You Cannot Audit What Your Brokers Filed: Reconciling ACE Entry Data Against Your Own Records
What changed: The retired AI Integration and Automation services, the free audit and the Carborundum wording are replaced by what GingerControl offers now (Classification Researcher, Tariff Calculator, the paid compliance audit, GingerOS); ACE data comes from your own exports.
You pull the line-level entry data your brokers transmitted to ACE (the HTS code, entered value, and duty on each CBP Form 7501) and match it field by field against your internal classification and valuation master, so every difference surfaces as a flagged exception instead of a surprise. GingerControl's compliance audit, a paid service quoted for its scope, starts that work by showing where your entities, brokers and records disagree and what to fix first, turning broker oversight from a trust exercise into a verifiable, repeatable audit.
Next step:Check every deadline on one entry
In this post
- Why can't trade compliance teams just trust their broker's spreadsheet?
- TL;DR
- The pain: you sign for entries you have never seen the data on
- Why it hurts: cost, audit risk, and decisions you cannot make
- What the data actually is: the 7501 and the ACE reports you can pull
- How reconciliation actually works, field by field
- Manual reconciliation versus an ingested data layer
- Where this fits in your reasonable-care program
- Frequently asked questions
- Pull broker and ACE entry data into one reconcilable layer
- References
Why can't trade compliance teams just trust their broker's spreadsheet?
Because the broker's spreadsheet is a summary of what they say they filed, not the line-level record of what ACE actually received, and the importer of record, not the broker, carries the legal liability for every code and value on the 7501. Until you reconcile the filed data against your own master, you are signing for entries you have never independently checked.
TL;DR
Broker entry data reconciliation in ACE is the process of pulling the line-level data your customs brokers transmitted (HTS, entered value, and duty on each entry summary) and matching it against your own classification and valuation records, so discrepancies surface as flagged exceptions before CBP finds them. For a global trade compliance team overseeing 3 to 8 brokers and tens of thousands of entry lines a year, the problem is structural: the broker holds the filing data, the team holds the product master, and the two never meet until a CF-28 Request for Information arrives, typically with a 30-day clock. The fix is not a bigger spreadsheet. It is ingesting broker and ACE entry data into one reconcilable layer and comparing filed codes, values, and duty against your internal master at scale. GingerControl builds GingerOS, the AI trade compliance platform that runs on your process and the data you have, ERP or not, and reads the ACE exports, broker files and spreadsheets you already have; its compliance audit, a paid service, shows where your entities, brokers and records disagree, so you are not left to re-key 7501s into Excel one broker at a time.
This article is for the person who owns broker oversight and has quietly accepted that they cannot actually see what their brokers filed. I co-founded GingerControl and have spent the last few years building the data and AI layer that compliance teams use to ingest, classify, and reconcile trade data, so this is written from the build side of the problem, not as a sales sheet. The pain below is real, it is structural, and most teams have normalized it. The point of naming it is to show that it is fixable.
The pain: you sign for entries you have never seen the data on
Here is the position most global trade compliance teams are actually in. You have a customs broker, or three, or eight. They file your entries. Each entry produces a CBP Form 7501 entry summary that declares, line by line, the HTS classification, the entered value, the country of origin, and the duty and fees owed. The importer of record, which is you, is "liable for payment of all duties and meeting all statutory and regulatory requirements" under 19 C.F.R. 141.1(b). You carry the liability. The broker presses the button.
And here is the gap: you almost never see the line-level data the broker actually transmitted. You see an invoice for duty. You see a monthly summary the broker emails you. You see, maybe, a PDF of the 7501 if you ask for a specific entry. What you do not have is a clean, structured, all-entries-all-brokers dataset that you can lay next to your own classification master and compare. So when a colleague asks "what did we declare this part number as, on every entry, across every broker, last year?", the honest answer is: nobody knows without a manual hunt.
The reasons this happens are not anyone's fault. They are structural:
- The data lives in the broker's system, not yours. Over 99 percent of entry summaries are transmitted to ACE through the Automated Broker Interface (ABI) via electronic data interchange, per CBP's ACE guidance. The broker's software generates and sends the filing. Your visibility depends on whatever extract the broker chooses to share.
