Polyshares, Inc. describes itself as a data lab that buys licenses to company operational data and supplies it to AI labs. It says it is the counterparty on each deal, not a marketplace or agent, and charges the seller no fee. Its managing partners, per its team page, are Nathan Mathews and Ryan Lock. Business Insider places Lock in Dallas.

Polyshares publishes named partners, a data-handling page and four case studies from $110,000 to $450,000, but the companies are unnamed, so the deals can't be checked independently. Its standard license terms link returned a not-found page on October 21, 2026. Trustpilot search results that day showed 4 stars from 5 reviews; we couldn't confirm the review content, and five reviews is too few to judge.

Does Polyshares buy company data?

Yes, as a license rather than an outright purchase. Polyshares says it licenses defined operational data directly from the company, pays for it itself and keeps no equity or debt in the deal. The seller keeps the business and the underlying data. Polyshares says typical deals run $100K to $2M and above.

How does a Polyshares offer work?

Polyshares says intake asks for record counts, not files, then it reads an anonymized sample to price the record and makes an offer within one to two weeks of intake. Terms are papered after that. Payment is net 30 to 60 once the data is approved, anonymized and delivered. The license is exclusive by default unless agreed otherwise.

Polyshares Review 2026: What It Takes and How Data Is Handled

How Polyshares licenses company Slack, email, CRM, ticket and call data, who anonymizes it, and what its terms say, checked October 21, 2026

Polyshares buys licenses to a company’s operational record (Slack, Teams, email, Drive, CRM history, support tickets, call recordings and accounting exports) and says a specialist PII partner strips names, emails, phone numbers and account numbers before Polyshares or any lab sees it. Deals are typically $100K to $2M, exclusive by default, per its pages read October 21, 2026.

Onfolio may earn a referral fee if you sign up to Polyshares through our links. This does not affect our editorial assessment.

Polyshares says it is the buyer, not a broker: it licenses data directly from the company, pays for it, builds it into data sets and supplies AI labs. That structure matters for ops teams, because the contract you sign is with Polyshares, and the labs that end up with the data aren’t named. This review covers what it asks for, how the data moves and which controls it publishes.

What Polyshares takes

Polyshares calls its target “the working record of the business” and says the length and texture of that record matter more than headcount or revenue. Its homepage and data-licensing page name these systems.

Data type	Named by Polyshares	What Polyshares says is stripped or kept
Slack and Microsoft Teams	Yes	Names, emails, phone numbers, account numbers and addresses removed or replaced; roles, timestamps and the work itself kept
Email	Yes	Same rule
Google Drive, SharePoint, spreadsheets	Yes	Same rule
CRM history	Yes	Customer names, emails and account numbers stripped
Support tickets	Yes	Same rule
Design files	Yes	Not published beyond the general rule
Call recordings	Yes	Not published (no description of how audio is handled)
Accounting and ERP exports	Yes	Account numbers stripped
Data you hold for your clients	Excluded	The license “does not cover what you hold on behalf of your own clients”
HR records, payroll, health data	Not published	Not published

For a RevOps or IT lead, the list reads like a full workspace export: chat, mail, files, CRM and tickets. The pages don’t say whether each system can be licensed separately, though the data-handling page describes the license as covering “defined systems.”

How the data gets out of your systems

Polyshares doesn’t publish an extraction method. Its pages don’t say whether it uses connectors, OAuth access or files your team exports. The sequence it does describe runs like this.

Intake by counts. Polyshares says intake asks for record counts, not files or exports, and “nothing leaves your systems until Polyshares has made an offer and you have accepted it.”
Anonymization by a partner. A specialist PII partner removes or replaces identifying fields “before Polyshares or any buyer sees the record.” The partner isn’t named on the pages we read.
Sample review. Someone at Polyshares reads an anonymized sample to price the record. Polyshares says it doesn’t pay for something it hasn’t seen.
Offer. An offer arrives within one to two weeks of intake and is “papered right after.”
Delivery and payment. Payment is net 30 to 60 once the data is approved, anonymized and delivered.

Two points need clarifying before any export. Polyshares doesn’t say whether the PII partner works inside your environment or on data transferred to it. Its pages also don’t explain how the anonymized pricing sample reaches the partner and Polyshares if nothing leaves before an accepted offer. By comparison, Scale AI says it anonymizes inside the customer’s environment before extraction.

