How a Paralegal VA Handles Billing and Time Capture

It is the 27th of the month and a commercial attorney in Claremont has blocked out her evening for what her practice manager politely calls “the billing run.”

She opens Outlook. She opens her diary. She opens WhatsApp, because two of the week’s most substantive conversations happened there, on a Sunday, with a client who was panicking about a supplier. She opens the matter folder on the server and looks at the modified dates on twelve documents, trying to work backwards from a file’s timestamp to a defensible entry.

Somewhere around the 14th, the trail goes cold.

She remembers a call about the restraint clause. She does not remember whether it ran twenty minutes or fifty. She writes 0.5, because 0.5 is the number you write when you are guessing and you do not want to be accused of guessing upwards. She finds three emails to counsel that clearly contain advice, clearly took time, and were clearly never recorded anywhere. She writes 0.3 for all three together, which is generous to the client and unkind to herself, and moves on because there are still nine matters to go and it is already 21:40.

By 23:15 she has produced a bill. It is honest. It is also, by any reasonable estimate, between fifteen and thirty percent smaller than the work she actually did.

Nobody stole that money. No client refused to pay it. It simply never made it onto a page, and the profession has spent a hundred years treating that as a personal failing — a discipline problem, a character problem, something a better lawyer would have handled — rather than what it plainly is.

It is a staffing problem. And it is the single most expensive one in South African legal practice.

The Money That Was Never Recorded

Start with the number everyone quotes and almost nobody sits with.

Clio’s Legal Trends Report puts the average law firm utilisation rate at 38% for 2025 — roughly three billable hours captured out of an eight-hour working day. Realisation sits at about 88%, which means the average practitioner actually invoices around 2.6 hours of that day. Some benchmarking work puts the figure lower still, at 2.9 hours recorded and less after collection losses. A solo practitioner running at 26% utilisation captures about 2.1 billable hours a day, against roughly 3.6 at a firm of twenty or more people.

Read that comparison again, because it is the one that matters. The attorney at the twenty-person firm is not smarter, more disciplined or harder working. She has the same twenty-four hours. She captures roughly seventy percent more of them because somebody else in the building is doing the capturing.

The decay curve on time entry is well documented and brutally consistent. Research compiled through the American Bar Association’s practice-management writing found that recording time once a day, at the end of the day, costs about ten percent of billable revenue. Waiting a full twenty-four hours costs around twenty-five percent. Waiting a week costs between fifty and seventy percent. An American Bar Foundation study reached the same conclusion by a different route: weekly time entry captures materially fewer hours than daily or contemporaneous entry.

Practitioners who audit this themselves land in a similar place. Consultants who run the exercise — pick three timekeepers, have them keep an honest contemporaneous paper log for one day, compare it to what got billed — report a consistent gap of five to fifteen percent of billable time, concentrated in email, unscheduled calls and short tasks. A LexisNexis survey of attorneys found they billed roughly six hours out of a nine-hour day, a shortfall the company valued at around a third of potential annual billings per fee earner.

A firm that improves nothing about its marketing, its rates or its client mix, and simply captures its time on the day the work happens, gives itself a raise. The work has already been done. It has already been paid for in evenings and weekends. It is sitting unrecorded in an inbox.

Now do the arithmetic on a South African practice. Take an attorney billing at R2,200 an hour, working 220 days a year. A conservative twelve percent capture gap on 2.5 recorded hours a day is roughly 0.3 hours daily. That is 66 hours a year, or about R145,000 — per fee earner, every year, indefinitely, from work that was performed, delivered and never charged.

Multiply by three fee earners and you are looking at the cost of a senior associate, evaporating annually into the space between doing the work and writing it down.

Why Timekeeping Always Loses

Every managing director has sent the memo. Contemporaneous entries. Daily timesheets. No month-end batching.

The memo is correct. It also decays within about three weeks, at every firm, in every jurisdiction, and it is worth being precise about why, because the reason is not laziness.

Timekeeping loses every priority contest it enters. At 16:00 on a Thursday, with a client waiting and a deadline moving, revenue work beats administrative work — and time entry, however much it determines whether the firm survives, presents itself in that moment as administrative work. A timekeeping policy asks a practitioner to interrupt billable activity in order to record billable activity. It asks the person under the most pressure to add a task. Policies that require people to act against their immediate incentives do not fail occasionally. They fail structurally.

