Prepare a company data summary

Describe the work behind the records

A company data summary explains what you hold before anyone handles the records. Write a short description of the business, the work people did and the history you retained. State whether the company still operates, is winding down or has closed.

Name the record families you could discuss: support tickets, project documents, email, code or media. Explain how they relate. A ticket linked to its resolution tells a different story from an isolated message. Leave customer names, employee details, passwords and confidential excerpts out of this summary.

Keep a dated copy. If you later exclude a system or discover a rights restriction, update the summary before using it again.

Make a small inventory

Use one row for each record family. Keep estimates visibly separate from counts you have checked.

Defined.ai says its partner assessment considers quality, volume, organization and available metadata. Its page publishes no fixed minimum volume; requirements depend on the dataset and use. A count alone cannot settle fit. Defined.ai partnership programs, checked November 8, 2026.

Keep the inventory short and check it against the records. Do not describe a complete archive if attachments are missing or an export covers only part of the stated period.

Separate existing records from new work

Mark whether you already hold the material or would need to record it. These lead to different discussions about effort, access and permissions.

Sieve describes both licensing existing content and arranging new physical-work capture. Its partner process includes rights checks. Someone with an existing video library should not assume that a request to film new tasks uses the same terms. Sieve partners, checked November 8, 2026.

For a new collection request, ask who supplies equipment, pays preparation costs and obtains permissions before agreeing to the work. Keep that request separate from the inventory of records you already hold.

List the questions you cannot answer yet

Show the gaps in your summary. Add a short list covering client-owned material, third-party content, personal information and restrictions in existing agreements. Use the data rights checklist to prepare that list with the people who hold the relevant documents.

Choose an internal person to resolve each question. Record a next step such as finding a contract, checking an export or asking counsel. Leave the item unresolved until that work is done.

Agree the sample process before sharing

A program may ask to inspect a sample. SuperAnnotate describes reviewing a sample before a valuation and preparing data for buyers the seller approves. That is its published process; ask any program to explain its own. SuperAnnotate data partnerships, checked November 8, 2026.

Before sending a sample, ask what it needs, who will see it, which uses are allowed during evaluation and what happens if there is no deal. Agree how the sample will move and how access ends. Keep those answers with the summary.

When the summary is ready, compare the listed programs or use the local qualification check. Bring any proposal back to the licensing terms checklist. A plausible match starts a conversation; it does not establish an offer.

For a full process, follow the company-data licensing checklist. A closing business also needs a clear authority and custody record.