ZLS.app is a sponsor and presenter at the 2026 NAPD Tech Expo, the National Association for Public Defense’s two-day virtual conference on technology in public defense, held September 17 and 18, 2026. Our session runs Friday, September 18 at 11:30 a.m. Eastern, and it is a live demonstration: one felony case, from day-one assignment through trial, showing what happens when AI lives inside the case management system instead of beside it.

NAPD is the largest membership association of public defense professionals in the country, and the Tech Expo exists because the association’s members have been clear that the technology gap is now a constitutional problem and not just an operational one. We are glad to be in that conversation, and we intend to use our time in it showing real software doing real work rather than talking about a roadmap.

What Is the NAPD Tech Expo?

The NAPD Tech Expo is a free, fully virtual, two-day conference for the public defense community, running 8:00 a.m. to 5:00 p.m. EDT on September 17 and 18, 2026. It is open exclusively to NAPD members, and individual membership is $50 per year, so for most attorneys the entire cost of attending is a membership they likely want anyway.

The program is organized around four areas that map closely to what defender offices are actually struggling with: parsing terabytes of body cam footage and mobile device data, using large language models ethically for research, drafting, and investigative leads, protecting attorney-client privilege through sound cybersecurity, and presenting persuasively in remote hearings. Registration is through NAPD.

What ZLS.app Is Demonstrating: One Felony Case, Day One to Trial

Our session follows a single felony file from assignment to trial and shows the case file assembling itself into a trial notebook along the way. Every public defender knows the starting point: 400 pages of discovery, hours of body cam, and a trial date closer than it should be. Here is what we will show handling it.

AI-aided file view. Large discovery dumps get grouped and summarized document by document, so a 400-page production arrives as an organized, readable case file instead of a PDF you scroll.

Advanced transcription tied to real people. Body cam, interview, and jail call audio become time-synced, searchable text, with each voice tied to an actual person record in the case management system rather than “Speaker 2.” Because those are records, anyone in the office can then surface prior videos where the same officer or witness appears.

Witness extraction from police reports. OCR and AI read the reports and add every witness and officer to the case file as a person in the system. Nobody retypes names from a PDF at 9 p.m.

Agentic AI in the case file. The ZLS.ai assistant creates case notes and calendars court dates directly, inside the file, without a second system to update.

Office-wide discovery indexing. Every page and every minute of footage the office has received becomes searchable in context, on a public-defense-specific implementation of current frontier models.

Office knowledge. A curated base of your office’s own motions and policies that ZLS.ai draws on, so the AI works like a member of your team instead of a generic chatbot.

A trial notebook that builds itself. Every statement a given witness made anywhere in the discovery is gathered in one place, so cross preparation starts with the statements already collected.

Why AI Inside the Case File Beats AI Beside It

Because context is the whole advantage. An AI tool you upload 400 pages into knows 400 pages. It does not know the docket, the deadlines, the confidential informant who testified in two other cases down the hall, or the suppression motion your office won on this exact kind of stop last year. That institutional knowledge already exists in a defender office. It has simply never been readable by anything at once.

ZLS.ai works from what the office already has, because the data is a byproduct of the work rather than a separate thing to feed. It also means AI output lands on the case file it belongs to, with a record of who reviewed it and when, which is the part that makes the whole thing defensible when a supervisor or an ethics opinion asks.

We Will Also Be Honest About the Limits

Technology can give defenders leverage. It cannot give them parity by itself, and we are not going to claim otherwise. AI does not reduce a 180 felony caseload to something constitutional, does not hire an investigator, and does not fix a funding formula written decades ago. It does not decide what to file or whether a client’s version holds up. The attorney stays in charge of the case, every output gets verified against its source, and the guardrails that make that possible are part of the product rather than a policy taped over it.

Join Us Friday, September 18 at 11:30 a.m. Eastern

ZLS.app is case management built by public defenders for public defenders, from a team that includes former defenders and the former CIO of the Los Angeles County Public Defender’s Office. If you want to see what AI inside your case management system means for your next client, register for the NAPD Tech Expo and join us Friday morning.

And if there is a workflow you want stress-tested or a question you think a vendor should have to answer live, send it to us before the 18th. We would rather build the session around what your office needs to know.

Sign up here: https://hubs.li/Q04smZls0