Privacy and security controls

The controls below are Polyshares’ own claims from its data-handling page and homepage, read October 21, 2026.

Control	What Polyshares says	Ask for in writing
De-identification	A specialist PII partner removes names, emails, phone numbers, account numbers and addresses; roles and timestamps stay	The partner’s name, its method, where it runs, and a sample output
Re-checks	If a sample still reads as identifying, it goes back for another pass	The acceptance test for “anonymized” and who signs it off
Internal access	Two named people see a company’s identity alongside its data	Their roles and how access is logged
Buyer view	Buyers see de-identified listings in a buyer portal, with a code instead of the company name	Which buyers or buyer types can browse listings
End buyers	“We do not disclose which labs take which data”	Named end buyers, or a right to approve or be notified
Use limits	No resale or use outside the license scope	The resale and sublicensing clause
Security certifications	Not published	A SOC 2 report or equivalent
Transfer and encryption	Not published	Transfer method and encryption in transit and at rest
Retention and deletion	Not published	Retention period, deletion certificate, and what happens to models already trained
Sub-processors	PII partner not named	A sub-processor list

The data-handling page also says buyers never contact your customers or staff and that Polyshares is the sole counterparty. We couldn’t check these commitments against a contract, because the standard license terms link returned a not-found page on October 21, 2026.

Commercial terms in brief

These are Polyshares’ published terms, read October 21, 2026.

Term	What Polyshares says
Stated deal size	$100K to $2M, “and above for large records”
Published case studies	Four unnamed licenses: roofing $450,000, general contractor $300,000, solar $275,000, global marketing agency $110,000
Payment timing	Net 30 to 60 after the data is approved, anonymized and delivered
Exclusivity	Exclusive by default unless agreed otherwise; all four case studies are “perpetual exclusive” licenses
Fee to seller	None; no fee or commission
Headcount floor	None published

Three items on Polyshares’ own pages don’t line up. The homepage says “no marketplace, no listing,” while the data-handling page refers to listings in a buyer portal. The data-handling page describes “a defined term,” while every case study is a perpetual license. And the “standard license terms” link returned a page-not-found message on October 21, 2026, so we couldn’t read the standard terms.

A blog post by Franklin Bryant IV (October 8, 2026) says Ryan Lock published “market standard” terms including two-year exclusivity, a 30-day acceptance window, net-60 payout and third-party de-identification. It isn’t clear whether those are Polyshares’ own contract terms or general market terms, and we couldn’t confirm them on Polyshares’ site.

What data you can and can’t license

Polyshares’ own exclusion is the starting point: its license doesn’t cover data you hold on behalf of clients. For a marketing agency or MSP, that removes a large share of the workspace.

Usually firm-owned and licensable once scrubbed	Usually not yours to license, or high risk
Your SOPs, playbooks, templates and internal knowledge base	Client ad accounts, client ESP lists and client deliverables under NDA
Internal tickets and project history, with client names removed	Client Slack Connect channels and shared client folders
Your own campaign process docs and CRM pipeline records, with contact details stripped	Contact-level CRM personal data (names, emails, phones)
Design files and marketing assets you made for yourself	HR systems, performance reviews, compensation and health data
Internal chat about how work gets done, de-identified	Call recordings made without all-party consent where it’s required

The general rules behind the right-hand column:

Personal data. CCPA/CPRA has covered employee and B2B contact data since January 1, 2023. GDPR applies to EU and UK people in your records.
Employee chat. NLRA Section 7 protects employee discussions of pay and working conditions, which often sit in Slack.
Call recordings. Polyshares lists call recordings as operational data. Consent rules differ by state, and Brewer v. Otter.ai (August 2025) alleges wiretap violations. Treat recordings as high risk.
Platform terms. Slack’s API terms (August 2025) bar bulk export and LLM training via the API, per Hunton. Google Workspace API policies restrict using data to train non-personalized models.
Your own privacy notice. The FTC said in February 2024 that retroactively changing privacy terms to allow AI training can be deceptive.