The task-switching residue makes it worse. You review an agreement for forty minutes, get pulled onto something urgent, come back to it for another fifteen after lunch. What gets recorded is the forty you remember. The fifteen evaporates. Then there are the matters you touch for a moment — reading an incoming notice, a two-paragraph status update to a client, checking whether the other side has filed. Each is trivial alone. Across a busy litigator’s day they routinely add up to an hour.

Research on legal work specifically finds that research sessions are the hardest category of all, because they are exploratory: the attorney does not know at the outset how long the session will run or which matter it properly belongs to. Travel is the second-hardest, not because it is difficult to reconstruct but because it is administratively neglected — time to a hearing or a client’s offices is billable under most fee agreements, and most practitioners default to recording only the time on site.

None of this is a discipline problem. It is a design problem. The person best placed to do the work is structurally the worst placed to record it, and the profession keeps trying to solve that with willpower.

The firms that have solved it did something else. They put a second person in the loop.

What a Paralegal VA Actually Touches: The Anatomy of a Bill

“Billing support” sounds like one task. It is at least seven, and only two of them require a practising certificate.

Here is the honest anatomy of getting work from performed to paid:

1. Activity capture. Gathering the raw evidence of what happened — calendar entries, sent items, call logs, document modification records, court diary, travel. This is retrieval, not judgement.

2. Draft narrative construction. Turning that evidence into a proposed entry with a description a client can read and a taxing master could defend. “Research” is a rejected entry. “Researched recent authority on the enforceability of restraint-of-trade clauses in the medical sector; prepared two-page note for client” is a paid entry.

3. Matter and code allocation. Assigning each entry to the correct matter, the correct fee earner, the correct rate and — where the client requires it — the correct task code.

4. Attorney review and approval. The practitioner reads the draft entries, corrects what is wrong, adds what is missing, deletes what should not be charged, and signs off. This layer is non-delegable and always will be.

5. Pre-bill assembly. Compiling approved entries with disbursements, sheriff’s fees, counsel’s fees, courier and search costs into a draft bill, checking it against the fee agreement and the cost estimate that was issued at the start of the matter.

6. Invoice dispatch and record. Sending it, filing it, updating the WIP and debtors reports, making sure the matter file contains what it needs to contain.

7. Follow-up. Structured reminders at defined intervals, escalation lists, aged-debtor reporting, and flagging the accounts that need a partner’s phone call rather than another email.

Six of those seven are administrative execution against attorney-set rules. One is professional judgement. In most South African practices, all seven currently sit on the desk of the person billing at R2,200 an hour — which is why five of them get done badly at 22:00 on the 27th.

A paralegal VA takes layers one, two, three, five, six and seven. The attorney keeps layer four, where she belongs, and it takes her twenty minutes rather than four hours, because she is reviewing a draft instead of building one from an empty screen.

The moment of professional judgement in a month’s billing occupies perhaps forty minutes. The other twelve hours are retrieval, formatting, arithmetic and follow-up. The profession has known this for a century — it is why paralegals exist. What changed is that the person doing the retrieval no longer has to sit down the corridor.

Section 35: The Compliance Layer Nobody Has Time For

South African billing carries a statutory obligation that a great many practices treat as a formality and a growing number of clients treat as leverage.

Section 35(7) of the Legal Practice Act 28 of 2014 requires an attorney — or a trust account advocate — to give a client a written cost estimate notice when instructions are first received, or as soon as practically possible afterwards. The notice must set out the likely financial implications, including fees, charges, disbursements and other costs; the hourly rate; an explanation of the client’s right to negotiate that rate; and an outline of the work envisaged at each stage.

Section 35(8) requires that every aspect of the notice be explained verbally as well. Section 35(9) requires the client’s written agreement to the envisaged services and estimated costs.

Then comes the part that concentrates the mind. Section 35(10) provides that non-compliance constitutes misconduct. Section 35(11) provides that where a practitioner has not complied, the client is not required to pay legal costs until the Legal Practice Council has reviewed the matter and determined what should be paid.

