| name | budget-and-fee-manager |
| description | Matter budgeting and ongoing WIP/variance monitoring. Build phase-based fee estimates at matter setup, run bottom-up budgets by jurisdiction or workstream, calculate contingency, and structure AFA arrangements (fixed fee, capped fee, phased fixed fees). Ongoing monitoring: WIP tracking against budget, proportionality assessment (spend vs progress), variance commentary with root cause analysis, forecast-to-complete, realisation monitoring, write-off analysis. Trigger on: 'build a budget', 'fee estimate', 'what will this cost', 'WIP review', 'budget vs actual', 'how are we tracking against budget', 'we're over budget', 'realisation is poor', 'what's our ETC', 'budget for the German workstream', 'model the financial impact of this scope change', 'draft a fee adjustment', 'write-off analysis', 'how much contingency', 'AFA structure', 'fixed fee estimate', 'budget update', 'forecast to complete'. |
Budget and Fee Manager
Purpose
Build, monitor, and adjust fee budgets across the full matter lifecycle. At matter setup: translate scope into a phase-based fee estimate with contingency and AFA structure. In flight: run WIP reviews that explain variance, not just report it. When spend has outpaced delivery: assess recoverability and present options. When scope changes are confirmed: model the financial impact.
This skill produces the financial analysis that status-report-drafter summarises. It receives confirmed OOS conclusions from scope-change-controller and models their financial impact. It does not execute billing (billing-cycle-manager handles that) and does not produce the client-facing financial summary (status-report-drafter consumes this skill's output and presents it).
Identifier Block — Required Before Any Output
Stop and confirm before proceeding:
Client: [Name] Client number: [Number]
Matter: [Name] Matter number: [Number]
Report date: [Date] Prepared by: [LPM name]
If any identifier is missing, ask for it. Do not produce output without the complete block.
Operating Modes
Mode 1 — Budget Build
At matter setup: produce a phase-based fee budget from agreed scope. Triggered by a partner asking for an estimate, a scoping brief, or the structured output of matter-intake-scoping.
Input: scope description (email, brief, or matter-intake-scoping output), jurisdiction list, team structure, indicative timeline, AFA preference if known.
Mode 2 — WIP Review
Month-end or mid-matter: assess actual spend against budget, explain variance, produce a forecast-to-complete. The standard financial health check.
Input: WIP data (pasted figures, uploaded Excel/CSV, or text description), progress narrative or self-reported completion percentages, phase/workstream structure matching the budget baseline.
Mode 3 — Realisation Alert
Spend is disproportionate to progress. The partner needs options now — this is not a scheduled review.
Input: current WIP position, progress assessment, original budget, any known causes. Often arrives as a forwarded email from a partner or billing manager flagging the position.
Mode 4 — AFA Tracking
Matter running on a fixed fee, capped fee, or other alternative fee arrangement. Track burn against the cap, calculate the breach point, and identify the decision trigger — the point at which the partner must act, before the cap is breached.
Input: agreed fee or cap amount, recorded WIP to date, estimated completion percentage or remaining task description.
Mode 5 — Fee Adjustment
A confirmed scope change (from scope-change-controller) requires financial impact modelling and a fee adjustment communication.
Input: original budget baseline, confirmed OOS description with estimated scope, any commercial context (relationship sensitivity, recovery appetite).
Domain Knowledge — Budget Build
Estimating approach
Three approaches, applied in order of available information:
Analogous estimating — use a comparable completed matter as the reference. Adjust explicitly for scope differences — do not apply a comparable budget without documenting what changed and why. The adjustments are the analytical content; the analogous figure is only the anchor.
Bottom-up estimating — break into phases and workstreams, estimate each component, aggregate. Most accurate when scope is defined. Requires agreement on who does what at which grade. Build at the workstream × phase intersection, not as a lump sum. Lump sums cannot be monitored meaningfully mid-matter.
Parametric estimating — use a unit rate (£X per entity restructured, £Y per jurisdiction, £Z per regulatory filing) calibrated against historical data. Useful for large multi-jurisdiction programmes where per-jurisdiction bottom-up is impractical. Always state the parametric rate and its source.
For most matters: analogous for the initial estimate, bottom-up once scope is agreed and team structure is confirmed.
Phase structure
Budget at phase level at minimum. On complex matters, budget at workstream × phase.
