When Lone Star Legal Aid launched its Artificial Intelligence Initiative, the goal was ambitious: explore how AI could help legal aid attorneys work more efficiently, improve internal operations, and make reliable legal information easier for the public to find.
Throughout spring and summer 2026, that vision moved closer to reality.
LSLA continued advancing three distinct AI tools—Juris, LSLAsks, and NAVI—while building the infrastructure, partnerships, and safeguards needed to move them from development into sustainable use.
At the center of the initiative are three chatbots designed for very different users and needs.
Juris is an internal legal research assistant designed to support LSLA attorneys in finding and working with legal information more efficiently.
LSLAsks is an internal administrative assistant designed to help staff quickly find information related to organizational policies, procedures, HR, IT, and other operational needs.
NAVI is the initiative’s public-facing component. Designed for people seeking civil legal help, NAVI aims to make it easier for users to identify their legal issue, understand legal information in plain language, find trusted resources, and connect with appropriate legal aid and community services.
During the spring, Juris and LSLAsks moved from development into practice, launching internally and becoming part of LSLA’s day-to-day workflows. Their launch, however, does not mark the end of their development. Like all three tools, Juris and LSLAsks are being designed around continuous iteration—using feedback from the people who actually use them, real-world testing, and input from subject-matter experts to identify opportunities for improvement.
Built-in feedback tools within the interfaces allow users to share their experiences as they work with the chatbots, helping the project team evaluate performance and refine the tools over time. This approach recognizes that responsible AI deployment is not a one-time event, but an ongoing process of testing, learning, evaluation, and improvement.
Meanwhile, work continues on the more complex task of preparing NAVI for public use.
NAVI represents more than adding a chatbot to a legal aid website.
The goal is to create a tool that can bring together trusted legal information from multiple organizations and help Texans navigate it in a way that makes sense for their individual circumstances. Generative AI can help translate complicated legal terminology into more understandable language, while geographic routing can help users locate information and resources relevant to their county and community.
That work has increasingly become collaborative. LSLA has worked with Texas Legal Services Center, law libraries, immigration-focused organizations, and other access-to-justice partners to identify and organize reliable information that can support the platform.
The underlying principle remains important: AI should help people navigate trusted legal information—not replace attorneys, legal judgment, or the expertise of the organizations creating that information.
One of the most significant developments this spring happened largely behind the scenes.
As LSLA prepared for long-term deployment, the project team decided to pursue on-premises AI infrastructure rather than relying exclusively on cloud-hosted AI services.
The approach is intended to give LSLA greater control over its technology and data, strengthen privacy protections, reduce dependence on consumption-based AI costs, and create an infrastructure that can support the organization’s AI work well beyond the initial grant period.
That decision reflects a lesson that has become increasingly clear as the project has evolved: building an AI tool is only part of the challenge. Legal aid organizations also must consider security, governance, cost, evaluation, staff training, communications, maintenance, and long-term sustainability.
The experience with Juris and LSLAsks is also reinforcing another lesson: deployment is the beginning of a new phase, not the end of development. Maintaining useful AI tools requires organizations to create processes for gathering feedback, evaluating performance, incorporating expert review, updating information, and responding to what they learn from real-world use.
As the technology has progressed, interest in LSLA’s work has grown beyond the organization.
Throughout the spring and summer, LSLA has shared lessons from the project with legal aid organizations, access-to-justice partners, researchers, and others exploring responsible uses of artificial intelligence.
Data Analytics Director Ashley Oborn has participated in national conversations surrounding AI and access to justice, including Legal Services Corporation’s AI Peer Learning Labs. LSLA’s experience has also drawn interest from researchers examining how legal aid organizations can deploy AI responsibly.
That interest continued into the summer. LawDroid invited Oborn to serve as the inaugural guest for its A2J Conversations webinar series, where she discussed LSLA’s approach to developing AI tools with attorney-vetted content, safeguards, and access-to-justice outcomes at the center of the work.
LSLA has also proposed sharing the lessons learned from operating all three tools at the next Innovations in Technology Conference through a session titled “From Pilot to Sustainable Practice: Operating Three AI Chatbots in Legal Aid.”
The proposed presentation reflects how much the conversation has changed. The question is no longer simply, Can legal aid build an AI chatbot? Increasingly, the questions are: How do we operate it responsibly? How do we know whether it works? How do we improve it based on what users experience? How do we earn staff and public trust? And how do we sustain it?
The project has also contributed to growing recognition of LSLA’s broader data and technology work.
This summer, LSLA was named a finalist for the 2026 Top 100 Data & Analytics Team Award, recognizing innovation, leadership, and impact in the data and analytics profession. The chatbot initiative has also been considered for recognition through the American Legal Technology Awards.
Perhaps just as meaningful is the growing interest from people who want to learn from the work. Students, researchers, legal technologists, and access-to-justice professionals have reached out to LSLA specifically about Juris, LSLAsks, and NAVI.
There is still work ahead—particularly as NAVI moves toward public use. At the same time, LSLA is learning from Juris and LSLAsks now that both tools are operating internally and becoming part of staff workflows.
Going forward, continuous iteration will be a central part of LSLA’s approach to all three tools. User feedback collected through built-in features, observations from real-world use, performance evaluation, and input from attorneys, subject-matter experts, and other stakeholders will help inform ongoing refinements. As the technology, available information, and users’ needs change, the tools will need to evolve with them.
That approach reflects an important transition for LSLA’s AI initiative. What began as an effort to explore three potential uses of artificial intelligence is becoming a broader strategy for integrating and maintaining AI responsibly within legal aid.
The next phase will continue to focus on deployment, testing, evaluation, security, staff adoption, partnerships, maintenance, and sustainability, while keeping the project’s original purpose at the center of every decision: using technology to strengthen access to justice.
As Juris, LSLAsks, and NAVI continue to evolve, we’ll continue sharing what we’re learning along the way.