For AI labs — Gainfluence

01Catalog overview

What is in the pipeline.

Listings are anonymized. Company names and details are shared with verified buyers.

We have 75 open deals across 30+ industries. Each is at one of three stages: submitted, valued or in negotiation.

Supply is sourced from operating companies as well as closed ones, so the catalog includes current data, not only historical archives. New listings are added as sellers submit.

Open the catalog

02Data types and formats

What you can license.

Default formats are listed below. Delivery formats can be agreed per deal.

TODO: Confirm default delivery formats.

A

Workplace data

How real teams communicate, plan and get work done.

Email
Slack and Teams chat
Documents
Calendars
CRM (Salesforce, HubSpot)
Support tickets
Project trackers (Jira, Linear, ClickUp)

Formats MBOX / EML, JSON message threads, original documents with extracted text, CSV or JSON exports per system.

B

Source code

Full repositories with the context around them.

Complete repos with commit history
Linked issues and pull requests
Code review threads
Design files
Internal technical docs

Formats Git bundles with full history, JSON for issues and PRs, original design files with exported previews.

C

Gameplay

What players did, and the games they did it in.

Gameplay footage
Player input logs
Telemetry and event data
Game builds
Engine projects (Unity, Unreal, others)

Formats Video in original resolution, time-aligned input and telemetry as Parquet or JSON, project archives.

D

Industry data

Operating records from specialised sectors.

E-commerce
Real estate
Construction
Professional services
Healthcare operations
Media and film
Other sectors

Formats Structured exports (CSV, Parquet, JSON) and original documents, with a data dictionary.

E

Anything else that is hard to find

Scarcity drives value. If few others hold it, ask us.

03Packaging

How we prepare each dataset.

Deduplication: Exact and near-duplicate records removed within and across systems.
Structure: Threads, conversations, repos and tickets kept intact, with relationships between them preserved where the source allows.
Metadata: Timestamps, authorship roles, system of origin and thread or project IDs, with personal identifiers replaced by consistent pseudonyms.
Redaction: Personal information removed or masked before delivery. The method is documented.
Documentation: A datasheet for each dataset: contents, date range, counts, schema, collection context and known gaps.

04Provenance and rights

Documentation that comes with every license.

You should know where the data came from and who had the right to license it.

Trust & compliance

01Chain-of-title summary showing who held the rights and how they passed to the licensor.
02Seller representations on ownership and authority to license.
03Redaction and PII-handling report.
04Datasheet with schema, counts and date ranges.
05License agreement with scope, term and exclusivity.

05Samples

Evaluate before you commit.

Verified buyers can request a redacted sample of a listing along with its datasheet. Samples are drawn from the same pipeline as the full delivery, so what you evaluate is what you get.

Samples are shared under NDA, through a data room or encrypted transfer.

TODO: Confirm sample size limits and whether a sample agreement is required.

Licensed, documented, model-ready datasets.