Standard phases for transactional work:
- Scoping and setup — intake, team briefing, kick-off, matter plan build, system setup
- Execution phases — named by legal milestone or deliverable, not "Phase 1/2/3"
- Coordination — multi-jurisdiction oversight, LC management, status cadence. Routinely underestimated; always budget explicitly
- Completion and close — final documents, signing logistics, post-close filings, matter close
Coordination uplift benchmarks (apply to execution phase subtotal):
- 2–5 jurisdictions: 10–15%
- 6–10 jurisdictions: 15–20%
- 11+ jurisdictions: 20–25%
These are starting points, not formulas. Scope complexity, counterparty behaviour, and client communication intensity all push the number up. Flag that it is an estimate with an explicit assumption, not a fixed percentage.
Contingency
Contingency is a named reserve against identified risks — not padding and not a rounding provision. Keep it separate from the base estimate in the budget table so it can be consumed, returned, or re-held against a different risk as the matter evolves.
Ranges by complexity:
- Low (single jurisdiction, well-defined scope, familiar client and matter type): 5–10%
- Medium (2–5 jurisdictions, some structural unknowns, standard matter type): 10–15%
- High (6+ jurisdictions, novel structure, regulatory uncertainty, new client or counterparty): 15–25%
Justify the contingency percentage in 2–3 sentences — name the two or three specific risks it is held against. Unjustified contingency gets cut. Justified contingency with named risks is defensible.
Management reserve is distinct from contingency. It covers scope changes that cannot be anticipated at outset. Do not bundle with contingency. If the partner asks for it to be included, label it separately.
AFA structures
Fixed fee — agreed total for defined scope. Risk sits with the firm. Requires tight scope definition and a functioning scope change mechanism. Build an internal T&M budget regardless — the fixed fee is the billing ceiling, but the internal budget tells you how much margin is being consumed.
Capped fee — T&M up to an agreed cap. Client bears downside risk to the cap; firm bears upside risk beyond it. Budget internally against the cap as the ceiling.
Phased fixed fees — fixed fee per phase, agreed at or before each phase gate. Reduces uncertainty for both parties when later phases cannot be scoped at outset. Recommended for multi-phase transactions with significant structural unknowns.
Success fee elements — base fee plus a contingent element tied to outcome. Budget internally against base only; the contingent element is potential upside, not a cost target.
AFA monitoring rule — at any point on an AFA matter the system must answer: what percentage of the agreed fee has been consumed, and what percentage of the work is complete? If the first number is materially higher than the second, the matter needs attention. Track this explicitly at every WIP review on AFA matters.
Domain Knowledge — WIP Monitoring
The proportionality test
This is the most important metric in any WIP review. Not "how much have we spent?" but "how much have we spent relative to how much work is done?"
60% of budget consumed with 60% of work complete: on track. 85% of budget consumed with 50% of work complete: the trajectory will not land within budget. 30% of budget consumed with 80% of work complete: investigate before declaring it good news.
Run the proportionality test explicitly for every workstream in every review. State the result as a number — do not describe it.
Variance threshold and commentary
Flag every workstream where spend exceeds budget by more than the agreed threshold (default 10–15%; confirm with the partner at matter setup and record in the matter plan). Above threshold, provide root cause commentary using the four-question framework. Below threshold, note the position without extended commentary.
Four-question variance analysis — for every overrun above threshold, answer all four:
- Root cause — the specific cause, not a generalisation. "Germany is complex" is not a root cause. "Entity count was 7, not 3 as scoped" is.
- Pattern — one-off or systemic? A one-off event (counterparty delay, unexpected regulatory requirement) is manageable. Systemic means the remaining budget assumption is also wrong.
- Recoverable? — can the overrun be absorbed within the remaining budget for this workstream, or has the total projected cost already exceeded budget?
- Scope signal? — does this overrun indicate the original scope assumptions were wrong? If yes, hand off to scope-change-controller. Variance caused by scope creep is not a financial problem — it is a scope management problem that happens to have a financial consequence.
Forecast-to-complete (FTC)
The forward-looking metric. Not what has been spent — what will be spent in total, and whether that is within budget.
Calculate two ways and present both:
- Burn rate method — (actual to date ÷ estimated % complete) = projected total. Compare to budget.
- Remaining work method — estimate the cost of tasks that remain, add actual to date. Use the task list from matter-plan-builder if available.
Always produce FTC as a range: lower bound (remaining work completes as scoped at normal efficiency), upper bound (current burn rate continues). A point estimate falsely implies precision. The range is the honest answer and the useful one.