That is not a filing requirement. That is a live commercial risk sitting on every matter opened without a properly issued and acknowledged estimate — and it is exactly the kind of task that gets skipped in the first busy week of an urgent instruction, because the substantive work feels more urgent than the paperwork about the substantive work.

The courts have been sharpening the same point from the other direction. In Sports Tavern & Restaurant and Others v Executor Estate Late Santos (2025), the appeal court criticised a R20,000 day fee claimed in an unopposed application, and reminded the taxing master that instances of overreaching should be reported to the registrar and referred onward to the Legal Practice Council. The direction of travel is unmistakable. Bills are being read more closely, by clients and by courts, and vague entries attract scrutiny rather than payment.

A paralegal VA handles this as a standing process rather than a good intention:

Cost estimate notice drafted from a firm template within twenty-four hours of a new matter being opened, populated with the correct rate, scope and stage outline, and sent to the attorney for approval and signature.
Written client acknowledgement chased, received and filed to the matter, with the file flagged as incomplete until it lands.
A revised estimate triggered automatically when a matter crosses an agreed threshold — say seventy percent of the estimated fee — so that scope creep is raised in writing before it becomes a fee dispute.
Disbursement records maintained contemporaneously, with supporting vouchers attached, so a bill of costs can be drawn without a two-week archaeology project.
Narratives written to a standard that survives a client’s line-by-line reading, because that is now the standard they are read to.

For firms doing recoverable work, the same discipline pays twice. A bill drawn from complete, specific, contemporaneous entries survives taxation. A bill reconstructed from memory gets taxed down, and the difference comes out of the practice.

The Human in the Loop: Why Passive Capture Is Not the Whole Answer

There is a category of software that promises to solve all of this by watching you work.

It is genuinely good, and it is worth being fair about what it does well. AI timekeeping platforms run passively, monitor activity across email, documents, calls and research, and present the practitioner with draft entries to review rather than a blank timesheet. Early adopters report capturing between ten and thirty percent more billable time and cutting narrative-writing effort dramatically. One litigation firm found the software surfaced dozens of short client calls a month that had never been billed. That is real money, and any practice serious about capture should be looking at these tools.

But the tools do not close the loop, and the reasons are instructive.

A time entry is a representation, not a log. When an attorney submits a bill, she is asserting that the work described was done, that it was reasonable, and that it was necessary for the client’s matter. Software can observe that a document was open for fifty minutes. It cannot know that thirty of those minutes were spent on a related matter for a different client, that the tab was left open over lunch, or that the last twenty minutes were the practitioner teaching herself something she should already have known and cannot in conscience charge for. Those judgements are professional, and the person making them carries the consequences.

Allocation is contextual. A single research session frequently spans three matters that share a legal question and nothing else. Passive capture assigns time to whatever it can see. A human who knows the file assigns it to whatever is true.

Narrative quality is a trust document. Institutional clients, insurers and government departments increasingly enforce billing guidelines that dictate narrative detail, prohibit block billing and require task codes. Vague entries, block-billed entries and miscoded entries are rejected, and each rejection means rework, delay and often a write-off when someone decides fixing it is not worth the effort. BigHand’s annual survey found 58% of firm leaders reporting that outside counsel guidelines had lengthened their billing and collection cycles, with two-thirds reporting delays of sixty days or more. Software can format an entry. It cannot decide what level of detail a particular client will accept without also feeling billed for the billing.

Privilege has an edge. A narrative must contain enough detail to justify the charge and not so much that it discloses privileged strategy on a document that may travel. That line is judgement, and it moves depending on who is reading.

The workable arrangement is not human or machine. It is machine captures, human verifies and constructs, attorney approves.

Automation is very good at answering “what happened.” It is unreliable at answering “what should be charged, described how, to whom, and why.” The second question is the one that gets paid.

This is where a trained paralegal VA sits, and why the role is not a downgrade from software. She uses the software. She reads its output against the matter file, corrects the allocations, writes narratives in the firm’s voice, flags the entries that need the attorney’s eye and presents a clean, decision-ready pre-bill. The attorney’s job shrinks from four hours of construction to forty minutes of review.

The South African Advantage, Applied to a Clock Problem

Most arguments for South African support staff run on cost and English. Both are true and both are secondary here. For time capture specifically, the decisive variable is the clock.

