Junto Law partner compensation & collections dashboard

01The problem

Junto Law pays attorneys on splits of collected revenue. All the underlying data lives in PracticePanther, but it's scattered across disconnected reports (productivity, origination, billing analysis) that never come together in one place. Every month, Swati manually pulls reports, cross-references invoices against payments, applies split percentages in spreadsheets, backs out expenses, and calculates what each person is owed: hours of careful work PracticePanther forces on her, with no way to show anyone how a number was derived. PracticePanther's own team has already told the firm that the report Junto needs would be a custom build.

The firm simplified its comp model on May 1, 2026 (one firm-wide split), which should have made this easy, but PracticePanther still can't answer the basic question: "Here's what happened this month, and here's what you owe each person."

The partners have a related problem: they can't see their own numbers. They can't tell what they've collected, what they're owed, or which of their invoices are sitting unpaid, so nobody chases the invoices they're personally waiting on.

02What success looks like

• Junto closes a month in one sitting, with zero spreadsheets. She opens the dashboard, reviews each partner's payout statement, resolves anything the system flagged, pays through Melio the way she does today, and marks each payout paid: the ledger closes the loop, all from a single screen of final numbers.
• Every partner logs in and sees their own numbers: what they've collected, what they're projected to be paid, which invoices are outstanding, and exactly how the math worked, without asking Shayn or Swati.
• Shayn sees the whole firm at a glance: collections over time, who's producing, who's originating, and where the money is stuck.
• Overdue invoices get chased by the people incentivized to chase them: the partner who won't get paid until it's collected.
• The numbers are trusted. Every dollar is traceable from payment → matter → split rule → payout. Anything ambiguous is flagged for a human, never silently guessed.
• Trust money is never mistaken for revenue: shown and labeled, excluded from "collected" until applied to an invoice.

Success metrics

03Users & roles

Two ledger distinctions, not new roles: Shayn also appears as a partner line in the comp data (his 10% originations, any work he collects on), and Junto itself is a first-class entity in the ledger (the 40% bucket plus expense recoveries; see §4).

Owners invite new people and change roles. Adding a new attorney to the firm should take minutes, not a development request.

04Compensation rules the system must encode

Standard split

• Working attorney → 50% of what's collected for their work
• Originating attorney → 10% of what's collected on their matter, paid whether or not they did the work
• Junto always keeps 40%, no exceptions; this is only possible if the 10% goes to the originator independently of who worked (see examples)
• Splits are computed on collected money, not billed. Payouts go out the month after collection.

Worked examples: $1,000 of fees collected (confirmed in writing, July 10):

Cash application order:

1. Expenses are reimbursed to the firm first, 100%; they never enter the split pool ("that always goes back to us first")
2. What remains is fee collection, split per the table above
3. A payment that partially covers an invoice allocates the same way: expenses first, then pro-rata across the fee lines (confirmed in writing, July 10)

Junto's bucket

Junto is tracked as its own entity in the ledger, like a partner. It accrues 40% of every fee collection plus 100% of expense recoveries, monthly. Recoveries are reimbursements of money the firm fronted, not earnings: every view labels them "recovered." Shayn's personal lines (50% for work he collects on, 10% for his originations) stay separate from the Junto bucket.

Transition policy

Settled in writing, July 10: all payments received from July 1, 2026 onward split at the current firm-wide rule, regardless of when the underlying work was billed. No legacy-rate engine is needed. Payout statements exist from the July 2026 close forward; earlier months sync for analytics only and are never restated as payout obligations. May and June payouts were made under the old splits, and the tool doesn't second-guess them.

Draw + quarterly true-up (currently Matt Jones)

• A recurring draw (biweekly or monthly; e.g., ~$50k per quarter)
• At quarter end, the system computes what the standard split would have paid: earned more than drawn → bonus for the difference; earned less than drawn → deficit carries forward and nets against the next quarter's true-up
• The carryforward chain is always visible, never a side calculation.

Configurability requirement

Comp is managed on a settings page, not in code:

• Define comp types (Standard Split, Draw + True-Up, and future variants)
• Set percentages, draw amounts, and true-up cadence per type or per person
• Assign each person a comp type with an effective date
• A per-matter split override as an escape hatch for one-off deals
• A firm-level payout effective date, now set: the July 2026 close. All history syncs for analytics, but payout obligations are only computed from that close forward

The bar: onboarding the next "Matt Jones-level" partner, possibly with different numbers, requires zero code changes.

05V1 features

06Explicitly out of scope for v1

• Sending payments: Melio remains the manual last step, by choice; the paid toggle in F3 is the deliberate human verification step
• Replacing PracticePanther data entry: PP remains the system where time, invoices, and payments are recorded (see roadmap, §10)
• Client-facing anything
• QuickBooks and card-statement (Capital One) reconciliation: the v2 profitability work (§10)
• Auto-drafting or auto-sending client collection emails (roadmap, §10; v1 ships the data they'd be built from)
• Editing PracticePanther data from the dashboard

07What we need from Junto

1. API access: done (app created in Shayn's account, July 6)
2. Light data hygiene going forward:
   • An "Originating Attorney" custom field on clients in PracticePanther; matters inherit it (origination follows the client relationship, per §8). We'll pre-fill it from the current "assigned to" convention for Swati to review, and it becomes part of new-client setup.
   • Attach the matter to payments whenever possible when recording them.
3. Swati to validate. One walkthrough of her current close process, then one month-end run in parallel: tool vs. her spreadsheet. Her numbers are the benchmark; payouts cut over only when the tool matches them. With legacy splits retired, the parallel run is pure new-split math on July collections.
4. Swati marks payouts paid in the tool as the last step of each close, after paying through Melio (F3).
5. During onboarding week (non-blocking): decide digest recipients, cadence, and lookback, and who owns collections follow-up. The tool makes all of these configurable.

