CompTable
Compensation intelligence for hospitality operators and the CPAs serving them.
The only pre-payroll platform that models tip pools, service charges, and traditional wages side-by-side, with jurisdiction-correct compliance findings, IRS Form 8846 credit computation, and deterministic math a CPA can stand behind.
Built for two populations in parallel: restaurant and brewery owner-operators who need to know the cost and compliance consequence of every compensation decision before payroll runs — across the size range from single-location independents to larger multi-location groups — and the CPAs and bookkeepers who carry hospitality clients through tax season and Client Accounting Services expansions.
The Problem
Restaurant wage law is uniquely complex — and it changes every year.
No other small-business category sits at the intersection of federal wage law, every state's own tipped-minimum-wage rules, dozens of city ordinances, multiple legally distinct compensation models with separate tax treatment, and an IRS credit most operators never claim. Get any layer wrong and the cost shows up in penalties, back wages, or unreclaimed credits — sometimes years after the decision was made. Independent and small-multi-unit operators don't have an in-house labor-economics team to model it. CPAs serving them don't have the cross-jurisdiction reference data to scan a book of clients at a glance.
Three legal compensation models
Traditional tipped wages, mandatory service charges, and tip pools each carry distinct FLSA, state, IRS, and workers'-comp treatment. Switching models without understanding the trade-offs creates immediate compliance exposure and surprise tax bills.
Every state, dozens of localities
The federal tipped minimum is $2.13/hr. States set their own floors. Seven prohibit tip credit entirely. Roughly 50 cities override state law. Denver, Seattle, NYC, San Francisco, and Chicago each operate at fundamentally different cost structures than their state-default peers.
The FICA credit most operators miss
IRS Form 8846 reduces employer tax liability dollar-for-dollar on FICA paid on tipped wages — but is forfeited the moment an operator switches to a service-charge model. Most operators who switch never see the line item disappear from their books.
The Architecture
Deterministic engine. AI surfaces. A hard line between them.
Every dollar CompTable displays comes from a deterministic, pure-TypeScript calculation engine — not from a language model. The engine runs the same inputs to the same outputs every time, regression-tested at cent precision against the workbook our founder maintained as a restaurant operator for years before the platform existed. The engine is composable: each compensation model is a configuration of typed income streams with per-stream tax treatment, distribution, and eligibility. Custom scenarios are first-class.
AI handles the surfaces where natural language is genuinely useful — roster and menu ingestion, plain-English narration of Pro Forma comparisons, scenario suggestion from a conversational goal. The line between AI and the calculation engine is enforced at the type-system level so unvalidated language-model output cannot reach the math. The compiler refuses to build code that would let it. Trust is enforced, not promised.
Deterministic calculation engine
Pure functions, zero runtime dependencies, every dollar traceable to a single call site. Regression-tested against the original operator workbook on every commit. Conformance gates in CI block any change that drifts a known scenario by a cent.
AI as input and narration only
Roster ingestion, scenario suggestion, plain-English narration of engine output. AI never produces a dollar amount, a compliance finding, or a recommendation. Every AI call passes through a redaction allowlist before any model sees the data.
Type-system enforced boundary
Unvalidated language-model output cannot reach the calculation layer. The compiler enforces the contract. A no-advisory phrase scan blocks the platform from ever telling an operator what they should do — facts, math, and citations only.
Compensation Models
Five templates. Infinite custom configurations.
Each template encodes the FLSA, state, IRS, and workers'-comp treatment of a real compensation pattern operators run today. Operators load their roster, pick a template, and immediately see employer cost, per-employee take-home, Form 8846 credit, workers'-comp premium, and the full compliance findings list — at jurisdiction-correct numbers, with statute citations on every flag. Custom scenarios compose any combination of income streams, pools, distributions, and eligibility rules the engine supports. The engine is comp-model-agnostic by design — a values-driven brewery, a family-owned restaurant, and a chef-owner concept all get the same precision on whatever model fits their philosophy.
Traditional
Tipped minimum wage plus tip income. The federal-tip-credit baseline for any U.S. restaurant operating under FLSA defaults.
Full-House Service Charge
A mandatory service charge split across BOH, FOH, and benefits pools through a tiered point system with hours-weighted shift differentials.
BOH & Benefits Service Charge
Service-charge revenue flows to back-of-house staff and a benefits pool, allowing operators to model service-charge adoption without eliminating FOH tip income.
Full-House Tip Pool
No tip credit. Tipped FOH at full minimum wage with a 75/25 FOH/BOH split distributed by hours-weighted tiers — the modern compliant tip-pool pattern.
Tip Credit Elimination
Compliance model for California AB 1228 and similar state-law transitions. Correctly prices the move from tip-credit to full-minimum-wage compensation.
Custom Scenarios
A parameter-driven editor builds any combination of income streams, pools, distributions, and eligibility rules — supporting multi-unit operators with non-standard pay structures.
Local Pulse
The dataset competitors cannot replicate.
Every active operator workspace contributes anonymized, partitioned snapshots into a shared peer-benchmark engine. Local Pulse lets an operator see how their payroll percentage, FOH/BOH disparity, benefits load, retention break-even, and full compensation profile compare to other operators like them — and lets a CPA see the same view across their entire book of restaurant clients.
The architecture is anonymity-first. K-anonymity gating at the read layer (no result returns unless a minimum cohort of comparable peers exists). A four-tier geographic expansion ladder — locality, county, state, federal — applied automatically until the anonymity threshold is met, with the expansion level disclosed in the UI. Partitioning by revenue band, headcount band, service type, and compensation model. No foreign key from any anonymized snapshot back to its source workspace — anonymity is a schema property.