Recall the decay curve: roughly ten percent lost at end of day, twenty-five percent at twenty-four hours, fifty to seventy at a week. Every hour between the work and the entry costs money. Which means the single most valuable property in a billing assistant is that they are awake and working while the work is happening.

South Africa sits at GMT+2. For a Johannesburg or Cape Town practice, that is not an overlap question at all — it is the same working day, the same court diary, the same 16:00 crunch. A VA who catches the practitioner at 16:30 with six draft entries and one question (“the call with counsel — was that the Mahlangu matter or the other one?”) is intervening inside the window where recall is still accurate. A support resource operating eight or ten hours behind is, by construction, always working on yesterday, which is precisely the twenty-five percent loss the data describes.

The same timezone position is why South African firms have become the default support base for UK and European practices, where GMT+2 delivers a full working-day overlap with London and most of Europe, with the South African assistant starting earlier and covering the first two hours of the UK morning before the office fills.

The rest of the case is straightforward:

Language and register. South African legal English is the same legal English. A South African paralegal VA writes “restraint of trade,” “letter of demand,” “notice of motion,” “rescission,” and knows what a Rule 41(a) notice is without a glossary.
Legal system familiarity. South Africa’s Roman-Dutch and common-law hybrid means a locally trained paralegal understands the structure of a matter file, the role of the taxing master, the distinction between party-and-party and attorney-and-client costs, and the practical mechanics of a bill of costs. That is not knowledge a general offshore administrator has.
Regulatory context. POPIA is a serious statute with a serious regulator, and South African professionals are trained under it. For a firm handling privileged client material, the difference between a support person who has been trained on data protection and one who has not is not a nice-to-have.
Cost against quality. The comparison that matters is not VA hourly rate versus VA hourly rate. It is the cost of the support against the revenue it recovers. A capture improvement of ten percent on three fee earners at South African commercial rates typically covers a dedicated paralegal VA several times over — which is why the sensible way to evaluate this is as a revenue recovery function, not an overhead line.

The structural context makes it more urgent, not less. As of February 2025 South Africa had 14,242 sole practitioner firms, against just 21 firms with more than fifty attorneys. Three-quarters of the country’s law firms have between one and ten fee earners. Reporting on the sector notes that 77% of small firms say they spend too much time on administration, and that South African SMEs lose an estimated 202 days a year to administrative work, at an average cost above R530,000 — with invoice generation and processing among the largest single components.

That is a market of thousands of practices where the person who wins the matter, does the work, writes the bill, chases the payment and reconciles the trust account is the same person. The capture gap is not an inefficiency in that model. It is the model.

Building It: What Handover Actually Looks Like

The failure mode when a firm brings in billing support is predictable. Someone hands over “the billing” as a single undifferentiated task, nothing is documented, the first month’s bill goes out with three errors, and the attorney concludes it is faster to do it herself. Which it now is, permanently, because the process was never built.

Here is a sequence that works.

Week one — capture only, no drafting. The VA builds a daily activity sheet for each fee earner from calendar, sent items, call logs and document activity. She does not write narratives yet. She writes a list of events and sends it at 16:30 with the standing question: what is missing, and what should not be there? The attorney’s answers are the training data.

Week two — draft narratives against a house standard. The firm supplies ten model entries: five excellent, five that were rejected or taxed down, with the reason. The VA writes drafts to that standard. The attorney corrects. Two rounds of this establishes voice more reliably than any written brief.

Week three — matter allocation and rate logic. The VA takes over assigning entries to matters, fee earners and rates, and flags ambiguity rather than guessing. A single standing rule handles most of it: if you are not certain which matter it belongs to, do not allocate it, ask. Wrong allocations are far more expensive than questions.

Week four — the full cycle. Pre-bill assembly, disbursement reconciliation, cost estimate compliance check, dispatch, and the follow-up cadence. The attorney’s involvement settles into two touchpoints: a daily five-minute confirmation, and a monthly review of the assembled pre-bill.