08Decisions: settled July 10, 2026

Answered in writing

1. The 10% goes to the originating attorney even when someone else does the work. In Junto's words: "that's what keeps Junto's 40% intact." The worked examples in §4 stand exactly as written.
2. Legacy splits are retired. "Old splits are not being considered anymore… All payments received now will be split as per new rule." Every payment received from July 1, 2026 onward splits at the current firm-wide rule, whatever the billing date (§4, Transition policy). This deleted real scope: no legacy-rate engine, no pre-May rate list, no legacy review flag.
3. Origination follows the client relationship, not the matter. A new matter opened for a client another partner brought in keeps the 10% with the client's originating partner.
4. Shayn's own originations get the standard 10%, shown as his own partner line, never folded into the firm's share.
5. Only operating-account money counts as "collected." Trust is shown but excluded until applied to an invoice; trust deposits are client money until then.
6. Payout effective date: the July 2026 close, run in parallel in early August, cutting over once the tool matches Swati's numbers.
7. Partial payments: expenses reimbursed first, then pro-rata across the invoice's fee lines.
8. Expenses: fully excluded from the split pool; the firm recovers 100%, and first.
9. Refunds and write-offs: clawed back from the next payout, surfaced in the review queue.

Agreed on the July 10 call

10. Payout paid status is a manual toggle (F3). Swati marks each payout paid after sending it through Melio: a deliberate human-in-the-loop verification step, chosen over another integration.
11. Expenses-by-partner view (F6): a partner's fronted costs shown next to their aging invoices, so the firm can push collections or require retainers before fronting more.
12. Junto is tracked as its own entity, like a partner: the 40% bucket plus expense recoveries, read as a monthly firm waterfall (F6), always separate from Shayn's personal lines (§4).
13. Downloadable payout verification report per close (F3): every component behind every number, per partner plus firm roll-up. Trust, but verify.
14. Collections follow-up: v1 ships the data, not the automation. Digests and exports carry client name, client email, invoice number, amount, and days past due in a clean, machine-friendly format (F8); auto-drafting and auto-sending are roadmap (§10).
15. Trust balance view (F10), with an explicit accuracy caveat: balances are provisional for the first month or two, reconciled against PracticePanther during the first close or two.
16. CSV export in v1 (F11): every table view; statements and digests also as PDF.
17. Ask your numbers in Claude (F12): Owner/Admin only in v1; answers read from the same reporting views as every other surface.
18. Confirmed as already in scope: invoice paid/unpaid status flips on the hourly sync (F1); the firm-wide outstanding number is a monthly total with month-over-month trend (F6); every payout number drills to its components one click deep ("Eugenia owed $3,000: show how it's calculated and where from").

Still open, deliberately: none of it blocks the build

• Analytics tags and dimensions (client / investor / fund, work type): post-v1; shapes the analytics layer (§10).
• Digest recipients, cadence, lookback, and who owns collections follow-up: Junto's internal call during onboarding week; the tool makes them configurable.
• Invoice-PDF pull into digests: a week-1 API feasibility check.

09Target timeline: dashboards live in 4–5 weeks

Every date here is a target, not a commitment: deliberately aggressive, and confirmed or revised at the end of week 1 once we've seen the live API.

Three honest calendar notes (kickoff July 13):

• Dashboards go live on sprint time (target: weeks 4–5); payout cutover goes live on calendar time. Closes happen at month-end. The goal is to parallel-run the July close in early August; sync (week 1) and the comp engine (week 2) are scheduled to be ready in time. If the live API or the first exception rate pushes that, the fallback is the August close in early September. Either way, cutover happens when the tool and Swati's spreadsheet agree, not before.
• The under-30-minutes close target depends on the exception rate the first parallel run reveals (Shayn's own estimate: payments tie out cleanly "90% of the time"). The first close calibrates it; the review queue is designed to drive it down month over month.
• Trust balances are validated during the parallel-close window and treated as provisional until reconciled against PracticePanther (F10).

10After v1: examples, not a roadmap

There is no committed month-by-month roadmap: priorities reset every sprint, and Shayn's list wins. The theme stays constant: find the tedious, manual, repetitive work in running the firm and hand it to software and AI. Candidates already visible:

• Analytics layer: tag-driven insights: where revenue comes from, which clients and work types pay best, per-client profitability
• Collections tooling: Claude-drafted client reminders built straight from the v1 digest exports (client, email, invoice, amount, days past due), then sent automatically: the workflow Shayn sketched on the call, automated end to end
• Profitability dashboard: QuickBooks and card statements (Capital One) reconciled against PracticePanther expenses, catching spend that never made it into PracticePanther (Delaware Secretary of State fees, for example); per-partner and per-client profitability
• QuickBooks payment matching: auto-verifying the payout paid toggle against money actually leaving the account
• Intake & document automation: engagement letters, new-matter setup, recurring document generation
• AI on the practice itself: first-pass trademark research memos, contract review summaries, drafting support: AI-prepared, attorney-finished
• Progressive PracticePanther replacement: time entry → invoicing → billing inside Junto's own system, retiring the ~$700/month PracticePanther spend. A real project rather than a feature, but the v1 database is designed as Junto's canonical schema (Appendix A), so this path extends the dashboard's foundation instead of restarting.