Sell Data to AI
Home Data Asset Score Pricing API documentation US labs and CROs list
For brokers
How to become an AI data broker Data broker business model Buyer programs compared Qualify a company AI training data companies
For data companies
Firmographic data providers Company data API
Seller guides
How to sell data to AI companies Is it legal? FAQ and glossary About
Check domain/company
Industry guide: financial services and mortgage

Financial Services and Mortgage Firms: GLBA, Customer Data, and What Remains to Sell

Last checked: 7 October 2026. Written for companies, not consumers.

Lenders, brokers, servicers and finance firms run on customer information that federal law protects. That rules out most of what sits in a loan file. It does not rule out the way your firm works: the procedures, checklists and approvals that move a file from application to closing.

$100K to $5MMode published company range
$100k+micro1 “qualified” published tier
20+ / 30+published employee minimums (Mode / micro1)
60 to 90 daystypical close, as practitioners cite
Why it is wanted

Buyers want the process, not the borrower

Financial work is rule-heavy, document-heavy and full of exceptions. That makes a well-run firm’s procedures a clear record of expert judgment.

micro1 names “Finance & accounting” as one of its example categories and describes it as “Financial processes, SOPs, reconciliations, approvals, reporting workflows, and operational documentation.” Under “Legal & contracts” it lists compliance procedures. Its page also says “Operational data from every industry can contribute” (as published, checked 7 October 2026). Grepped, in its program for individual professionals, lists financial analysts (“Models, memos, forecasts”) and insurance adjusters (“Claims, underwriting, appeals”) among the experts it pays.

None of those descriptions asks for customer accounts. They describe how the work gets done. In a mortgage shop, that is the route a file takes: who reviews what, which conditions come back, how an exception gets approved, what the closing team checks the day before funding. In a wealth or lending firm, it is the monthly reconciliation, the approval chain for a credit decision and the compliance review that follows.

Practitioners say raw data is the cheapest tier and that evaluations built from it are worth roughly ten times more. A documented sequence of decisions, with the reason for each, is the kind of material that can become a test. A pile of statements is not, and it carries most of the legal risk.

Condition clearing

Underwriting checklists, condition categories, and the internal notes on how each kind of condition is usually resolved, written without borrower details.

Closing coordination

Pre-closing checklists, title and settlement handoffs, and the steps your team follows when a date slips.

Reconciliations and approvals

Month-end close procedures, approval chains and the escalation rules for breaks and exceptions.

Compliance and QA

Policy manuals your firm wrote, quality-control review procedures and vendor-management checklists.

Scope

What is mostly off the table, and what may remain

Be blunt with yourself here. For a financial firm, the customer side of the business is largely excluded, and the operational side has to be cleaned first.

Mostly off the table

  • Loan files and loan-origination records tied to borrowers.
  • Credit reports, bank statements, tax returns, pay stubs and appraisals with property owners named.
  • Account numbers, Social Security numbers, balances and transaction histories.
  • Recorded calls and email threads with borrowers or clients.
  • Investor, agency and vendor guides you license but do not own.
  • Exam findings and regulator correspondence, unless counsel clears them.

May remain, after review

  • SOPs, policy manuals and desk procedures your firm wrote.
  • Blank checklists, templates and decision trees for underwriting, processing and closing.
  • Workflow and ticket metadata from Jira or ServiceNow, with any customer reference stripped.
  • Internal process discussions in Microsoft Teams or Slack, after scrubbing.
  • Training materials and onboarding guides for new staff.
  • Vendor-management and QC procedures, with vendor pricing removed if it is confidential.
Risk in a financial firm’s dataMitigation to discuss with counsel and the buyer
Borrower details buried in free-text notes and chatExclude customer-facing channels entirely; scan internal ones; agree audit rights over the de-identified copy.
Your privacy notice did not mention this useHave counsel read the notice against the proposed scope before any sample leaves.
Investor or vendor material inside your SOPsSeparate your own text from licensed guides; send only what you own.
A consent warranty you cannot back upNarrow the representation in the agreement to what is true; cap and time-limit any indemnity.
Record-retention duties on originalsConfirm the deal touches only a copy; originals stay under your retention schedule.
Eligibility

Which published rule applies to a lender or finance firm

No program publishes a separate rule for financial services. The general headcount, record and country rules apply.

