Case studies/PROTECT/eFiling site

eFiling site: searching every county from a browser

The PROTECT eFiling site lets DA staff look up people and cases across Wisconsin counties from a web browser. Data refreshes every night from each county's PROTECT database. I designed the search, results, error states and detail views, mostly from the existing XCQ component library, which was built from IBM's Carbon kit.

Web toolLibrary reuseData freshness & trust
Client
State of Wisconsin · DA IT program
Role
Sole UI/UX designer
Timeline
2024 – 2025
Team
Me + business analysts + developers
How it works

From nightly sync to an open case

Nightly sync

Each county's PROTECT database syncs overnight to one web-accessible source.

Search

Staff search by name, date of birth or case number across all counties at once.

Scan results

Matches appear in a dense table with county and aliases visible.

Open the record

Person and case details open in the browser. No desktop app needed.

Research

Trust depends on knowing how fresh the data is

I worked with the BA to understand how DA staff cross-reference records outside their own county, and built on what we'd already learned during XCQ. The new question for eFiling was data freshness: the data is a nightly copy, so staff needed to know when it was last updated and whether any county was missing.

BA requirement sessionsXCQ findings reusedEdge-case mapping (stale data, failed syncs)Stakeholder reviews
What we heard
Staff search by surname, and aliases decide whether two records are the same person.
What it changedSurname-first sorting and an alias column in the results table.
What we heard
A missing county looks the same as “no record”, which could lead to a wrong conclusion.
What it changedA partial-sync banner that names the counties that didn't update.
What we heard
Not every staff member has the full desktop app.
What it changedDetail pages that work entirely in the browser.
What I designed

Four parts of the interface

01

Cross-county search

One search across every county, without picking a county first. Each result shows which county it came from.

02

Dense results table

Sortable columns, clear status and enough density for people who work through long lists daily.

03

Error and validation states

Actionable messages for failed county syncs, inline validation, and helpful empty results.

04

Person and case details

Aliases, case list, hearing dates and the source county, all in the browser.

Figma file

Happy path, error states and detail view

The screens are built from XCQ library instances. That's the governance story: a second product shipped with only one new component, the sync banner, which then went back into the shared library. Names are fictional.

PROTECT · eFiling site / Search
100%
# Results / Synced
PROTECT|XCQ SearchPerson LookupCase SearchReportsData synced 2:00 AMKR
All 72 counties synced · today 2:00 AMWhat's included?
Search
Full name ↑DOBSexCountyAliases
Larsen, Erik J03/14/1981MDaneAKA Erik Larssen
Larsen-Holt, Priya A11/02/1990FBrownAKA Priya Holt · AKA P. Larsen
Larson, Dana K06/27/1975FEau Claire—
Lars, Theo M01/09/1968MMarathonAKA Ted Lars +2 more
Lindqvist, Maya R08/30/1996FOutagamieAKA Maya Larsen
# States / Errors
PROTECT|XCQ SearchPerson LookupCase SearchReportsData synced 2:00 AMKR
2 counties didn't sync last night: Iron, Florence.
Results from those counties may be missing. Try again after 6 AM, or contact DA IT if it continues.
Search
No matches for “Larsenn”
Check the spelling, or search by last name only.
# Person detail
PROTECT|XCQ SearchPerson LookupCase SearchReportsData synced 2:00 AMKR
Larsen, Erik J
DOB 03/14/1981 · M
← Back to results
Known aliases
Erik LarssenE. J. Larsen
CountyCase #StatusNext hearing
Dane2024CF001180OpenApr 22
Brown2022CM000418Closed—
Source: county PROTECT databases · synced today 2:00 AM
Interactive fileClick a layer or anything on the canvas to inspect it
Design decisions

Why these choices

Cards or a dense table?

OptionCards with photos and summaries
ChosenA tight table with strong hierarchy

Staff scan long lists. Cards would multiply the scrolling for no gain.

Hide aliases in detail, or show them?

OptionShow aliases only on the detail page
ChosenShow aliases inline in results

Cross-county matching depends on aliases, so they need to be visible while scanning.

Generic error or specific?

Option“Something went wrong. Try again.”
ChosenName the failed counties and the next step

A specific message prevents a missing county being mistaken for “no record found”.

Web or desktop only?

OptionRequire the PROTECT desktop app
ChosenBrowser-based, no installation

Lighter to roll out and available to staff without full desktop access.

Reflection

Good legal search is about order, not simplicity

Putting the right data in the right order mattered more than making the screen look minimal.

  • Density is a feature for legal professionals.
  • Designing for nightly batch data means designing for stale data, failed syncs and partial outages.
  • Being open about data freshness builds trust in government tools.
  • A solid library lets a second product ship quickly and consistently.