Raii Data Group

How it works

A 12-month partnership. We do the work. You get paid.

Licensing data to AI labs is new to almost every company we talk to. Under one agreement we handle the parts you don’t do every day — audit, packaging, rights diligence, buyer access, negotiation — and leave you in control of every decision that matters.

The model

We’re a broker, not a vendor.

We don't buy your data

We don't take a position in it, resell it, or mark it up. Our incentive is to get you the best license at the best price, because our 25% is a share of it.

We don't hold your data

Samples and deliveries run through your infrastructure or a buyer-provided secure channel. We never take custody, so there is no Raii copy to secure, breach, or audit.

We don't become the licensor

You sign the license directly with the buyer. Warranties, indemnities, and obligations run between the two parties who actually own and use the data.

The economics

$0 upfront. Raii earns 25% of every license it sells.

The 25% is fixed in the partnership agreement before we contact a single buyer, so there’s nothing to renegotiate at closing. It is due only when the buyer pays you, and only on licenses Raii sold.

Upfront

$0

Retainer

$0

Fee to buyers

$0

Raii's share

25%

Term

12 mo

Why 12 months?

Deals take four to seven months from first contact to signed license. Buyers need to trust that the data is exclusively represented before they commit diligence time. And we only invest in packaging and running a process when we know we’ll be in the deal. Twelve months, exclusive, gives all three enough room.

Why 25%?

It covers the audit, the packaging, the rights diligence pack, the buyer process, and the negotiation. Comparable agents and brokers take 15–35%. Most companies net far more with us than selling alone, because we run competitive processes with multiple buyers at the table.

Raii’s share is 25% of license fees, paid when the buyer pays. Renewals and additional buyers for the same asset during the term are on the same basis. Third-party costs you choose to incur (outside counsel, de-identification vendors) are yours and are always approved by you in advance.

The 12-month timeline

From signature to first license in four to seven months.

Timelines depend on how license-ready the data is and how fast the buyer’s legal team moves. This is what a typical partnership looks like.

  1. Month 0

    Sign the partnership

    A 12-month exclusive partnership agreement. It fixes Raii's 25% share of license fees, confirms there is nothing upfront, and sets the ground rules: you approve every document, you sign every license, and we never take custody of the data.

  2. Weeks 1–3

    Audit and data sheet

    We inventory what you have — volume, structure, time span, how it was collected, who holds the rights — and write the one-page data sheet buyers see first. You get a written read on what is sellable and a price range.

  3. Weeks 3–6

    Diligence pack

    We build what buyers ask for before they commit: provenance and rights chain, schema and data dictionary, PII posture, sample protocol, and a license term sheet. Rights issues get fixed here, with your counsel if needed. You approve everything.

  4. Months 2–6

    Buyer process

    We approach the labs and data vendors whose requirements match, under NDA and non-circumvention. Qualified buyers get sample access through your systems, never ours. We run a competitive process rather than taking the first offer.

  5. Close

    Negotiate and sign

    We negotiate price and terms alongside you and coordinate legal review. You sign directly with the buyer and are paid directly by the buyer. Delivery happens on your infrastructure or a buyer-provided secure channel. Raii invoices its 25% when the buyer pays.

  6. Ongoing

    Renewals and additional buyers

    Most licenses are time-bound and non-exclusive. We manage renewals, refresh deliveries, and additional buyers for the same asset throughout the partnership, each on the same 25% basis.

Clean rights

What “clean rights” means to us.

Every dataset we represent has to clear this checklist. If yours doesn’t yet, that is often fixable — the diligence-pack phase exists to fix it.

  • Documented origin

    You can state how each record was created: by your employees, by your users under terms you can show, or under a third-party license you can produce.

  • Rights to license for ML training

    Your terms of service, employment agreements, or upstream licenses permit sublicensing the data for model training. If they're silent, we'll help you assess the risk — but silence is not consent.

  • No scraped or ambiguous sources

    The dataset does not include content copied from third-party websites, APIs, or platforms without an explicit license to redistribute it.

  • Personal-data posture is known

    You know whether the data contains PII, what consent or legal basis applies, and whether de-identification has been performed and how.

  • No conflicting exclusivity

    You have not already granted an exclusive license or representation, and no customer contract prohibits you from licensing derived or aggregate data.

  • Regulated-data compliance

    Health, financial, children's, or biometric data has the relevant regulatory basis (HIPAA, GLBA, COPPA, BIPA, GDPR) and your counsel has signed off on the intended use.

Ready to talk?

Apply for a partnership and we’ll follow up within two business days.