- Entry summaries cannot be filed through the ACE Portal. As CBP states, "entry summaries may only be filed via EDI, not via the Portal." The portal is for review and reporting, not filing, so the importer's window into its own entries is a reporting window, downstream of the broker's transmission.
- Each broker formats data differently. Three brokers means three spreadsheet layouts, three sets of column names, three definitions of "value." Reconciling them by hand is a data-cleaning project before it is a compliance project.
- The product master and the filing data never touch. Your classification decisions live in an item master, a PLM system, or a classification database. The filed codes live in ABI. Nothing automatically compares the two.
Quotable insight: The structural flaw in broker oversight is that the party with the legal liability (the importer of record) is not the party that holds the filing data (the broker), and the two datasets never meet until CBP forces a comparison. A 7501 declares the HTS, entered value, and duty on every line, yet most trade compliance teams cannot lay all their filed entries next to their own classification master without a manual, broker-by-broker hunt. Reconciliation closes that loop before a CF-28 does it for you.
Why it hurts: cost, audit risk, and decisions you cannot make
The trust-based status quo is not free. It has three concrete costs.
1. Duty leakage you never catch. If a broker classifies a part under a slightly wrong HTS code, you may be overpaying duty (money walking out the door every shipment) or underpaying it (a liability accruing silently until liquidation or audit). Without reconciliation, neither shows up. You find out the expensive way.
2. Audit exposure on a 30-day clock. Before an entry liquidates, CBP can issue a CF-28 Request for Information to verify that goods were "classified, cleared, and declared properly." The CF-28 typically requires a response within 30 days from the date printed on the form, and CBP frequently questions the declared value or whether the HTS code matches the product. If you have never reconciled what was filed, the CF-28 is the first time you learn what your broker actually declared, and now you are reconstructing it under deadline. Worse, CBP "may view non-response as evidence of inadequate compliance systems," which influences future audit selection.
3. Decisions you cannot make. Sourcing, tariff-engineering, and FTA decisions all assume you know your own baseline. If you cannot say with confidence what was classified as what, at what value, on which entries, you cannot quantify duty exposure by product or entity, you cannot prove an FTA claim was filed correctly, and you cannot give finance a defensible number. The fragmentation stalls the strategic work the team exists to do.
The reasonable-care standard makes this sharper. Under 19 U.S.C. 1484, the importer of record must use reasonable care in classifying and valuing merchandise. Reasonable care is hard to demonstrate for data you have never reviewed. Reconciliation is, in practice, the mechanism that turns "we trust our broker" into "we verify our broker," which is the posture CBP expects.
What the data actually is: the 7501 and the ACE reports you can pull
The encouraging part is that the data exists and you have a right to it. The importer of record controls access to its own ACE Portal account, and ACE exposes reports that give you line-level entry data for in-house audit. The raw material for reconciliation is already there; the obstacle is assembling and comparing it at scale.
The two anchors are the entry summary form and the ACE reports.
| Data source | What it contains | Who controls it | Use in reconciliation |
|---|---|---|---|
| CBP Form 7501 (Entry Summary) | Line-level HTS code, entered value, country of origin, duty and fees, importer of record, entry number | Filed by the broker via ABI/EDI | The system of record for what was declared per line |
| ES-003 Entry Summary Line Tariff Details (ACE report) | Every formal entry line: HTS codes assessed, duties paid, liquidation status, and tariff-type breakdown (IEEPA 9903.01/02, Section 301 9903.88, Section 232, Section 122) | Importer of record (ACE Portal account) | The importer-side line-level extract to reconcile against the internal master |
| Liquidation report (ACE) | Whether entries liquidated as filed or with rate changes | Importer of record (ACE Portal) | Catches CBP rate changes and confirms the final duty position |
| Rejected Entry report (ACE) | Entries CBP rejected for data errors | Importer of record (ACE Portal) | Surfaces filing failures the broker may not have flagged |
| Internal classification and valuation master | Your governed HTS code, declared-value basis, and origin per part number | The trade compliance team | The reference standard the filed data is reconciled against |
The ES-003 Entry Summary Line Tariff Details report is the workhorse. It lists every formal entry line you filed, the duties paid, the HTS codes assessed, and the liquidation status, and it breaks each line out by tariff type. CBP and the trade documented it widely in 2026 as the report importers use to identify IEEPA tariffs paid and to verify CAPE processing, but its value is broader: it is the importer-side, line-level mirror of what the broker transmitted. That is exactly the dataset you reconcile against your own records.
