Software consulting · Contracting · Law firms

The systems your firm runs on, kept running and moved forward.

RNDBI is a software consulting and contracting group for law firms. Data solutions, legal software support, migrations, customizations, integrations and instrumentation, process automation: the practice, accounting and document systems, and the Microsoft platform underneath them.

Start a conversation See the work

THE SYSTEMS WE SUPPORT, CONNECT AND MODERNIZE AT ONE FIRM

What we do

Seven lines of work. Most engagements cross more than one, because a firm's systems do.

Data solutions

Get the numbers right, then get them out

SQL Server design and repair, reporting with SSRS and Power BI, a Fabric fact layer when the firm is ready for one. Reconciliation of accounting, time and billing data that has drifted apart.

Legal software support

Keep the practice systems running

Administration, troubleshooting and hardening of practice management, accounting, case management and document management systems. Vendor tickets escalated with the evidence attached.

Migrations

Move without losing the ledger

Platform, version, tenant and server migrations, planned around the firm's calendar, rehearsed before cut-over, and reconciled afterwards on the authoritative keys.

Customizations

Make the system fit the firm

Forms, workflows, reports, stored procedures and extensions built inside the products the firm already owns, documented so the next administrator can read them.

Integrations & instrumentation

Connect the systems, and watch them

Integrations across practice, accounting, document and identity systems; sync monitoring, reconciliation checks and alerts, so a failed job is known the hour it fails rather than at month end.

Process automation

Take the repetition off people

Power Platform, PowerShell and scheduled jobs for intake, provisioning, reporting and reconciliation. Every automation proposes and logs; anything that reaches a client record is held for a person.

Digital assistants

Assistants that show their work

Assistants that work inside the firm's own systems, answer in plain language, and show the query behind every number. Each one is scoped to the data its job needs and writes nothing without a person's approval, so the routine follow-up gets done without a partner doing the following up.

Show, not tell

Small pieces of the work, built the way we build them for firms. Every figure below is fictional and says so.

01 · Prebill

A prebill that posts itself

Time entries post, hours and fees accumulate, disbursements and trust apply, and the balance due is the arithmetic, not a typed number. The stamp appears only when the prebill agrees with the ledger.

The layout is a prebill review sheet of the kind a law firm circulates before invoices go out. The same data drives the report in the next sheet.

Fictional firm, fictional matter

Doe & Roe LLP

PREBILL · DRAFT 3
PERIOD 2026-08-01 TO 2026-08-31

ClientBlackacre Holdings

MatterMTR-0417 · Lease dispute

Billing attorneyJ. H. Roe

Date	TK	Description	Hours	Rate	Amount
08-03	JHM	Review lease amendment and correspondence from landlord's counsel	1.8	450	810.00
08-04	ATR	Assemble exhibit binder; confirm service list	0.6	285	171.00
08-07	JHM	Draft response to notice of default; call with client	2.4	450	1,080.00
08-12	ATR	Research cure-period provisions; memo to JHM	3.1	285	883.50
08-19	JHM	Settlement conference preparation	0.9	450	405.00
08-26	JHM	Attend settlement conference; report to client	1.5	450	675.00

Hours0.0 Fees$0.00 Disbursements (filing fee, courier)$0.00 Trust on account (to apply at invoice)$0.00 Balance due$0.00

Agrees with ledger Example · fictional data

02 · Report

A report that answers back, and shows the query

Ask it in plain language -- which matters slipped, where realization moved, what changed since last month. The question becomes a query against the firm's own data, and every cell below re-runs on one governed set of facts, so the tiles, the chart and the table cannot disagree with each other.

The model writes the query; the database returns the number, and the query travels with it. A figure can still be wrong -- a query can ask the wrong question -- but it cannot be made up, and it can be checked, because the query sits beside the answer.

Once a query is accepted it is saved with the report and the model leaves the loop. That is what makes a schedule dependable: it lands at seven on Monday and runs the way it ran last Monday.

Published as a report for the people who read it, kept as a notebook for the people who maintain it -- and the same facts drive a live WIP screen for the billing desk, where any difference from the month-end figure is a timing difference, not a data one.

