TBF Insure
Working prototypeBuilt by NexAI Advisors · peer data is synthetic

Coverage Benchmark

Describe a business in one sentence. The system classifies it, reads its current insurance program, and shows where every coverage line sits against a peer cohort — with the evidence behind each finding on the page.

18 coverage lines, including 6 yacht and marine lines that general commercial benchmarking tools do not carry.

Start a benchmarkHow it works, and what the data is

The flow

Three steps, each of which shows its work

  1. 01

    Classify the business

    A plain-English description matched against the Census NAICS index. Returns a ranked list with the index terms that fired and a confidence you can act on, not a single unexplained code.

  2. 02

    Read the current program

    Key the lines in, or upload a declarations page. Extracted values carry per-field confidence and the source text they came from; anything the parser will not assert goes to a review queue.

  3. 03

    Benchmark against peers

    Every line placed against a cohort matched on NAICS, revenue band and state. Limits, premiums and adoption rates as quartiles, with gaps ranked by severity and tied to the statistic behind them.

Worked examples

Five accounts, already loaded

Each opens straight on its finished report. They were chosen because each one breaks something a naive benchmark would get wrong.

  • Dock & boat-lift contractor, Fort Lauderdale

    237990

    Eighteen employees on the payroll and no workers compensation. The engine should call that first and loudest, ahead of the equipment schedule it is also missing - and the cohort here is deep enough that it never has to widen.

    Revenue
    $3.2M
    Band
    $1M - $5M
    Lines benchmarked
    12
    Critical gaps
    2
    Open this report

  • Management consultancy, mid-market

    541611

    A well-run $12M firm with no EPLI and no cyber. Its exact NAICS class is thin in the corpus, so this is the scenario where the cohort widens and has to explain itself in plain English rather than quietly benchmarking against five policies.

    Revenue
    $12M
    Band
    $5M - $25M
    Lines benchmarked
    12
    Critical gaps
    1
    Open this report

  • 80ft motor yacht on charter, Fort Lauderdale

    487210

    Hull, P&I, crew and tenders all in place and proportionate, including a $50,000 hull deductible that is normal for the value insured and correctly left alone. The vessel is chartered and has no charter liability: most pleasure forms exclude charter outright, so that is the one critical hole in an otherwise tidy program.

    Revenue
    $850K
    Band
    $1M - $5M
    Lines benchmarked
    6
    Critical gaps
    1
    Open this report

  • Full-service restaurant group, Broward County

    722511

    Written on a business owners policy, so the packaged general liability and property must not be reported as missing - and they are not. What does get flagged is the retention: $25,000 against a $1M package, twice the share of insured value its peers keep.

    Revenue
    $2.4M
    Band
    $1M - $5M
    Lines benchmarked
    12
    Critical gaps
    0
    Open this report

  • Structural steel fabricator, $28M

    332312

    The control case. A properly built mid-market manufacturing program, bought at or above peer median across the board - the engine should find nothing critical or material here, which is the only way to trust it when it does.

    Revenue
    $28M
    Band
    $25M - $100M
    Lines benchmarked
    12
    Critical gaps
    0
    Open this report

Read this before you read the numbers

What is real here, and what is not

Real

  • The NAICS reference data: 1,012 codes and 20,373 index terms across 20 sectors, from the US Census Bureau, 2022 NAICS index file.
  • The classification, extraction and benchmarking logic, end to end.
  • The coverage-line taxonomy, including the marine lines.

Synthetic

  • Peer figures come from a synthetic corpus of 11,000 modelled policies. They are not market data: no carrier, filing, bordereau or real policy is represented here. The corpus is generated by a documented model (seed 20260826) that varies adoption, limits, premiums and deductibles by NAICS sector, revenue band and state, weighted toward Florida and the South Atlantic. It is built to be internally coherent and byte-for-byte reproducible so the analysis behaves the way it would on real data - but every peer statistic on this page is modelled. Replace this corpus with bordereau or rating-bureau data before quoting any figure to a client.
  • Corpus peer-corpus-2026.08-r1, 11,000 generated policies, seed 20260826.

The interesting question is not whether these distributions are right — they are not meant to be. It is whether the method holds once a real book of business is behind it. The methodology page sets out what that swap involves.