Platform

Plain English explains every CBP rejection.

Built on gpt-4o. Reads your filing, your party data, and the CBP response. Streams numbered fix steps. Gated behind your team's enable flag — toggle off anytime.

The AI features

One model. Three surfaces.

  • Today's brief

    One sentence at the top of your day. Auto-generated when you open the Compliance Center. Capped at 140 characters so you can read it without scrolling.

    With AI enabled
    "Three drafts waiting on you. Run an AI pre-flight before submitting — one rejection blocking re-file."
    Rule-based fallback (AI off)
    "5 cards in your queue. 1 needs attention before EOD."
    ≤140 charsRefreshes each login

  • AI Coach — rejection mode

    When CBP rejects, the Coach reads the error code, your filing, and party data — then streams numbered fix steps.

  • AI Coach — pre-flight mode

    Reviews the entire draft for UFLPA risks, PGA flags, HTS issues, and rule-based gates before you submit.

HTS Classifier

Describe goods, get the right HTS code.

polyester athletic socks, women's↵
  • 6115.96.6010Stockings, socks — Knit — Of cottonBest match
    Best match — polyester is a synthetic fiber and athletic socks are typically knit.
  • 6115.95.6020Stockings, socks — Synthetic fibersAlternative
    Alternative — but 'cotton' subheading often used for blended athletic socks.
  • 6115.99.1410Stockings, socks — Other textile materialsAlternative
    Fallback for blends with no dominant material.

Sources: USITC HTS revision 14, 2026. Last updated 2026-04-12.

Under the hood

How the AI actually works.

  • The model

    gpt-4o via OpenAI. We send only the filing data plus the CBP response. We do not send your AR data, banking, billing, or other unrelated tenant data.

  • The gate

    Gated by your team's enable flag — admin can toggle AI off per-team or per-feature. Rule-based fallback is always available.

  • Your data

    Your data is not used for model training. OpenAI's API tier we use has zero-retention by contract.

FAQ

Common questions.

  • Is the AI required to use MyCargoLens?

    No. AI is opt-in per team. The rule-based gate runs on every filing regardless.

  • How long does AI Coach take to respond?

    First token typically 1.5–3 seconds. Full numbered fix steps in 6–12 seconds. Streams as it generates.

  • What if the AI gets it wrong?

    You're the final reviewer. Every fix step is suggested, not auto-applied. You always click submit.

  • Does the AI know my templates?

    Yes — pre-flight reads your saved templates to suggest field reuse and flag inconsistencies.

  • Can I disable AI for sensitive filings?

    Yes. Mark any draft as 'AI off' and it bypasses Coach + pre-flight regardless of team setting.

Want AI on your next filing?

Start freeSee pricing