The specific pain: administration billed at a fee earner's rate
Every hour a solicitor or associate spends on work that isn't legal work still gets counted against the same clock a client's matter runs on. Opening a new file means creating a matter number, building a folder structure, and putting together a checklist of what the file needs — done properly, that's twenty minutes a fee earner spent instead of turning to the substantive point they were meant to be looking at. An engagement letter gets drafted from whichever template is closest, with the scope and the rate structure typed in by hand and checked against the intake notes a second time to catch a transposed figure. The client care pack that has to go out alongside it — the letter, the privacy notice, the complaints procedure, the identification requirements — gets assembled from four separate documents because nobody built a single pack that does it on its own.
Then the matter runs, and the drag continues. A document a client promised three weeks ago still hasn't arrived, and nobody chases it until a fee earner notices the gap themselves, usually the night before it's needed. Correspondence lands in a shared inbox and someone has to work out which of forty open matters it belongs to before it can be filed. Time slips because the twenty-minute call at 4:45 on a Friday never made it into the system, and a week later nobody can reconstruct what it was for. None of this needs a law degree. All of it currently gets done by someone who has one, because nobody built the pipeline that would do it instead.
What the system does, day to day
Once it's live, the pipeline sits inside the systems your firm already runs and triggers off events that already happen — a matter gets approved, a document lands in an inbox, a billing cycle date arrives — rather than waiting for someone to remember a task. Its job is the same handful of things, run the same way every time:
- Opens the matter and builds the file the moment an intake is approved.
- Assembles the engagement letter and client care pack from the intake data already on file.
- Collects the checklist of documents a matter type needs, chases whatever's missing on a schedule, and checks each arrival against the checklist for completeness.
- Reads incoming correspondence — email, portal messages, scanned post — and files it to the right matter automatically instead of waiting for someone to sort it.
- Prompts a fee earner to log time against activity the system can see, rather than asking anyone to reconstruct a week from memory.
- Prepares the billing run and drafts a first-pass narrative from logged time and activity, for review before it goes anywhere.
- Runs a closing checklist and archives the file once a matter is done.
Every one of those steps runs the same sequence on the fortieth matter as it did on the first. That consistency is the entire value of a fixed pipeline over a person doing it by hand: it doesn't skip a step because it's Friday afternoon, and it doesn't handle this matter differently because the fee earner running it is new.
Pipeline walkthrough: opening a matter and getting the client care pack out
- Trigger. A fee earner marks a new matter approved in the practice management platform, with the intake data already captured from a form, a call, or an intake conversation.
- File setup. The pipeline creates the matter number, builds the folder structure in the document management system from the template for that matter type, and populates the document checklist the matter needs.
- Engagement letter and client care pack. The engagement letter drafts itself from the matter type's template, merging in the client's name, the scope of work, and the rate structure already agreed — every field a fee earner would otherwise type by hand. The client care pack — engagement letter, privacy notice, complaints procedure, identification requirements — assembles alongside it as one pack instead of four separate lookups. Nothing goes out until the fee earner responsible for the matter reviews and approves it; the draft sits in front of them, not in the client's inbox.
- First collection request. Once approved, the pack goes out to the client along with a first request for whatever's still outstanding on the checklist.
- Completeness checks as documents arrive. Each item that comes back gets matched to its checklist line and checked for completeness — signed where a signature is required, dated where a date matters, the correct document rather than a similarly named one. Anything that doesn't match is flagged to the fee earner instead of marked received on a guess.
Pipeline walkthrough: keeping the file current while the matter runs
- Correspondence intake. Every inbound email, portal message, and piece of scanned post gets read for the reference number, the parties named, and the subject line, matched to the open matter it belongs to, and filed there directly — the same folder and the same access restrictions that already apply to that matter, untouched.
- Ambiguous items held for a person. Anything that doesn't match cleanly — no reference number, two matters with the same client name, a document that could plausibly belong to either — gets flagged to the fee earner instead of filed on a best guess.
- Document chasing continues on schedule. The same checklist from matter opening keeps running for the life of the file. An outstanding item that passes the firm's follow-up interval triggers a chaser naming exactly what's missing, sent again on the same schedule until it's resolved or a person tells the pipeline to stop.