Then the ongoing rhythm, which is where the money actually is:

Frequency	What the VA does	What the attorney does
Daily, 16:30	Draft entries circulated for confirmation	Confirm, correct, add missed items (5 min)
Weekly	WIP report, unbilled-time flag, matters approaching estimate threshold	Read; decide on any revised estimates
Monthly	Pre-bill assembled, disbursements reconciled, narratives finalised	Review and approve (30–40 min)
Monthly + 7 days	Invoices dispatched, debtors report updated	Nothing
Monthly + 21/35/50 days	Structured follow-up, escalation list prepared	Phone calls on flagged accounts only

And the boundary, which matters more than anything above it. A paralegal VA does not give legal advice, does not sign anything, does not decide what may properly be charged, does not touch the trust account, does not authorise transfers between trust and business accounts, and does not communicate substantive positions to a client. Trust accounting in South Africa sits under the Legal Practice Act and the LPC’s rules, and the responsibility is the practitioner’s and non-transferable. A VA can prepare, reconcile, schedule and flag. The signature, the transfer and the advice stay with the attorney.

Any provider who is vague about that line should be treated as a risk rather than a bargain.

Why “Managed, Not Matched” Decides Whether This Works

The billing function has an unusual property: it is simultaneously the most valuable thing to delegate and the least forgiving of turnover.

A billing assistant who leaves after four months takes with her the accumulated knowledge of forty matters, three clients’ billing preferences, the firm’s narrative voice and every judgement call the attorney taught her. The replacement starts at week one. Do that twice and the practice has spent a year of management time to end up where it started — which is exactly what firms report after cycling through freelance marketplace hires.

This is the distinction VAConnect built the business around, and it is the reason the model is described as managed, not matched. The difference is what happens after placement.

A marketplace matches. You post a role, you screen a list, you hire, and from that moment you are the manager, the trainer, the quality control and the contingency plan. If the person underperforms, that is your problem. If they leave, that is your problem. If they are ill during the last week of the month, the bill does not go out.

A managed agency retains responsibility. VAConnect sources through VAJobs.co.za, trains through VAVarsity, monitors workload and wellbeing through Atomic Energy, and runs monthly performance reviews under VAPIness rather than waiting for a client to complain. If a VA is not performing to the agreed standard, VAConnect replaces them and manages the transition — no fee, no friction. There is backup cover, so the last week of the month is not a single point of failure. The firm has been operating since 2008, moved to the managed model in 2014, and has delivered over 250,000 hours of VA work with a team of 35-plus behind the placements.

For a function where the cost of a gap is a month of unbilled work, that structural difference is not a marketing distinction. It is the whole risk profile.

The Competitive Gap Is Wider Than It Looks

Here is what should be uncomfortable about all of this.

Two practices of identical size, identical rates and identical quality of legal work will produce materially different revenue depending on nothing more than who writes down the time and when. Not who is the better lawyer. Not who wins more. Who captures.

The firm running end-of-month reconstruction is operating at something like 26–38% utilisation with a further ten to twenty-five percent shaved off by the decay curve, invoicing late, and absorbing write-downs on entries it cannot defend. The firm running same-day capture with a dedicated paralegal VA is invoicing more of the same work, invoicing it sooner, and defending it better — while the attorney gets back roughly one full working week per year that was previously spent on billing runs.

Compounded across a year, that gap is not a rounding error. It is the difference between a practice that can afford to hire and one that cannot.

The most expensive hour in a law firm is the one that was worked, delivered, valued by the client and never written down. It costs the full rate and produces nothing.

The work has already been done. Someone just has to write it down while the memory is still warm — and that person does not have to be, and should never have been, the person billing at R2,200 an hour.