Underspend scepticism
A workstream significantly below budget is not automatically good news. Confirm which of three explanations applies before reporting it:
- Work is genuinely simpler than expected — confirm, return surplus to contingency
- Work has not started or is significantly delayed — flag as a progress risk
- Time has not been recorded against this matter code — recoverable, but needs to be confirmed and corrected before the WIP report is finalised
Never report a material underspend without a confirmed explanation for it.
Self-reported completion scepticism
The last 10% of a workstream routinely consumes 30% of its remaining budget. When a team reports near-completion, ask what specific tasks remain and who owns them. "Tidying up" and "nearly there" are not task descriptions — they are early warning signs.
When self-reported completion is above 80% AND remaining work is described vaguely ("just X left", "nearly there", "final documents"), produce this challenge explicitly in the variance commentary — required, not optional:
"[Description] is not a task description. Confirm specifically: what documents or deliverables remain, who is drafting them, whether any client or counterparty inputs are outstanding, and whether there are any dependencies on TU2/TU3/other workstreams. Until confirmed, ETC upper bound assumes [X]% of workstream budget remains."
Apply explicit scepticism even when the proportionality test shows a workstream on track — a −3pp gap at 85% self-reported completion is not confirmation of good health if the remaining task description is vague.
Realisation
Realisation = fees billed ÷ fees recorded at standard rates. A realisation problem means the firm is working more than it is recovering.
Monitor:
- Write-offs — fees recorded but not billed. Flag when cumulative write-offs exceed 5% of matter budget. Distinguish instructed write-offs (partner decision, often correct and appropriate) from incurred write-offs (inefficiency or unbillable WIP).
- Rate discounts — agreed discounts should be reflected in the budget as the discounted rate, not the rack rate. A budget built at rack rate for a matter running at a 15% discount is wrong before it starts.
Query/chase loop
When WIP data contains an anomaly that cannot be explained from available information:
- Identify the anomaly precisely
- Draft a specific query to the responsible team — produce a draft email or message, not a list of action items. Action items tell the LPM what to do. A draft query does it.
- Incorporate the explanation into the variance commentary once received
- If the explanation reveals OOS work, hand off to scope-change-controller immediately
Query draft format — required for every unexplained anomaly:
To: [Name / role]
Re: [Matter] — [Workstream] WIP query
[Workstream] shows [£X] recorded against [£Y] budget at [Z]% self-reported completion [/ zero progress / no time entries since [date]]. Before the WIP review is finalised, I need the following confirmed by [date]:
1. [Specific question — what happened, what is the cause]
2. [Specific question — what remains and who owns it]
3. [Specific question — are there any cost items that should be assessed for write-off]
If any of the above indicates work outside the agreed scope, this will be referred to scope-change-controller immediately.
Financial disclosure sequencing
Do not communicate specific overrun amounts to the client until: (a) the WIP position has been reconciled, (b) write-offs have been processed and the net billing figure is confirmed, (c) the commercial response has been agreed with the partner. Until then: "we are monitoring fees in [workstream] — we'll provide a full update in the next financial report." Provisional overrun figures that may change damage trust more than a delayed but accurate number.
Output Format
All outputs from this skill are produced as .docx files unless the user explicitly requests otherwise. Skill outputs are matter records — they belong in the matter folder.
Mode 1 — Budget table
Required column header row (use exactly):
| Phase | Workstream | Grade | Est. hours | Rate (£/hr) | Subtotal (£) | Notes |
Summary rows below the detail:
| | | Base estimate total | | | £[X] | |
| | | Coordination uplift ([X]%) | | | £[X] | |
| | | Contingency ([X]% — [brief justification]) | | | £[X] | |
| | | Total budget | | | £[X] | |
For AFA matters, add beneath the total:
"AFA structure: [Fixed / Capped / Phased fixed]. Agreed fee / cap: £[X]. Internal margin at budget: [X]%. Partner review required if margin falls below [X]%."
Named-firm attribution rule: Never reference a named firm anywhere in skill output — in documents, tables, or conversational text. This includes attributing rates, policies, practices, or organisational structures to any named law firm. The skill does not know any firm's actual structure, rates, or policies. Use "assumed — confirm with Pricing", "confirm with Finance", or "firm policy — confirm before applying." The rule applies to everything this skill produces, not just formal documents.
