AI Audit report · October 2026 · Demo
Where AI pays off at Halden & Rocha Advogados
We mapped the firm's matter lifecycle from first inquiry to archive, interviewed people in every area, and ranked each problem by impact. Each finding comes with a fix: an existing tool to adopt, or a project we can build.
Recoverable time
≈ 375 h
Per month, across all proposed solutions, at current volumes.
Findings
12
8 pain points and 4 opportunities across 11 workflow phases.
Highest risk
1 person
All court deadlines depend on one paralegal reading notices by hand (P08).
Engagement scope
- Client
- Halden & Rocha Advogados (fictional)
- Practice
- Civil litigation and corporate advisory
- Team
- 6 partners, 22 associates, 9 paralegals, 5 staff
- Volume
- ≈ 40 new matters / month · 620 active
- Interviews
- 14 people across intake, legal, finance and admin
- Duration
- 2 weeks on site and remote
How to read this report
- 01 · Workflow mapHow work moves through the firm today, phase by phase, with owners, tools and volumes.
- 02 · Pain pointsYellow notes mark problems, blue notes mark opportunities, each pinned to the phase where it happens.
- 03 · SolutionsGreen notes sit on each finding with the proposed fix, effort and expected savings.
Now
0–30 days · adopt existing tools · ≈ 115 h / mo- S04 → P04Client portal with automatic reminders20 hExisting tool · 1 wk setup
- S06 → P06Drafting copilot with house templates70 hTool + build · 2 wks pilot
- S07 → P07Document management with versioning15 hExisting tool · 3 wks rollout
- S11 → P11Collections reminder schedule10 hExisting tool · 1 wk setup
Next
Months 1–4 · sprint builds · ≈ 215 h / mo- S01 → P01Unified intake with AI triage25 hTool + build · 2–3 wks
- S03 → P03Engagement letter automation30 hhenrilab build · 2–3 wks
- S05 → P05Firm knowledge base (RAG)60 hhenrilab build · 5–6 wks
- S08 → P08Court notice reader and deadline engine25 hhenrilab build · 4 wks sprint
- S09 → P09Client status assistant35 hhenrilab build · 4–5 wks
- S10 → P10Passive time capture40 hTool + build · 3 wks
Later
Months 4–6 · build on the base · ≈ 45 h / mo- S02 → P02Automated conflict search30 hhenrilab build · 3–4 wks
- S12 → P12Archive-to-knowledge pipeline15 hhenrilab build · 2 wks, after S05
Recommended next step
Start with a 4-week AI Automation Sprint on deadline automation (S08). It removes the firm's largest operational risk and creates the document pipeline that the client status assistant and knowledge base will reuse.