Turning these scattered reports and broker extracts into one comparable layer is the engineering work, and it starts from exports: the ACE reports come from the importer's own ACE Portal account, and the broker extracts from each broker.
If a broker sends client files directly to an outside processor, confirm client-record sharing permissions before adding that feed.
How reconciliation actually works, field by field
Reconciliation is not "look at the broker's spreadsheet harder." It is a structured, three-input comparison: the broker's filed data, the importer's ACE extract, and the internal master, matched on a common key (usually part number plus entry line) so that every field either ties out or flags.
A practical reconciliation map:
| Field on the 7501 | Reconcile against | What a mismatch means |
|---|---|---|
| HTS classification (10-digit) | Your governed code in the item master | Filed code differs from your decision; possible over/underpayment and audit risk |
| Entered value | Your declared-value basis (transaction value plus statutory additions) | Value drift; valuation exposure CBP probes via CF-28 |
| Country of origin | Your origin determination | Origin mismatch; wrong duty rate, broken FTA claim, or 301/232 scope error |
| Duty and fees | Recomputed full tariff stack for the filed code and origin | Math error or missed Chapter 99, 301, 232, or 122 layer |
| Chapter 99 / special tariff lines | Your exposure model by tariff type | Missing or extra special tariff; refund or liability signal |
| Liquidation status | Your open-entry tracker | Rate changes or approaching deadlines you did not know about |
The pattern that breaks naive reconciliation is the one-to-many line: a split-code product (a Chapter 91 watch, for example) can legitimately decompose into multiple component-level HTS codes, so a single part number may map to several filed lines. A reconciliation that assumes one part equals one code will mis-flag every split-code entry. The comparison logic has to handle one-to-many from the start.
GingerControl's HTS Classification Researcher is built to produce the internal master side of this comparison the right way. It follows the same reasoning a licensed customs broker uses (GRI analysis, Section and Chapter Note review, and CROSS ruling research), autonomously detects when a composite product needs a GRI 3(b) essential-character judgement, and records which component gives it its essential character. For split-code goods dutied part by part, such as Chapter 91 watches, GingerControl's Tariff Calculator lists each component and its duty. That gives you a defensible, audit-ready reference code to reconcile the filed code against, instead of reconciling one unverified number against another.
GingerControl's classifications are for general reference, educational, and planning purposes; before a code goes on an entry, the customer's licensed customs broker reviews it. GingerControl does not provide legal advice, file entries or replace licensed customs expertise. Reconciliation outputs are research for the importer or their licensed broker and counsel to review and act on, not a determination filed on their behalf.
Manual reconciliation versus an ingested data layer
Most teams attempt reconciliation in a spreadsheet, periodically, on a sample of entries, when something prompts it. The structural alternative is to ingest broker and ACE entry data into one reconcilable layer and run the comparison continuously and completely. Here is the honest comparison.