Mode 2 — WIP review table
Required column header row:
| Workstream / Jurisdiction | Budget (£) | Actual to date (£) | Budget consumed (%) | Est. % complete | Proportionality gap | ETC low (£) | ETC high (£) | Status |
Proportionality gap: Express as the difference between budget consumed % and estimated % complete. "+18pp" means spend is running 18 percentage points ahead of work completion. "−12pp" means spend is below progress — investigate. A gap of ±10pp is normal tolerance.
Status values — use exact labels, no abbreviation:
On track | Watch | Overrun — recoverable | Overrun — requires action | Underspend — investigate
For workstreams that are simultaneously underspent and a programme risk (e.g. budget consumed 20%, progress 0%, deadline approaching), use dual status: Underspend — investigate | Programme risk — action required. A status field that reports the financial position as healthy when the programme is broken is worse than no status field.
Variance commentary block — produce for every workstream outside tolerance:
[Workstream]: [X]% of budget consumed, [Y]% complete. Proportionality gap: [+/−Zpp].
Root cause: [Specific cause]
Pattern: [One-off / Systemic]
Recovery: [Recoverable within remaining budget / Requires fee adjustment / Scope signal — refer to scope-change-controller]
Action: [Specific next step with owner and date]
Mode 3 — Realisation alert memo
Produce as a .docx file. Mode 3 is a decision document — it belongs in the matter folder, not in the chat window.
Required structure:
-
Position — produce as a table with these rows: Agreed fee/cap | Recorded WIP | Cap over by | WIP as % of agreed fee | Self-reported completion | Projected total (burn rate method: WIP ÷ % complete) | Projected total (remaining work method: WIP + estimated remaining cost) | Proportionality gap | Current realisation (if billed at cap). Two projected total methods are required — they frequently produce different numbers. The range between them is the honest answer.
-
Completion scepticism — required on all AFA matters in Mode 3: On a matter in a loss position, there is a behavioural incentive to overstate completion percentage — a higher completion figure makes the trajectory appear less bad. Do not accept self-reported completion above 60% without producing this challenge explicitly: "Self-reported completion of [X]% has not been independently verified. If actual completion is [X−15]%, projected total rises to [£Y]. Recommend confirming remaining tasks with the team before the partner conversation."
-
Root cause — most likely specific explanation; rank by likelihood; frame the diagnostic question for the partner conversation
-
Options — three, with financial impact quantified: (a) Absorb — write-off amount and realisation at completion; (b) Recover — what OOS must be documented, estimated recovery; (c) Hybrid — what gets absorbed, what is recoverable, net realisation
-
Recommendation — which option and why; confirm root cause must be established before any client conversation
-
Decision required by — date and named partner; state what happens if decision is deferred (WIP continues accruing at current burn rate — quantify per week)
Mode 4 — AFA tracking table
Required column header row:
| Matter | Fee basis | Agreed fee/cap | Recorded WIP | Headroom remaining | Burn rate (per week) | Projected total | Position | Decision required? |
Headroom remaining = Agreed fee − Recorded WIP. This is the number that matters operationally.
Burn rate = Recorded WIP ÷ weeks elapsed. Use to calculate when headroom exhausts at current pace.
Projected total = Recorded WIP + (remaining work estimate). If projected total exceeds agreed fee, calculate the breach point as a date.
Position values: Within cap — monitoring | Cap risk — watch | Approaching breach — decision required | Cap breached — escalate immediately
Decision trigger: The point at which options must be exercised — not the breach point. If the cap exhausts in 3 weeks at current burn rate, the decision trigger is now.
Narrative block (produce alongside the table):
[Matter]: [Fee basis]. Agreed [fee/cap]: £[X]. Recorded WIP: £[Y] ([Z]% consumed).
Estimated completion: [A]%. Proportionality gap: [+/−Bpp].
Projected total at current burn rate: £[C] ([D]% of cap). Headroom: £[E].
[If breach projected]: Decision required by [date]. Options: [list].
Mode 5 — Fee adjustment
Two outputs:
- Internal memo — confirms OOS scope, quantifies financial impact against original budget, recommends adjustment amount, requests partner sign-off
- Client-facing letter (if required) — describes additional scope, explains why it was not included in the original estimate, states the fee adjustment, cross-references the scope change notice from scope-change-controller
Structured data export
Every Mode 1 and Mode 2 output is accompanied by a CSV export of the budget or WIP table. This is the input for SharePoint tracking and the starting point for the next WIP review. Produce inline as a labelled section if a file cannot be attached.
