These guidelines help you get the most accurate matches, use the AI effectively, and maintain a clean workflow.

BlockNote image

Writing effective search queries

  1. Be specific with specialty terms — Use the most precise subspecialty you can. "Orthopedic spine surgery" will return better matches than just "orthopedics." Check Salesforce for common subspecialty naming conventions.

  2. Always include location — The 75-mile rule is central to matching. Provide a city/state or ZIP code as close to the actual venue as possible — not the client's home address.

  3. Include side preference when it matters — If the case specifically needs a plaintiff-side or defense-side expert, include this in your search. It helps the scoring algorithm prioritize the right candidates.

  4. Use requirements text for nuance — The free-text requirements field is processed semantically. Add details like "must have published research on spinal fusion outcomes" or "experience testifying in federal court" — the AI uses this for semantic matching.

Using the AI Chat effectively

✅ Do

  • Start broad, then narrow — Begin with a general search, then ask follow-up questions to refine: "Now filter those to defense-side only"

  • Ask for comparisons — "Compare the top 3 experts from my last search"

  • Use context from prior messages — The AI remembers the conversation: "Check that first expert for conflicts with Johnson & Johnson"

  • Request specific information — "What's Dr. Smith's retainer fee?" or "Does she have a CV on file?"

  • Switch models for different tasks — Use Fast models for lookups, Recommended for searches, Research for background questions

⛔ Don't

  • Don't ask the AI to modify data — It can only read, not write. Use the UI for actions like selecting experts or closing requests.

  • Don't paste very long documents — The context window is large but not unlimited. Summarize lengthy case details before pasting.

  • Don't expect real-time information from standard models — Only Perplexity Research models have web access.

  • Don't chain too many actions in one message — Break complex workflows into separate messages: search first, then conflict check, then compare.

Request intake tips

  • Complete all fields — The more information you provide upfront, the better the AI matching. Location and specialty are critical — everything else improves accuracy.

  • Use consistent naming — When entering plaintiff and defendant names, use the exact legal names. This improves conflict check accuracy.

  • Add context in notes — If there's a deadline, specific judge, or unusual requirement, put it in the notes. The AI summary will pick it up and the case manager will see it immediately.

  • Set priority correctly — Rush priority triggers faster attention from case managers. Use it for genuine urgency only.

Managing your workflow

Inbox hygiene

  • Update statuses promptly — Move requests from New to In Progress when you start working on them. This prevents duplicate effort.

  • Close completed requests — Don't leave matched requests open. Close them after expert selection to keep the Inbox clean.

  • Use filters daily — Filter by In Progress to see your active workload; filter by New to pick up unassigned requests.

Conflict checking

  • Always check before selecting — Run a conflict check for every expert you're considering, even if they seem like a clear match.

  • Check all parties — Include plaintiff, defendant, and the involved law firm. Missing a party is the most common cause of missed conflicts.

  • Verify edge cases manually — Name-based matching may miss aliases, subsidiaries, or individuals with common names. When in doubt, verify with the expert directly.

Expert selection

  • Review the full profile — Don't select based on the summary card alone. Open the slide panel and check credentials, fees, and testimony history.

  • Document your reasoning — Add notes to the request explaining why you selected a specific expert. This creates a paper trail and helps with future similar cases.

Data quality

You directly impact match quality by keeping Salesforce data clean:

  • Ensure expert addresses are complete — Missing addresses mean excluded experts

  • Use consistent subspecialty terms — "Cardiology" vs. "Cardiology - Interventional" matters

  • Update the Can Travel field — If an expert says they'll travel to new states, update Salesforce so the system picks it up

  • Flag quality/margin tags — These directly affect scoring and ranking