| Approach | Ingests broker plus ACE entry data into one dataset | Field-level match of filed HTS, value, duty vs internal master | Reference code with GRI reasoning | Handles split-code one-to-many entries | Anomaly detection across all entries | Audit trail for CF-28 response | Coverage |
|---|---|---|---|---|---|---|---|
| A reconciliation layer built on your own data | Yes, from ACE exports and broker files | Yes | Yes, if the master holds reasoned codes (for example from GingerControl's Classification Researcher) | Yes, with one-to-many logic | Yes | Yes | All entries, all brokers |
| Manual spreadsheet reconciliation | No, manual re-keying per broker | Sample only, by hand | No, depends on whatever is in the master | Rarely, usually mis-flagged | No | Manual, scattered | Whatever you have time for |
| Trusting the broker summary | No | No | No | No | No | No | None |
| Generic BI dashboard on ACE exports | Partial, you build and maintain the pipeline | Only if you code the comparison rules | No | Only with custom logic | Threshold rules only | Reporting only | All entries, no classification logic |
Bottom line: For a global trade compliance team overseeing multiple brokers and tens of thousands of entry lines a year, the deciding factor is not dashboard polish, it is whether the filed HTS, value, and duty are actually matched against a defensible internal master, including split-code entries, across every broker. GingerControl's Classification Researcher supplies GRI-reasoned reference codes for that master, and its compliance audit shows where your entities, brokers and records disagree. A generic BI dashboard is best suited to teams that already have a clean, governed classification master and only need to visualize ACE exports.
Where this fits in your reasonable-care program
Reconciliation is not a one-time cleanup. It is the feedback loop that keeps broker oversight honest. The mature version runs on a cadence: ingest the period's entries from ACE and from each broker, reconcile against the governed master, route every exception to a human for disposition, document the resolution, and feed systematic errors back into broker SLAs and the classification master. Done this way, the CF-28 stops being a fire drill, because the data is already reconciled and the audit trail already exists.
This piece is the problem diagnosis. If you want the architecture of an automated classify-screen-compute pipeline that produces the governed internal master in the first place, that is a separate design decision covered in the GingerControl write-up on customs compliance workflow orchestration, and classification at enterprise volume through one call is covered on the Classification + Tariff Calculation API page. Where the reconciliation you need is not against your classification master but against your accounts-payable records, the parallel discipline of matching the PO, invoice, and entry catches the overpaid duty and undervaluation exposure that live between what you paid and what your broker declared. Here, the point is narrower and prior to both: you cannot govern, audit, or improve what you cannot see, and right now you cannot see what your brokers filed.
Frequently asked questions
How do you reconcile what customs brokers filed in ACE against your own records?
You pull the line-level entry data the broker transmitted (the HTS, entered value, and duty on each 7501) using ACE reports such as ES-003, then match each field against your internal classification and valuation master so discrepancies flag automatically. Run as a standing process on one comparable dataset, it covers every entry instead of a manual sample; GingerControl's compliance audit, a paid service quoted for its scope, is a starting point that shows where your entities, brokers and records disagree.
Why can't I just rely on the monthly summary my broker sends?
A broker summary reports what the broker says it filed, not the line-level data ACE actually received, and the importer of record carries the liability for every code and value regardless. For a trade compliance team that owns broker oversight, the gap between the summary and the filed 7501 is exactly where duty leakage and audit risk hide. Reconciling the filed entry data against your own master is what turns the summary from assumed to verified.
What ACE report shows the line-level data I need to reconcile?
The ES-003 Entry Summary Line Tariff Details report in the ACE Portal lists every formal entry line you filed, the HTS codes assessed, duties paid, liquidation status, and a tariff-type breakdown (IEEPA, Section 301, Section 232, Section 122). For an importer of record running in-house audits across all brokers, ES-003 is the importer-side mirror of what was transmitted. Export it from the ACE Portal and line it up with each broker's extract on part number and entry line.
The CF-28 audit trends and response preparation guide explains why classification and origin records need to be ready for an information request.
How does reconciliation help me respond to a CF-28 Request for Information?
A CF-28 typically gives you 30 days from the date on the form to prove your HTS and value were declared correctly, and if you have never reconciled, you are reconstructing the filing under deadline. A standing reconciliation means the filed data is already matched to your master and the supporting reasoning already exists. GingerControl's HTS Classification Researcher produces the GRI reasoning for the internal reference code, with the Section and Chapter Notes and CROSS rulings behind it, the documentation a CF-28 response needs.
Does GingerControl file entries or act as my customs broker?