ProgramPublished company payoutPublished eligibilityWhat it means for finance firms
micro1 Enterprise Data Partnership“$100k+ qualified”, “$500k+ large-scale”, “$1M+ highly unique”30+ employees, mature operations, documented processes, modern software tools, primarily English; U.S. prioritized, then other Western marketsLists finance and accounting processes and compliance procedures among its examples.
Mode company data“$100K to $5M”20+ full-time U.S. office employees; several years of records the company owns; U.S.-based teams strongest fitMost finance staff are office staff, so the headcount test is usually simple.
Grepped“$20K to $5M”Any vertical; also pays individual professionals for expertiseLists financial analysts and insurance adjusters among the experts it seeks.

Last checked 7 October 2026. Sources: each program’s own website (micro1 data partnerships page, data.mode.inc, grepped.ai). Figures are published ranges, not offers. Miro Advisory also publishes indicative ranges for operating datasets ($100K to $1M+); it has no button here.

Independent site. Some links are referral links: if your company signs with a buyer through them, the buyer may pay us a fee. You are not charged, and we never see your data.

What the buyers say they do, and what it does not settle. micro1 says sensitive and confidential information is scrubbed, originals are deleted after processing, no customer information is exposed, and the company keeps ownership of its underlying data. Mode says it buys an agreed copy, the originals stay with the company, and it de-identifies before onward delivery (both as published, checked 7 October 2026). Those are the buyer’s steps, applied after the data reaches the buyer. Your own duties as a financial firm apply before that moment. A buyer’s cleaning process does not answer whether you were allowed to hand the data over. See de-identification before selling data.
Sector law, in one honest box

GLBA comes first, then your notices and contracts

The Gramm-Leach-Bliley Act (GLBA) is the federal law most financial firms will hear about first. Its privacy rule governs when a financial institution may disclose customers’ nonpublic personal information to nonaffiliated third parties, and what notice and opt-out duties come with that. Its safeguards rule requires a security program for customer information. An AI data buyer is a nonaffiliated third party. That is why customer data is mostly off the table on this page.

  • Your privacy notice. What did you tell customers you share, and with whom? Ask counsel whether the proposed scope fits it.
  • State laws and regulators. States may add their own rules and examiners. CCPA/CPRA in California has its own treatment of GLBA-covered data. Ask how they apply to your firm.
  • Investor, servicer and vendor contracts. These often carry confidentiality and data-use terms of their own.
  • Record retention. Ask which records you must keep, and confirm the deal involves only a copy.
  • Employees. Internal chat includes staff messages. See is it legal to sell company data for the general picture.
Worked example

A conservative scope for a mid-size lender

The point of a written scope is to make the exclusions visible before anyone discusses money. This one carries no price, because a buyer sets any price only after review.

Illustrative, not an offer. Fictional company.

Stillwater Ridge Home Lending (fictional): 60 employees, U.S.-based retail mortgage lender

Systems in scope
SharePoint (policies and procedures), Jira (internal process tickets), Outlook (one shared operations mailbox, internal-only threads), Salesforce (pipeline stage history with every customer field removed), QuickBooks (close procedures, not ledgers).
Date range
2018 to 2025, the years with stable systems and written procedures.
Included
Processing and underwriting SOPs; condition-clearing checklists with no borrower data; about 2,400 internal process tickets with names, loan numbers and addresses stripped; internal question-and-answer threads among staff after scrubbing; QC review procedures; month-end close checklists.
Excluded
Loan files and loan-origination borrower records, credit reports and credit data, bank statements and income documents, borrower call recordings, borrower email, investor and agency guides, exam correspondence, vendor pricing.
Order of steps
Counsel reviews the exclusion list against GLBA and the privacy notice. A manifest (systems, dates, counts, exclusions) goes to more than one buyer. Samples follow only under NDA.
By firm size

A broker, an agency and a servicer: how the rules land

Size decides which published minimum you meet. The type of firm decides how much is left once customer information is taken out. These three walkthroughs show both, with no prices.

Illustrative, not an offer. Fictional company.