DIY Coordination vs Generic Freelancers vs VAConnect

Dimension	DIY (attorney does it)	Generic freelancer / marketplace VA	VAConnect Paralegal VA
When time is recorded	End of month, reconstructed from memory	Whenever briefed; typically weekly batches	Same day, drafted by 16:30 in your timezone
Expected capture loss	25–70% against contemporaneous entry	10–25%; depends entirely on hire	Approaches the same-day floor (~10%) and falls as the process matures
Cost of the hour spent	R2,200+/hr of fee-earner time	Low hourly rate, high management load	Managed rate; management load carried by the agency
Legal context knowledge	Complete	Usually none; must be taught from zero	SA legal system familiarity; VAVarsity-trained
Narrative quality	Excellent when there is time; thin at 22:00	Highly variable; generic descriptions invite write-downs	Written to a house standard built from your own accepted and rejected entries
s35 cost estimate compliance	Depends on remembering during a busy week	Only if explicitly briefed and monitored	Standing process: drafted, chased, filed, threshold-triggered revisions
Disbursement records	Reconstructed at bill-of-costs stage	Ad hoc	Maintained contemporaneously with vouchers attached
Debtor follow-up	Sporadic; usually only when cash is tight	Rarely included	Fixed cadence with escalation list and aged-debtor reporting
Cover when the person is unavailable	Bill goes out late	None — you absorb it	Backup cover; billing cycle protected
If it is not working	It is still your problem	Rehire, retrain, restart from zero	Replaced and transitioned at no additional cost
Who owns quality	You	You	VAConnect — monthly reviews, proactive flagging
Ramp to useful output	n/a	6–12 weeks, if the hire sticks	Meaningful output in week one; full ramp 2–4 weeks

The practical next step is smaller than it sounds. Pick one fee earner. Run a single honest week — a contemporaneous paper log of everything done, compared against what actually got billed for those five days. Annualise the difference.

Almost nobody who runs that audit comes back arguing about whether the gap exists. The argument shifts to what to do about it, which is the argument worth having.

Book a 30-minute discovery call with VAConnect and we will scope what a paralegal VA would take off your desk in the first thirty days — starting with the entries you are currently writing at 22:00 on the 27th.

Sources

Clio, Legal Trends Report and Law Firm KPIs / benchmarks (2025): utilisation 38%, ~3.0 billable hours per eight-hour day, realisation 88%, ~2.6 hours invoiced; median total lockup 93 days.
Accounting Atelier, Law Firm Financial Benchmarks (2026): solo at 26% utilisation captures 2.1 billable hours daily versus 3.6 at firms of 20+; larger firms outperform through systems, not effort.
Ann Guinn / American Bar Association practice-management research, as compiled by Bill4Time: ~10% revenue loss at same-day entry, ~25% at 24 hours, 50–70% at one week.
American Bar Foundation study on time-entry frequency, cited in TimeSolv’s billable hours guide: weekly entry captures fewer hours than daily or contemporaneous entry.
Ajax, How Law Firms Stop Losing Billable Hours (2026): audited leakage of 5–15% of billable time, concentrated in email, unscheduled calls and short tasks; task-switching residue; timekeeping loses every priority contest.
US Tech Automations (2026): legal research as the hardest category to capture accurately; travel time as systematically under-recorded; email as the largest single leakage source.
LexisNexis attorney survey: ~6 billable hours recorded out of a 9-hour working day.
Legal Practice Act 28 of 2014, s35(7)–(11): written cost estimate notice, verbal explanation, written client agreement, non-compliance as misconduct, and suspension of the client’s obligation to pay pending LPC determination.
Sports Tavern & Restaurant and Others v Executor Estate Late Santos (HCAA 01/2023) [2025] ZALMPPHC 17, as reported by Cliffe Dekker Hofmeyr: overreaching to be reported to the registrar and referred to the LPC.
PointOne, Legal Billing and Timekeeping Best Practices (2026) and BigHand annual survey: narrative detail and block-billing rejections; 58% of firm leaders report OCGs lengthening billing cycles, two-thirds reporting delays of 60+ days.
LawSites / LawNext, Billing in the Age of AI (2025): early AI-timekeeping adopters reporting 10–30% more billable time captured and large reductions in narrative-writing effort.
Global Legal Market, Law Firm Consolidation in South Africa (2026): 14,242 sole practitioner firms as of February 2025 against 21 firms with more than 50 attorneys; 77% of small firms report excessive admin; SA SMEs losing ~202 days a year to administration at ~R532,801 average annual cost, with invoice generation and processing among the largest components.
Global Legal Market, Less Admin, More Growth: more than three-quarters of South Africa’s 10,930 law firms have one to ten fee earners.
VAConnect company data (vaconnect.co.za): founded 2008, managed model since 2014, 250,000+ hours delivered, 35+ team, VAJobs.co.za sourcing, VAVarsity training, Atomic Energy monitoring, VAPIness performance reviews, no-fee replacement guarantee, GMT+2 positioning, meaningful output in week one with 2–4 week full ramp.