No. GingerControl is an AI trade compliance platform; it does not file entries, act as importer of record or do customs business, and it works alongside your broker. Its classifications are for general reference, educational, and planning purposes; before a code goes on an entry, your licensed customs broker reviews it. For a GTC team, GingerControl's output is the research foundation your broker and counsel review and act on.
How does GingerControl handle split-code products in reconciliation?
A split-code product can legitimately map to several component-level HTS lines, so a one-part-equals-one-code reconciliation mis-flags every split entry. For goods dutied part by part, such as Chapter 91 watches, GingerControl's Tariff Calculator lists each component with its own duty, so the reference side of the comparison models the one-to-many relationship correctly. For teams importing electronics, machinery, or other composite goods, this prevents reconciliation from generating false discrepancies on legitimate split-code filings.
Can reconciliation quantify whether I am overpaying or underpaying duty?
Yes. By recomputing every duty layer for the filed code, origin and entry date (MFN plus Section 301, Section 232 and other Chapter 99 lines, with Section 122 on entries from February 24 to July 23, 2026) and comparing it to what was actually paid, reconciliation surfaces both overpayments to recover and underpayments to correct. GingerControl's Tariff Calculator recomputes every layer by entry date, so a trade compliance team can convert a pile of 7501 lines into a quantified duty position by product, entity, and tariff type.
Pull broker and ACE entry data into one reconcilable layer
You cannot audit what you cannot see, and right now the line-level data your brokers transmitted to ACE lives in their systems, not yours. Start with what you can pull today: the ES-003 report from your ACE Portal account and each broker's extract, matched against your internal master so every discrepancy flags before a CF-28 does. GingerControl's compliance audit, a paid service quoted for its scope, shows where your entities, brokers and records disagree and what to fix first, and the HTS Classification Researcher gives you a GRI-reasoned reference code to reconcile against. Reconciliation moves broker oversight from trust to verification.
GingerControl builds GingerOS, the AI trade compliance platform that runs on your process and the data you have, ERP or not. A project starts with Prototype Week, the step before a quote. Talk to our team →
References
- U.S. Customs and Border Protection, CBP Form 7501 (Entry Summary)
Data cited: Entry summary purpose (appraisement, classification, origin), line-level fields (HTS, entered value, duty), importer of record liability under 19 C.F.R. 141.1(b).
Source: CBP Form 7501: Entry Summary
Published: U.S. Customs and Border Protection (accessed June 2026) - U.S. Customs and Border Protection, How to Use the Automated Commercial Environment (ACE)
Data cited: Over 99 percent of entry summaries transmitted via ABI/EDI; entry summaries filed via EDI not the Portal; importer-of-record ACE Portal access; ES-003 and in-house audit reports.
Source: How to Use ACE
Published: U.S. Customs and Border Protection (accessed June 2026) - U.S. Customs and Border Protection, CBP Form 28 (Request for Information)
Data cited: CF-28 pre-liquidation review of classification, value, and declaration; response timeline expectations.
Source: CBP Form 28, Request for Information
Published: U.S. Customs and Border Protection - CBP Ruling HQ H290535 and HQ H350722
Data cited: Classifying specific goods beyond six digits for importation, and AI-assisted classification together with Form 5106 importer registration, constitute customs business requiring a licensed customs broker.
Source: CBP Customs Rulings Online Search System (CROSS)
Published: CBP (HQ H350722, Jan 16, 2026) - U.S. Code, 19 U.S.C. 1484 (Entry of merchandise) and 19 C.F.R. 141.1
Data cited: Importer-of-record reasonable-care obligation in classification and valuation; liability for duties and compliance.
Source: 19 U.S.C. 1484, Legal Information Institute
Published: U.S. Government Publishing Office
Tracking liquidation and protest dates in a spreadsheet?
Hand this to GingerOS.
Pick a 30-minute slot on the next page with Chen Cui, our CEO. Bring one notice or one SKU list (an ERP export, a spreadsheet or an email will do) and see what GingerOS does with it.
Reviewed by Michael Weick, LCB / CCS