Bayline Mortgage Brokers: 15 people

An independent broker with 15 office staff. That is below Mode’s published 20+ office-employee minimum and micro1’s 30+ (as published, checked 7 October 2026). Grepped says it accepts any vertical and also pays individual professionals. Even if a program accepted it, little would remain: most of a broker’s records are borrower files. What survives is a short list of intake SOPs, lender-submission checklists and training notes.

Illustrative, not an offer. Fictional company.

Granite Mile Insurance Agency: 55 people

A commercial lines agency with 55 office staff, so it clears both published floors. Its policyholder files carry personal and business information and stay out. What may remain is process: renewal workflows, submission-quality checklists, internal service tickets with client fields stripped, and staff training material. Carrier manuals and underwriting guides belong to the carriers, so they stay out too.

Illustrative, not an offer. Fictional company.

Northgate Loan Servicing: 300 people

A lender and servicer with 300 staff. It clears every published size line, and it also has the most exposure: payment histories, hardship files and borrower calls. Those are excluded. A scope could still cover default-operations SOPs, internal QC procedures, a decade of process tickets with loan numbers removed, and policy change histories. Investor, servicer and vendor contracts are read before anything is scoped.

Records inventory

What a financial firm holds, and what is off-limits

List every record type before you talk to a buyer. In this industry the off-limits column is long. The includable column is still worth writing down.

Record typeTypical systemUsually includable?Why
Policies, procedures and SOPs you wroteSharePoint, Google Drive, ConfluenceUsually yesYour own process documents, with little or no customer information.
Checklists and stacking orders (blank)SharePoint, loan origination system templatesUsually yesThey show the steps without anyone’s file attached.
Internal process ticketsJira, ServiceNow, AsanaPossibly, after scrubbingLoan numbers, names and addresses often sit in titles and comments.
Pipeline stage historySalesforce, HubSpotPossibly, metadata onlyStage timing can show the workflow once every customer field is removed.
Month-end close and reconciliation proceduresQuickBooks, NetSuite, XeroProcedures yes; ledgers rarelyProcedures describe work; ledgers carry customer and account data.
Internal staff chat about processMicrosoft Teams, SlackSelected channels after reviewStaff paste account details into chat; every channel needs checking.
Loan files, applications, credit reportsLoan origination system, document storageNoNonpublic personal information, the core of GLBA concerns.
Bank statements, tax returns, pay stubsDocument storage, email attachmentsNoCustomer financial data collected for one purpose.
Recorded customer callsZoom, phone systemNoCustomer voices and details, plus recording-consent questions.
Investor and agency guides, vendor manualsDownloaded PDFs, vendor portalsNoYou do not own them, and their terms usually limit reuse.
Regulatory exam correspondenceEmail, secure portalsNoOften confidential by rule or agreement; ask counsel before treating it as yours.

General guidance on typical records, not a legal classification. Your counsel decides what counts as nonpublic personal information in your firm.

Owner decisions

Settle these before you apply

A regulated firm should be able to answer each question in writing. If you cannot answer one yet, that is the first task, not the application.

Owner go/no-go checklist

  • Has counsel read our GLBA privacy notice with this use in mind, and told us what it allows?
  • Have we read our investor, servicer, warehouse-line and vendor contracts for confidentiality and data-use terms?
  • Can we produce a meaningful scope with every customer record removed? If not, stop here.
  • Do we meet a published size minimum, and have several years of records in systems we own?
  • Does compliance own the exclusion list, with sign-off before any sample leaves?
  • Are we ready to share a manifest with more than one buyer, and wait for offers?

Notice: customers and staff

  • Customers. Your privacy notice already told customers how you share their information. A plan that keeps customer data out avoids testing that notice. If any customer-derived data is in scope, ask counsel what the notice permits and whether anything must change first.
  • Loan officers and processors. Their internal messages and tickets are the material. Tell them what is in scope and what is excluded before export. See employees and selling company data.
  • Partners and counterparties. Investors, servicers and vendors may have contract rights over shared documents. Ask counsel whether any must be told or asked.

Common mistakes in financial services

Relying on the buyer’s de-identification

Buyer privacy steps happen after the data leaves you. Under GLBA, the disclosure question is yours to answer first.

Including recorded calls

Borrower and client calls carry voices, account details and consent issues. Transcripts do not remove the problem.

