When someone asks ChatGPT or Google's AI 'who's the best injury lawyer in Louisville,' does the firm get named? This audit checks that — and tracks it over time. The problem: it's been run three separate times by three sessions that each started over instead of building on the last.

A sweep of what rival Louisville PI firms are doing — their reviews, ads, and positioning — so you can see where SAIL stands. Two versions exist from the same day and were never merged.

A private vault inside Litify so only eight named leaders — you, Joe, Kyle, Nathaniel, Jennifer, Amy, Aliyah, and Janelle — can see sensitive finance, credential, and leadership-only SOP records. Marketing-expense records are locked to that same group. It's already built, deployed, and verified live: a non-leader was confirmed to have zero access.

A redesigned 'contact us' form for the sidebar of the firm's website. The new version is finished and sitting on a branch — it just hasn't gone live yet.

A small automation that QA-checks any website form before it goes live, so a broken form never ships again. It's written and working — it's just not registered in the skills index yet, so agents can't find it.

A simple form for logging what the firm spends on each marketing channel, so spend lives next to the matters it generates. The spec is finalized; the form itself isn't built yet.

A tab inside each Litify matter that shows the RingCentral call and text history for that client — calls right where the case lives. The component is built; it just needs to be dropped onto the page layout.

An audit found 1,422 loose reports in Litify — 310 of them stale and 50 with duplicate names. It's clutter that makes the real reports hard to find. The cleanup was scoped but never run.

Business accounts (insurers, employers, vendors) are entered inconsistently — same company, many spellings. This normalizes them so reporting and dedup actually work. A follow-on to a dedup project that already solved the same problem for addresses.

Three small data-cleanup scripts from the P05 project are broken (a bad query, wrong field names, and a parsing error). On their own they're minor — but they'll stop the next P05 run cold until they're fixed.

There are 11 old behind-the-scenes automations that reformat phone numbers (an earlier handoff thought there were only 5). Too many overlapping rules slows saves and invites bugs.