CPA firm administration and quality management archive, one firm

Summary

Firm administration from one US regional CPA firm, 2013 to 2025: operations, IT, marketing and quality management mailboxes, Teams channels and the internal policy library. It covers the firm's quality control system and its move to the newer quality management standards, peer review preparation, independence and client acceptance procedures, security incidents and software rollouts. Client tax and audit files, engagement workpapers, HR threads and partner compensation are excluded. The archive closes in 2025 and is not refreshed.

Collected to order. We collect it from US companies for your order. Scope, volume and timing are confirmed in writing before you sign.

Pricing on request

Request scope and pricing

At a glance

Planned scope. Volume, sources and timing are confirmed in writing before you sign.

Records: 4Memails, channel messages and policy documents
Size: 10.1 GB
Tokens, estimated: 1.1B
Years: 2013 to 2025
Firms: 1

Owner profile: One US regional CPA firm, 150 to 300 staff
Industry: Accounting and finance
Data type: Company operations archive
Headcount: 150 to 300 staff
Region: United States
Languages: English
Formats: JSONL
Updated: 2026-09-23

Sources

System	Kind	Records
Microsoft 365 mail	Email	3.06M messages
Microsoft Teams channels	Team chat	940K messages
SharePoint intranet and policy library	Projects and wiki	18.4K pages and documents

Good for

training agents on professional-standards workflows such as independence checks and client acceptance
evaluating question answering over a policy library that changed across 13 years
modeling IT incident response and change management in a mid-size professional firm
long-context summarization of multi-year quality remediation threads

Fields

Field	Type	Description
message_id	string	Stable message or document id
source	enum	mail, teams_channel or policy_library
mailbox_group	enum	operations, it, marketing, quality or firm_management
sent_at	timestamp	Send or publish time in UTC
author_role	enum	managing_partner, partner, director, manager, staff or it_admin
thread_id	string	Thread or channel conversation id
subject	string	De-identified subject or page title
body	string	Text with quoted replies and signatures stripped
policy_refs	array<string>	Policy library sections the message cites

Example record

{
  "message_id": "F01-M-2214983",
  "source": "mail",
  "mailbox_group": "quality",
  "sent_at": "2024-09-17T13:22:05Z",
  "author_role": "director",
  "thread_id": "F01-T-0771420",
  "subject": "2024 internal inspection results and remediation owners",
  "body": "Inspection is done. 18 engagements reviewed, 3 findings. Two audits released the report before the EQR reviewer signed off in the binder, and one review engagement had no documented independence confirmation for a new staff member. None of them change an opinion, but all three go in the QM 1 monitoring log as deficiencies. [PERSON_22] owns the release hold in the workflow tool by 10/31 and [PERSON_7] updates the independence checklist. Root cause write-up is due before the December partner meeting.",
  "policy_refs": [
    "QM-4.2 Engagement quality reviews",
    "QM-6.1 Independence confirmations"
  ]
}

One record, de-identified. Placeholders such as [PERSON_12] and [COMPANY_4] stay the same across the whole dataset, so threads still read correctly.

Licenses available

Exclusive
One buyer. Nobody else can license it during the exclusive term, usually 2 years.
Non-exclusive
Licensed to more than one buyer.
First look
One buyer gets 2 to 4 weeks to make an offer before anyone else sees the dataset.

What ships with it

Datasheet: what it is, what it is good for, verified size, sources, years, format and limits.
Provenance file: the owner shown by code, staff notice and consent status, exclusions, mapped to the EU AI Act Article 53 training-data summary and California AB 2013.
Manifest with a hash for every file.
QC report and de-identification report.
Your own copy, with unique canary records.

How it is de-identified

Default exclusions removed, files flattened to text, identifiers found in three layers and replaced with consistent placeholders, and the result measured against a 99% bar for each identifier type in each source.

The full method