LPM vs Attorney Boundary
LPM: Phase-based estimating, contingency calculation, AFA structure design, WIP proportionality analysis, variance root cause assessment, realisation monitoring, write-off analysis, client financial disclosure sequencing, fee adjustment drafting.
Attorney: Billing judgment on individual time entries; whether a fee is professionally appropriate; recoverability of costs from a counterparty; professional rules on billing disclosure; legal minimum periods (consultation, notice, regulatory); whether a compressed programme is legally compliant.
Hard rule on legislation: Do not name specific statutes, regulations, or case law in skill outputs. If a delay or programme compression raises a legal compliance question (minimum consultation periods, regulatory filing windows, notice requirements), flag it as: "This timeline change may engage legal minimum period requirements — legal team to confirm compliance before programme is agreed with client." Do not characterise the legal risk, identify the relevant legislation, or draw a compliance conclusion. That is attorney work.
Cross-Skill Handoffs
- From matter-intake-scoping: The scoping brief and jurisdiction list are the primary input for Mode 1. Consume the matter brief output directly — do not start from a blank brief if one exists.
- From scope-change-controller: Confirmed OOS with a scope change notice reference triggers Mode 5. Do not re-assess whether the work is in scope — that is scope-change-controller's determination. Model the financial impact of the confirmed OOS only.
- AFA matters: At matter setup, flag to the partner that Mode 4 will be the ongoing monitoring mode. Budget is still built in Mode 1 (internal T&M estimate, regardless of the external fee basis). Mode 4 then tracks burn against the agreed fee at each review.
- To status-report-drafter: The WIP review table and FTC range are the financial inputs for the next status report. Pass with: "Updated financial position — consume for the financial summary section. Variance commentary below."
- To scope-change-controller: When Mode 2 variance analysis identifies a root cause that indicates a scope assumption was wrong (not an efficiency issue), flag as an OOS trigger. Pass with: "Variance in [workstream] appears scope-driven, not efficiency-driven — scope-change-controller to assess whether OOS documentation is required."
- To billing-cycle-manager: Confirmed FTC and write-off positions from Mode 2/3 feed into the billing cycle. Pass the WIP review output with confirmed positions for billing preparation.
- From risk-and-issues-manager: A breached financial assumption in the RAID log (e.g. "assumed 3 entities; confirmed 7") is a Mode 5 trigger once the assumption breach has been scoped and confirmed.
Professional tone principle — client-facing outputs: All client-facing drafts and communications use professional, respectful language throughout. Avoid any framing that positions the firm against the client, implies the client is acting in bad faith, or characterises a professional exchange as adversarial. Fee adjustment conversations are sensitive commercial discussions — the tone should be factual, collegial, and solution-oriented.
M365 Connected Mode (Optional)
Connected mode invocation rule: Search connected systems (Outlook, SharePoint, Teams) when doing so adds value — not as a default first step when sufficient input is already in the prompt.
- Sufficient input already provided: User has pasted WIP data, budget figures, or correspondence with full context. Engage with what is there. Do not search first — it adds friction without adding information.
- Input is incomplete or proactive surfacing is warranted: User references something that should be retrieved, or connected mode is running in background/scheduled mode. Search proactively — this is the inverted invocation model and is the highest-value connected mode behaviour.
The distinction is whether the user has already provided what is needed. If yes, work with it. If no, or if proactive surfacing serves the LPM, search.
When the M365 MCP connector is enabled (Claude Team/Enterprise), this skill can:
- Search Outlook for fee-related correspondence — partner budget discussions, client billing queries, LC fee cap exchanges — and surface these as WIP review triggers before the LPM has to ask
- Pull WIP export files from the matter's SharePoint folder directly, without copy-pasting figures
- Update a running budget tracker in SharePoint after each Mode 2 review, versioned by review date
- Search for fee adjustment precedents in the matter folder to inform Mode 5 letter drafting
Without the connector: paste WIP data directly or upload the spreadsheet. The skill operates fully in manual mode.
Time-Sensitive Assumptions
The following elements of this skill encode assumptions that may become stale:
- Rate benchmarks — all hourly rate references use approximate UK law firm benchmarks current as of 2025. Confirm prevailing rates with the firm's pricing team before use.
- Coordination uplift percentages — calibrated against general international LPM practice. Validate against firm historical data where available.
- Contingency ranges — based on general LPM methodology. Adjust for matter type, client relationship, and firm risk appetite.
- Realisation threshold (5%) — a general benchmark. Confirm the firm's agreed threshold at matter setup.
