil legal aid crisis that technology has promised to solve for over a decade. Approximately 92% of the civil legal needs of low-income Americans go unmet, according to the Legal Services Corporation's 2022 Justice Gap Report — a figure that...
The United States faces a civil legal aid crisis that technology has promised to solve for over a decade. Approximately 92% of the civil legal needs of low-income Americans go unmet, according to the Legal Services Corporation's 2022 Justice Gap Report — a figure that has remained stubbornly resistant to intervention despite sustained investment in legal technology. The question confronting funders, policymakers, and builders in 2024 is no longer whether technology can help close this gap, but which specific tools have earned their place in the ecosystem through demonstrated impact.
The Measurement Problem Comes First
Before evaluating what works, it's worth confronting the foundational obstacle: the access-to-justice technology space suffers from a profound measurement deficit. Most tools are evaluated on outputs — number of users, documents generated, sessions completed — rather than outcomes. Did the user successfully navigate their eviction proceeding? Was the protective order granted? Did the child support modification resolve favorably?
This gap is not accidental. Outcome measurement requires longitudinal tracking, court data integration, and follow-up infrastructure that most legal aid organizations and technology vendors lack the capacity to build. The result is an ecosystem where tools get funded for reach rather than resolution, creating incentives misaligned with actual access to justice.
What the Evidence Actually Shows
Document Assembly: The Strongest Signal
Among the technology interventions with the most credible evidence base, document assembly tools — which guide users through structured interviews to produce completed legal forms — have demonstrated consistent value. A2J Author, developed through a partnership between Chicago-Kent College of Law and the Center for Computer-Assisted Legal Instruction, has been deployed across hundreds of court self-help programs and documented meaningful reductions in form errors compared to unassisted filers.
The Michigan Legal Help program, which uses document assembly alongside plain-language guides, offers one of the more rigorous cases. A 2019 study conducted in partnership with the University of Michigan found that users who completed document assembly interviews had higher rates of successful court filings than those who attempted forms without assistance. Critically, the program also reported reduced burden on court clerks, suggesting efficiency gains at the institutional level rather than merely shifting complexity around.
Documate (now Gavel) and HotDocs similarly report significant deployment in legal aid contexts, though independent third-party evaluation of outcomes remains limited. The pattern holds: document assembly has the strongest evidence base in this sector, even if that evidence is incomplete by rigorous standards.
Chatbots and AI Tools: Promising Signal, Thin Evidence
The chatbot space presents a more complicated picture, particularly as the arrival of large language models prompted an explosion of AI-assisted legal guidance tools beginning in 2022 and 2023. DoNotPay became the most visible — and most cautionary — example. Its aggressive marketing claims about replacing lawyers were contradicted by documented failures, culminating in a 2024 FTC consent order (finalized in February 2025, with $193,000 in monetary relief) and multiple state bar investigations. The episode damaged credibility for the broader sector at a moment when legitimate tools were beginning to establish track records.
More credible work is happening at the nonprofit and court-embedded level. LawHelp Interactive, operated by Pro Bono Net, powers guided assistance tools across multiple states and has processed millions of guided interviews. The ILAO (Illinois Legal Aid Online) website serves as a comprehensive model, combining document assembly, legal information, and referral routing. ILAO reported over 5.7 million legal questions answered in 2022, though the organization is candid that measuring downstream outcomes remains a challenge.
RelativityOne and similar platforms have integrated AI tools aimed at legal professionals rather than end users, demonstrating efficiency gains in document review — but those gains accrue to law firms and legal departments, not to the self-represented litigant navigating a housing court.
The honest assessment of AI chatbots for direct client use is this: they are very good at delivering information and very uncertain in their ability to provide reliable guidance across the variability of state and local legal rules. The jurisdictional specificity required for legal advice — knowing that Illinois has a 10-day answer deadline in eviction cases, not the federal standard — remains a persistent failure mode for general-purpose large language models. Tools built with jurisdiction-specific training data and human review workflows, like those being piloted by several state court systems, show more promise than general deployments.
Court-Integrated Self-Help: The Underrated Model
Court-integrated self-help centers deserve greater attention than they typically receive in legal technology discussions. The Judicial Council of California's self-help center network, serving hundreds of thousands of litigants annually, combines in-person facilitation with digital tools in a model that addresses a key limitation of purely online approaches: many low-income users lack reliable broadband access, digital literacy, or trust in online systems.
The National Center for State Courts has documented that courts with integrated self-help programs show measurably lower rates of case dismissal due to procedural error among self-represented litigants. This is not a glamorous technology intervention — it often involves a laptop, a trained facilitator, and a well-designed document assembly tool — but the evidence of impact is more reliable than most venture-backed alternatives.
Funding Sustainability: The Structural Failure
The most serious challenge facing access-to-justice technology is not technical — it is financial. The dominant funding model relies on foundation grants, LSC appropriations, and IOLTA (Interest on Lawyers' Trust Accounts) revenue that fluctuates with interest rates. The IOLTA crash following the 2008 financial crisis wiped out funding for dozens of programs, and the inverse occurred when rising rates boosted IOLTA revenue after 2022.
This volatility is incompatible with the infrastructure investment that technology tools require. A document assembly platform needs ongoing maintenance as forms change, laws update, and platforms evolve. The 2020 sunsetting of Rocket Lawyer's nonprofit partnerships illustrated how quickly commercially backed tools can withdraw when unit economics shift.
State justice commissions have begun exploring court fee revenue as a more stable funding mechanism for self-help technology, following models piloted in Washington State and Nevada. The Access to Justice Commission in Washington successfully advocated for a portion of electronic filing fees to fund self-help resources — a structural approach that reduces grant dependency. This model deserves replication.
Recommendations for Funders and Builders
For funders: Require outcome metrics as a condition of continued grant support, but fund the measurement infrastructure necessary to generate them. The Legal Services Corporation's Technology Initiative Grants program has moved in this direction, but the field needs standardized outcome frameworks that allow cross-program comparison.
Prioritize multi-year operating support over project grants. A document assembly tool is worthless if the organization cannot afford maintenance two years after launch. The Heising-Simons Foundation and Schmidt Futures have provided multi-year technology grants in the legal aid space — this model should become standard practice rather than the exception.
For builders: Stop building new tools and start integrating with existing infrastructure. The field has too many competing platforms and too little interoperability. Investment in APIs that connect court case management systems with legal aid intake — a capability that Tyler Technologies is positioned to enable through its widespread court system deployments — would have greater impact than another consumer-facing chatbot.
Prioritize jurisdictional specificity over scale claims. A tool that reliably serves users in three counties is more valuable than one that vaguely claims national coverage but fails unpredictably. Legal problems are intensely local, and technology that respects that reality will outperform technology that ignores it.
The access-to-justice technology gap is real, but it is not primarily a technical problem. It is a measurement problem, a funding problem, and a prioritization problem. The tools to close it largely exist. The evidence to know which ones work is still being built — and building that evidence infrastructure may be the highest-leverage investment available to anyone serious about change.