Shipping documents you do not own

Investor and agency guides, carrier manuals and vendor documentation sit on every shared drive. They are not yours to license.

Missing free-text fields

Account numbers, loan numbers and income figures appear in ticket titles, notes and chat. Structured fields are the easy part.

Forgetting attachments

A clean email thread can carry a bank statement or a credit report as an attachment. Exclude attachments unless each one is reviewed.

Skipping compliance until the end

A late compliance review can cut the scope after a buyer has priced it. Put compliance on the scope from the first draft.

Before you sign

Questions a regulated firm should put to any buyer

Generic questions, not claims about any named program. They matter more for financial firms because the downside of a miss is larger.

  1. Consent representations. Will the agreement ask you to warrant that customers or employees consented? Limit it to what is true.
  2. Indemnities. If a customer detail survives de-identification, who pays? Is there a cap, and when does it expire? See indemnities and warranties.
  3. Audit rights. Can you inspect the de-identified copy, or receive evidence of the process, before onward delivery?
  4. Scope of use. Training only, or evaluation as well? Which downstream buyers may receive the copy?
  5. Exclusivity and resale. Is the license exclusive, time-limited or perpetual? Can the buyer resell?
  6. Deletion. When are the originals and the copy deleted, and will you get written confirmation?
  7. Payment and acceptance. One payment or milestones? What makes the data “accepted”?
  8. Termination and survival. What happens to the delivered copy if the agreement ends?
Timing and offers. Mode publishes that it generally expects about three months from the first conversation through payment (checked 7 October 2026). Practitioners cite 60 to 90 days to close. They also advise sharing a manifest and samples first, never the full dataset before a price, and getting more than one offer. Nothing here promises acceptance, an amount or a date.
FAQ

Financial services and mortgage: common questions

Can a mortgage company sell its loan files to an AI data buyer?

Almost certainly not as loan files. A loan file is full of nonpublic personal information: income, account numbers, credit reports, Social Security numbers. The Gramm-Leach-Bliley Act (GLBA) limits how financial institutions disclose that kind of information to unaffiliated third parties. What may remain is the process around the file: your SOPs, checklists and the internal steps your team follows, with borrower details removed.

Does the buyer's de-identification make it safe to share customer data?

Not by itself. Buyers publish their own privacy steps, such as scrubbing sensitive information or de-identifying before onward delivery. Those steps happen after the data leaves you. Whether you may disclose the data in the first place is your question to answer under GLBA, your privacy notice and your contracts. Ask your lawyer before any sample leaves.

Which financial-services records do buyers describe wanting?

micro1 lists financial processes, SOPs, reconciliations, approvals, reporting workflows and operational documentation among its example categories, and names compliance procedures under legal and contracts (as published, checked 7 October 2026). In a lender or finance firm, that points to how work moves: condition clearing, closing coordination, exception handling and quality control.

Do small mortgage brokers meet the published minimums?

Some do. Mode publishes 20+ full-time U.S. office employees for most businesses, and micro1 publishes 30+ employees. Grepped says it accepts any vertical. A 15-person brokerage would fall under the first two published floors. The eligibility checker compares your numbers with each rule.

Can we include recorded calls with borrowers or clients?

Treat them as out of scope. Recorded calls carry customer voices, account details and recording-consent questions, and turning them into transcripts does not remove the customer information. Internal calls among staff about process are a separate question for your lawyer.

Are investor guides and vendor manuals part of our data?

No. Agency and investor guides, carrier manuals and vendor documentation are owned by others, and their terms usually limit reuse. Only documents your firm wrote and owns belong in a scope. Your own procedures that refer to those guides can usually be reviewed separately.

Should we tell customers before selling operational data?

If no customer information is in scope, the main question is staff notice, not customer notice. If any customer-derived data is in scope, ask your lawyer what your GLBA privacy notice permits and whether anything must change first. This is general information, not legal advice.

How long does a deal take for a regulated firm?

Expect it to take longer than the headline numbers suggest. Mode publishes that it generally expects about three months from the first conversation through payment, and practitioners cite 60 to 90 days to close. A compliance review of the scope on your side adds time, and it should.

Check the published rules against your firm

The checker runs in your browser and saves nothing. It shows each program’s published rule next to your numbers, with a separate note for confidential or regulated records.

Related

Further reading