Washington Attorney General Consumer Complaints

Every consumer complaint filed with the Washington Attorney General's office is public record - and the office publishes the whole file as open data. This actor turns it into a charged-per-record lookup, filter and aggregate tool: pull the complaints filed against a business, sweep an industry, isolate the live cases, or range over the dollars the AG recovered.

Built for: vendors and partners screening a company's complaint history, consumer-protection and compliance teams, market researchers reading an industry's grievance mix, and anyone who needs the complaint record behind a business - not a review-site score.

What it covers

281,333 complaints filed against Washington businesses - 20,110 so far in 2026, running roughly 23-26k a year in recent years (16,005 back in 2019). Each row is one complaint:

• the business it was filed against (216,825 rows carry a name), with street, city, state and ZIP
• the industry category - Electronic Shopping (18,562), Auto Sales (17,443), Telecommunications (15,846), Retail Sales (15,423), Health Care (12,734), Broadband Providers (11,979), Contractors (10,982), and more
• the NAICS code and name (281,245 rows carry a NAICS), split from the source's compound text into a clean code and label
• the status - Closed (276,961), New (2,585), Open (1,787)
• the dollars recovered for the consumer - estimated and actual savings, present on 44,532 complaints
• the opened date/year, and the business's geocode (latitude/longitude) where the address resolves

The fine print that matters

Three source quirks are handled for you. The business address fields arrive whitespace-padded (businessZip = " ", businessState = " "), so every value is trimmed - a blank is a real blank, and a ZIP filter matches the digits you'd expect. The naics column is a compound string ("522291-Consumer Lending (includes Payday Lenders)") - the actor splits it into naicsCode and naicsName, so a naics=522291 filter returns Consumer Lending and nothing else. And businessId is the source's own 6-digit id - not a WA UBI (the UBI is 9 digits); it is surfaced as an opaque id and never claimed as a universal business key.

The recovered-dollars angle is the one no directory has: the AG reports what it actually recovered for consumers, and 44,532 complaints carry an amount - filter hasSavings or a minimum to find the cases with money behind them.

Typical questions

• "Every complaint filed against Comcast."
• "Auto Sales complaints over the last year."
• "NAICS 522291 (payday lending) complaints in WA."
• "Complaints where the AG recovered more than $5,000."
• "Aggregate complaints by category, NAICS or year."

Inputs

Input	What it does
mode	rows (default) / aggregate
status / business / category / naics	who and what
city / state / zipcode	where the business is
openedFrom / openedTo	opened-date range
minSavings / minActualSavings / hasSavings	the recovered-dollars angle
groupBy	aggregate over category / status / NAICS / state / year
maxResults	cap records (default 50)

Default run = the 50 most recently opened complaints - fast for the daily auto-test. For a targeted query add a filter; for a broad view use aggregate.

Example inputs

One company's complaint history - business is a case-insensitive substring.

{ "business": "Amazon.com", "maxResults": 5 }

The live cases - New and Open are the complaints still moving.

{ "status": "Open", "maxResults": 10 }

An industry over a date window - category is a substring, openedFrom an ISO date.

{ "category": "Auto Sales", "openedFrom": "2025-01-01", "maxResults": 10 }

Complaints with money recovered - hasSavings keeps only the rows where the AG recovered an amount.

{ "hasSavings": true, "minSavings": 5000, "maxResults": 10 }

Low cost

From $0.0001 per record, down to $0.00006 at Gold - billed only for the rows you use, at the low end of the store. Cost scales with what you pull, not with the size of the register, and each record is metered individually.

The traps are in the cleaning. The address fields are padded with runs of spaces, so a naive zip = '98004' match returns nothing and a filter that forgets to trim silently drops every row; the industry is stored as a compound code-and-name string that has to be split before a NAICS filter means anything; and the business_id looks like a business key but is not the state's UBI, so treating it as one silently mis-joins two different companies. Normalizing the padded address, the compound NAICS and the mis-labeled id into a schema where a business / category / zipcode query returns exactly the complaints you mean is the actual product. Every pull is integrity-checked against the register's known shape, so a degraded source fails loudly instead of returning bad rows.

Example output

One complaint - business="Amazon.com" returns records like this one:

{
  "platform": "wa-ag-complaints",
  "source": "wa-attorney-general-consumer-complaints",
  "mode": "rows",
  "groupKey": "",
  "groupCount": "",
  "groupBy": "",
  "complaintId": "733076",
  "businessId": "120419",
  "openedDate": "2026-09-15",
  "openedYear": "2026",
  "status": "New",
  "business": "Amazon.com",
  "category": "Electronic Shopping",
  "naics": "454100-Electronic Shopping & Mail Order Houses",
  "naicsCode": "454100",
  "naicsName": "Electronic Shopping & Mail Order Houses",
  "streetLine1": "PO Box 81226",
  "streetLine2": "",
  "city": "Seattle",
  "state": "WA",
  "zipcode": "98108-1226",
  "estimatedSavings": "0.00",
  "actualSavings": "0.00",
  "latitude": "",
  "longitude": ""
}

mode=aggregate, groupBy=category - one row per industry group (groupKey + groupCount):

Electronic Shopping   18,562
Auto Sales            17,443
Telecommunications    15,846
Retail Sales          15,423

Source

• Washington Open Data: Attorney General Consumer Complaints - complaints filed with the WA AG office, refreshed daily. Public open data. Records are the office's own published fields; a complaint is an allegation, not a finding of wrongdoing.