Fictional firm, fictional timekeepers

billing_review.py▶ ran 0.2s

[1]

# "fees by timekeeper for August, and flag WIP over 60 days"
month = mo.ui.dropdown(months, value="2026-08", label="Month")
facts = ledger.fees(month.value)   # generated query, shown with the result

[2]

Month Fictional data · Doe & Roe LLP

Fees billed$0

Realization0%

WIP > 60 days$0

Trust balance$0

Hours0.0

[3]

Fees billed by timekeeper

[4]

Timekeeper	Fees	Share
JHM	$38,700	26.3%
ATR	$39,400	26.7%
LKO	$30,150	20.5%
PDS	$23,850	16.2%
MRC	$15,300	10.4%

03 · Reconciliation

Three-way reconciliation, drawn

Bank statement, trust ledger and the sum of the client ledgers post entry by entry and land on one line. Then one entry is keyed wrong, the line breaks over the column that owns it, and the check catches it before month end.

Animated because a control is easier to trust once you have watched it catch something.

Fictional figures

04 · Digital assistants

It proposes; a person approves

An assistant that lives inside the systems the firm already runs, scoped to the data its job needs -- the billing desk sees time, billing and its own mailbox, and nothing else. It answers in plain language and shows the query behind every figure, so an answer can be checked rather than believed.

Anything that reaches a client record, a mailbox or the ledger stops at a gate with the action written out in full. Every question, answer and approval is logged. The assistant does the chasing; the firm keeps the judgement.

Fictional firm, fictional matter · built into a firm's own systems, not a feature of RNDBI Work

Example · fictional data Assistants · on duty

Billing deskIntakeService deskReporting

Kara · billingWhich matters have unbilled time over sixty days, and who owns them?

Billing deskEleven matters, $84,200 unbilled past sixty days. Nine sit with two timekeepers; the oldest is 114 days. ledger.wip_age · read 08:12:04 · 11 rows · show the query

Kara · billingRemind the two of them, and copy the billing partner.

Billing deskDrafted two reminders listing each matter and its age, copied to the billing partner. Ready to send. Held for approval · this sends mail about client matters ApproveEdit the draftNot now

Readstime, billing, matters, its own mailbox Writesnothing without approval Loggedevery question and answer

Expertise

The stack a law firm actually runs on.

Practice and finance systems, the data underneath them, the Microsoft platform around them, and the lifecycle work that keeps all of it current.

Practice & finance

Case managementAdministration, SQL, reporting, integrations
Legal accountingBilling, trust, AR/AP, GL reconciliation
Practice managementMatters, intake, workflow
Document managementDMS administration and migration

Data & reporting

SQL ServerSchema, performance, ETL, recovery
SSRSPaginated and scheduled reporting
Power BIModels, dashboards, row-level security
Microsoft FabricWarehouse, mirroring, one fact layer

Microsoft platform

Microsoft 365Exchange, SharePoint, Teams, governance
Entra ID & Active DirectoryIdentity, access, conditional access
AzureHosting, networking, cost control
Power PlatformPower Automate, Power Apps, connectors

Lifecycle

Legacy systems supportKeep it running while it is still needed
ModernizationRetire, replace or re-platform, in that order of preference
Process automationScripted, scheduled, logged, reversible
InstrumentationMonitoring and reconciliation checks

If a system you rely on is not on this list, ask. Most legal-industry stacks are a combination of the above, and the method carries.

How we engage

Four commitments, stated so a firm can check them against what it gets.

Read firstFindings come with their evidence.

We start by reading how the firm actually runs: the matters, the money, the systems, and the places where they disagree. What we write down carries the query, the log, or the screenshot that makes it true.

Change carefullyProposed, approved, then applied.

Production changes are proposed in writing, approved by the firm, rehearsed where a rehearsal is possible, and reversible where it is not. Nothing reaches a client record on our own say-so.

Enforce where it holdsControls live where they are hardest to bypass.

A rule in application code holds until someone is in a hurry. A rule in the database, the identity system, or the scheduler holds when the application, or the person, is wrong.

Leave it yoursDocumented, handed over, no lock-in.

Everything we build sits inside products the firm already owns, with documentation the next administrator can read. You should be able to end an engagement without losing anything.

RNDBI Work is our own operations platform for law firms, in invitation-only beta with a small number of firms. It is a product of the same method, not a condition of working with us. What is and is not in place is stated on its Security and Reliability page.

Tell us what is not working.

A report nobody trusts, a migration nobody wants to own, a system the vendor has stopped answering for. Write a few lines about it and a person replies, not a form.

Projects, support, everything elsehello@rndbi.com

Terms, privacy, security questionnaireslegal@rndbi.com