Risks for the ops team

Client confidentiality. Many 2026 agency contracts carry a “no client data used to train AI” warranty, and a service provider’s access to client data doesn’t authorize monetizing it (RICHT law firm, October 18, 2026). Check client contracts before scoping.
Exclusivity. Exclusive by default, with perpetual exclusive case studies, means you can’t license the same record again. Epoch AI’s interviews (January 12, 2026) found exclusive deals for AI training environments run 4 to 5 times the price of non-exclusive ones, so ask what exclusivity adds to the price.
Deletion vs trained models. Polyshares doesn’t publish deletion terms. Deleting data doesn’t remove it from models already trained on it.
Buyer opacity. Polyshares doesn’t disclose which labs take which data.
Payment timing. Your team does the export work before money arrives, and payment is net 30 to 60 after delivery.
Short deadlines. Business Insider (October 13, 2026) reported that six-figure offers expiring in 48 hours have become common across this market. In the same article, Lock compared the market to the early Texas oil boom and said his firm aims to give owners more transparency on value.
M&A diligence. A prior data license, its exclusivity and any change-of-control terms surface in due diligence (Frankfurt Kurnit via Mondaq, October 2026).

Pre-export checklist

List every system in scope and its record counts, matching the “defined systems” the offer names.
Pull client contracts and flag any AI-training, confidentiality or data-ownership clauses.
Remove client workspaces, Slack Connect channels, client ad accounts and ESP lists from scope.
Exclude HR, payroll, health data and call recordings unless counsel clears them.
Request the PII partner’s name, method and processing location in writing.
Ask how the pricing sample is produced and where it is processed.
Get the full license terms, including term length, exclusivity, deletion and resale limits.
Check whether your Slack and Google Workspace export method complies with each platform’s terms.
Confirm who signs, who pays and the payment date before your team starts the export.

Deciding whether to sell at all? onfolio.com’s Polyshares review covers it from the owner’s side: whether it’s legit, what you might get and what can go wrong.

Sources

Polyshares homepage, read October 21, 2026: https://polyshares.com/
Polyshares data-licensing page, read October 21, 2026: https://polyshares.com/data-licensing
Polyshares data-handling page, read October 21, 2026: https://polyshares.com/data-handling
Polyshares case studies, read October 21, 2026: https://polyshares.com/track-record
Polyshares team page, read October 21, 2026: https://polyshares.com/team
Polyshares answers page, read October 21, 2026: https://polyshares.com/answers
Business Insider, “AI labs are fueling a wild market for your Slack DMs,” October 13, 2026: https://www.businessinsider.com/ai-training-employee-data-emails-slack-dms-2026-10
Franklin Bryant IV, “data brokerage boom” (secondhand, unconfirmed), October 8, 2026: https://franklin.simplifyingbusinesses.com/blog/data-brokerage-boom
Epoch AI, “An FAQ on reinforcement learning environments,” January 12, 2026 (interview-based): https://epoch.ai/gradient-updates/state-of-rl-envs
Trustpilot search results for Polyshares, October 21, 2026
RICHT law firm on agency client data, October 18, 2026
Hunton on Slack API terms, August 2025
Frankfurt Kurnit via Mondaq on data licenses in M&A, October 2026
FTC guidance on retroactive privacy policy changes, February 2024
Brewer v. Otter.ai complaint, August 2025

Polyshares’ data handling page

FAQ

What is Polyshares?: Polyshares, Inc. describes itself as a data lab that buys licenses to company operational data and supplies it to AI labs. It says it is the counterparty on each deal, not a marketplace or agent, and charges the seller no fee. Its managing partners, per its team page, are Nathan Mathews and Ryan Lock. Business Insider places Lock in Dallas.
Is Polyshares legit?: Polyshares publishes named partners, a data-handling page and four case studies from $110,000 to $450,000, but the companies are unnamed, so the deals can't be checked independently. Its standard license terms link returned a not-found page on October 21, 2026. Trustpilot search results that day showed 4 stars from 5 reviews; we couldn't confirm the review content, and five reviews is too few to judge.
Does Polyshares buy company data?: Yes, as a license rather than an outright purchase. Polyshares says it licenses defined operational data directly from the company, pays for it itself and keeps no equity or debt in the deal. The seller keeps the business and the underlying data. Polyshares says typical deals run $100K to $2M and above.
How does a Polyshares offer work?: Polyshares says intake asks for record counts, not files, then it reads an anonymized sample to price the record and makes an offer within one to two weeks of intake. Terms are papered after that. Payment is net 30 to 60 once the data is approved, anonymized and delivered. The license is exclusive by default unless agreed otherwise.