This is the studio's most compounding data moat. The benchmark is only valuable with cohort breadth, and the cohort only achieves breadth with active, diverse adoption — from the single-location independent to the multi-location group, across every operator philosophy. The dataset accrues in our favor every day the product is in use.
Jurisdiction Data
1,397 effective-dated rate rows. Maintained by the team.
Every rate the platform displays carries a provenance badge: Maintained (authoritative, kept current by the CompTable team), Default — editable (a reference value the operator can override for their own situation), or Your Input. Hover any value to see the source and last-updated date. There are no hidden numbers and no "trust me" figures.
Rate rows are effective-dated. The platform can reproduce any historical pay period's compliance state and correctly prices rate-change crossings mid-year — a non-trivial property when a state minimum wage steps up on July 1 and an operator needs to backfill an audit. A database-level exclusion constraint prevents overlapping rate ranges, so the data is always queryable at a single timestamp.
Sources include the DOL state minimum-wage tracker, each state's labor department, the UC Berkeley Labor Center local-ordinance inventory, NCCI loss costs, and SSA wage base data. Every rate passes through a scraper plus a human review step before promotion to authoritative.
- ·FICA (6.2% SS / 1.45% Medicare)
- ·IRS Form 8846 FICA tip credit
- ·FLSA tipped minimum wage
- ·Federal income tax brackets (W-4)
- ·State minimum wage
- ·State tipped minimum wage
- ·Tip-credit rules and limits
- ·State income tax (CO, CA, NY shipped, more in progress)
- ·State UI taxable wage base + experience rate
- ·State PFML programs across the eleven active-program states
- ·Denver, Boulder, Edgewater
- ·Seattle, Tukwila, Renton, Everett, Burien, SeaTac, Bellingham, Kings County
- ·Portland Metro tiers
- ·NYC, Nassau, Suffolk, Westchester
- ·San Francisco Bay Area and Los Angeles cities
- ·Chicago; Minneapolis and Saint Paul tiers
- ·Flagstaff; Tucson
- ·Predictive-scheduling rules across six jurisdictions
- ·NCCI Class 9083 (Restaurants-Catering)
- ·44 NCCI states + 6 independent bureaus
- ·X-mod adjustable per operator
AI Scope
AI handles the surfaces. The engine handles the math.
CompTable's AI surfaces fall into three categories, all of them server-side and contracted on a zero-retention basis. Ingestion converts unstructured input — a pasted CSV, a photographed roster, a menu in PDF form — into validated structured records the engine can run on. Narration takes engine output and translates it into plain English for an operator who doesn't read spreadsheets. The proposer turns a conversational goal into a set of scenario configurations the operator can run against the deterministic engine. None of these surfaces produces a dollar amount, a compliance finding, or a recommendation. Every model output is gated at the API boundary by a structured-output schema and routed through a redaction allowlist that projects only safe fields. That separation is what makes the platform CPA-defensible.
Ingestion
Roster and menu data captured from CSV, paste, or image. AI extracts; the operator confirms; the engine runs.
Narration
Pro Forma comparisons, retention projections, menu repricing analyses, OT exposure, and wage compression — all translated to plain English alongside the numbers, never replacing them.
Proposer
Conversational goal extraction produces candidate scenarios the operator can run and compare. AI proposes; the engine evaluates; the operator decides.
Who It's For
Two channels. One platform.
CompTable is purpose-built for the operator-CPA relationship. Operators run scenarios against their actual roster. CPAs run scans across an entire book of restaurant clients. Workspaces can be owned by either party and handed across the bidirectional ownership boundary — the CPA can spin up an operator's workspace and transfer it, or the operator can invite the CPA into one they already run.
Hospitality Operators
Load your roster. The platform resolves your jurisdiction automatically. Run your current compensation model alongside every alternative and see employer cost delta, Form 8846 credit impact, per-employee take-home change, and workers'-comp premium — all on the same screen, at jurisdiction-correct numbers.
Built for the operator who doesn't have an in-house finance or HR analytics team — from the single-location independent restaurant or brewery to the larger multi-location group. The engine is agnostic to compensation philosophy. A values-driven brewery, a family-owned restaurant, a chef-owner concept, and a multi-unit group all get the same precision against whatever model fits their business.
The retention simulator quantifies whether switching models is worth the transition cost. The menu pricing engine shows what price changes offset a labor cost increase. The headroom solver finds the levers that close the gap without touching menu prices. Native shift scheduling catches OT compliance issues before they hit payroll.
Compliance findings cite the statute, link to the rule, and compute the dollar impact. The platform never recommends a model — it surfaces facts, math, and the citations behind them. Operators draw their own conclusions.
CPAs & Bookkeepers
One firm account manages every restaurant client under a single login. Each client's jurisdiction rates auto-update. Run a compliance scan across the whole book to flag exposure before any operator gets audited. Export PDF-quality Pro Forma reports to share at tax season.
For firms expanding into Client Accounting Services, CompTable is the wedge into hospitality. Firm-tier pricing bundles the platform into the CPA's service offering, deepening wallet share without adding a line item the operator has to evaluate separately. The bidirectional ownership model means the CPA can either provision the workspace for the operator or be invited in as a collaborator.
For partners and funders: this is our distribution wedge. CPA firms serving restaurants are the channel; CompTable is the leading hook; the productized double-entry ledger across the rest of the portfolio deepens the relationship after the wedge lands.
Built for the operators and CPAs who need to know their numbers.
Jurisdiction-correct wage modeling for every U.S. state, territory, and major locality. Facts and math, with statute citations. No advice.