- Time capture prompts. At the end of each day, the pipeline looks at what it can see happened against the matter — documents drafted, calls logged, correspondence sent — and prompts the fee earner with a suggested time entry and a duration drawn from that activity. The fee earner confirms, adjusts, or rejects it. Nothing posts to the ledger without that step.
Pipeline walkthrough: the billing run and closing the file
- Billing run trigger. A billing cycle date arrives, or a fee earner marks a matter ready to bill.
- Narrative assembly. The pipeline pulls every logged time entry for the matter, groups it by phase or activity, and drafts a narrative in the firm's standard bill format — a complete first draft built from what's already been logged, not a blank page.
- Review, not release. The draft bill routes to the fee earner or billing partner responsible for the matter. They edit the narrative, adjust the figures, and approve it. Nothing goes to a client until a person has signed off — the pipeline prepares the bill; it does not send one.
- Closing checklist. Once a matter is marked closed, the pipeline runs the closing checklist the firm has defined for that matter type: outstanding items resolved, final correspondence filed, the engagement letter and client care pack present in the file, any client property returned or accounted for.
- Archiving. Once the checklist clears, the file moves into archive storage under the firm's retention policy, and the matter's status updates across every connected system — no fee earner spending an afternoon boxing up a file by hand.
Where the pipeline stops
Every step above runs on data a person has already put into the system — a checklist the firm defined, a date a fee earner entered, a template the firm approved. That's the boundary that matters most: the pipeline never calculates a limitation period, a filing deadline, or any date that carries legal consequence. It can chase a date once a lawyer has already entered it — reminding a fee earner that a deadline recorded in the diary is three days out, following up on a document a court date depends on — but it never works out what that date should be. That calculation, and the diary process behind it, stays with a lawyer and the firm's own systems, every time.
Conflict checking works the same way. The pipeline can collect the information a conflict check needs — the parties on a new matter, gathered during intake — and get it into the record before the file opens. It does not compare those names against the firm's client list and decide a matter is clear. That check remains the firm's, exactly as it does today.
Where a step in this administration stops being fixed and starts needing a person to weigh something up — judging what a document discrepancy means, deciding whether a status update needs a partner's input before it reaches a client — that's the point where a rules-based pipeline hands off to something built to exercise judgment inside guardrails, which is what our AI agents for law firms are for. Most firms end up using both: a pipeline for the steps that never change, and an agent for the ones that occasionally do. We'll tell you honestly which parts of your matter administration need which, rather than building an agent where a simpler pipeline does the job. For what separates the two in plain terms, see what an AI workflow actually is.
Integration notes
None of this works unless it sits inside the systems your firm already runs, so here's how the connections are actually built.
Practice management platform. Matter creation, checklist items, and status updates read and write through your practice management platform's API wherever one exists. Where it doesn't, we typically work through a scheduled export or a controlled screen-level integration rather than asking a firm to change platforms.
Document management system. Folder structures, filed correspondence, and the documents collected against a checklist all land in your document management system using the structure and permissions your firm already has, not a second system that has to be kept in sync by hand.
Calendar and billing software. The pipeline reads dates a fee earner or your calendaring system has already set — it does not write new ones with legal consequence — and it writes time entries and draft narratives into your time and billing software for a fee earner to confirm.
Correspondence. Email, client portal messages, and scanned post are all read the same way, so a letter that arrives as a scan gets filed exactly as reliably as one that arrives by email.
Matter-level access and confidentiality. Every connection respects the access rules your firm already runs. A pipeline instance sees only the matters and documents a defined task requires, existing ethical walls and permission boundaries apply to it exactly as they apply to a person on staff, and nothing it touches trains a shared or public model. Confidential and privileged material is handled with the same care your firm already applies to it — narrow access, logged activity, no exceptions built in for convenience.
This is the same shape as our wider workflow automation work, applied to what a law firm specifically runs on, and it follows the same four-stage process as every Calfy build — discover, design, build, run — with a clear price agreed before work starts.