Every ticket you resolved is a worked example. Buyers pay for the resolution, not the volume.

Companies with a large help desk, outsourced contact centres, and the operations leaders who run support in software, telecoms, banking, insurance, utilities and travel.

6

buyers quoted below on what they want from companies like yours

6 of 7

kinds of record below that the valuation measures or values

1

lot like yours among those a lab buyer chose on 12 Oct 2026

What companies like yours hold

What you typically have, where it sits, and whether the valuation can measure or value it. Buyers price what exists together, over how many years, and how well it links.

  • Tickets with thread and resolution
    Where it sits, and what the evidence shows
    Zendesk, Intercom, Freshdesk, ServiceNow or Salesforce Service Cloud: the request, internal notes, escalations and the fix. The marketplace lot a lab buyer chose on 12 October held 57K.
    How the valuation reads it
    Measured from your tickets exportCounted with tickets and projects
  • Live chat and messaging
    Where it sits, and what the evidence shows
    Intercom, LivePerson, Genesys or Zendesk Messaging transcripts, including hand-offs from a bot to a person.
    How the valuation reads it
    Measured from a CSV or JSON export
  • Call recordings
    Where it sits, and what the evidence shows
    NICE, Genesys, Five9, Verint or Talkdesk; best on two channels, with disposition codes. That lot held about 83K sales-call recordings.
    How the valuation reads it
    Describe it in your offer
  • Quality scorecards
    Where it sits, and what the evidence shows
    Scores against a written rubric, coaching notes and calibration between reviewers.
    How the valuation reads it
    Measured from a CSV or JSON export
  • Help-centre articles and macros
    Where it sits, and what the evidence shows
    Public articles, internal troubleshooting guides and the playbooks agents work from.
    How the valuation reads it
    Measured from your documents exportCounted with documents and knowledge
  • AI assistant logs with corrections
    Where it sits, and what the evidence shows
    Conversations handled by an AI agent, with the human edits, ratings and escalations attached.
    How the valuation reads it
    Measured from a CSV or JSON export
  • Customer and account records
    Where it sits, and what the evidence shows
    The CRM that ties a ticket or a call to an account and an order.
    How the valuation reads it
    Measured from your customer records exportCounted with customer records

What buyers say they want from companies like you

In their own words, from their own pages or from reporting on what they bought. Each opens its source.

  • “Salesforce or HubSpot CRM, service histories, support tickets, and the record of how requests and issues got resolved.”
    NyneIts own page, 12 Oct 2026
  • “Support documentation, ticket workflows, QA processes, customer operations, and knowledge management”
    micro1Its own page, 12 Oct 2026
  • “Human feedback on AI outputs in real business environments to improve model performance”
    micro1Its own page, 12 Oct 2026
  • “AI performance data AI at work in your environment”
    TroveoIts own page, 12 Oct 2026
  • “Tickets, approvals, escalations, incident reviews, QA and audit trails.”
    SuperAnnotateIts own page, 12 Oct 2026
  • “material covering pauses, interruptions, overlapping speech and natural response”
    Frontier labs, buying from Ocular AIAs reported, 10 Oct 2026
  • “Org artifacts (tickets, specs, docs, support threads): $50 per artifact when bundled with a codebase”
    Turing (Project Lazarus)Its own account, 24 Mar 2026

What raises the value

  • The resolution recorded, not just the request: Nyne names "the record of how requests and issues got resolved".
  • Tickets linked to the order, the account or the code change that fixed them.
  • Internal notes and the escalation path kept in the thread.
  • Two-channel audio with natural overlap and interruption, and disposition or outcome codes.
  • Languages and accents beyond US English.
  • Human corrections attached to each AI failure, delivered as a continuing feed: "Continuous feeds: Recurring data, recurring revenue" (Troveo).

What lowers it

  • Replies that are only macros, and tickets closed without a resolution.
  • Mono audio with both speakers mixed together.
  • English call-centre audio, which already sells off the shelf at catalogue prices: $5 to $25 an hour of audio in one vendor's comparison (11 Oct 2026).
  • Consent that covers only quality and training purposes.
  • Volume alone: "More records do not automatically mean a higher offer." (Nyne).

A line that quotes a buyer names it. The others are our checklist, drawn from the factors buyers state in general: no buyer publishes how much each one moves a price. How often each factor matters across 156 kinds of data.

The rights to settle first

These stop a sale more often than price does. Each can be settled before any buyer sees a record.

The full rights checklist
  • Consent is the blocker for recordings. In the Spirit Airlines estate sale, 30.9M call recordings and 15.8M chat sessions were left out from the start (docket, as reported).
  • Voiceprint claims: a 2026 suit "accused Walmart of using an AI-powered customer-service phone system that allegedly created voiceprints from callers" (CX Today, 4 Oct 2026).
  • Card numbers and identity checks are spoken on calls and typed into chats; they have to be found and removed.
  • A contact centre's tickets and calls usually belong to its clients: customer data your contracts reserve for serving that customer stays reserved. Service agreements and data processing terms usually limit it to performing the contract.
  • Named agents' scores are employee data, and works councils or unions may limit their reuse.
  • AI assistant logs: the model provider's terms may restrict sharing outputs, and prompts can embed confidential instructions.

A recent purchase that fits

One of five lots a lab buyer went for on 12 Oct 2026, wanting to close in under a week. No price was given.

All five lots
Seller
Creator marketplace
People
55
History
5 years (2021 to 2026)
Size
About 2.1 TB
What was inside
  • 96 repositories
  • 313K CRM contacts
  • About 83K sales-call recordings
  • 57K support tickets
  • About 200K Drive files
  • 120K+ Slack messages
  • 58M analytics events

Source: A buyer-side referral network's announcement to its members, 12 October 2026 (reported to us).

What is already on offer from companies like you

Read this morning from the brokers' and marketplaces' own catalogues. A buyer compares your records with these; settled rights, more years and more linked systems are what set one apart.

Every lot on offer

11

lots on offer like yours

3 years

of history in the middle of them

1 of 11

say the rights are settled

What holders like you ask

We have millions of tickets. Does that make them valuable?
Volume is the factor buyers warn against: "More records do not automatically mean a higher offer." (Nyne). What they name is the resolution: "the record of how requests and issues got resolved". A ticket with its thread, internal notes, escalation and fix, linked to the account or the code change, is a worked example an AI agent can practise on.
Are our call recordings sellable?
Only with consent that covers this use. Natural two-way conversation is scarce: frontier labs buy "material covering pauses, interruptions, overlapping speech and natural response" from specialists (as reported, 10 Oct 2026). Ordinary English call-centre audio already sells off the shelf at $5 to $25 an hour, and the Spirit Airlines estate left its 30.9M call recordings out of its sale. Recordings in other languages, on two channels, with consent that names AI training, are the ones with room for a price.
We use an AI assistant in support. Is its log worth anything?
micro1 names "Human feedback on AI outputs in real business environments" among what it buys, and Troveo lists "AI performance data". The log matters when each failure carries the human correction or the escalation outcome, and more as a continuing feed, which can be licensed again each period.
We are a contact centre working for clients. Can we sell?
Usually not your clients' records. Customer data your contracts reserve for serving that customer stays reserved: service agreements and data processing terms usually limit it to performing the contract. The client signs for its own tickets and calls. Your own quality rubrics, coaching guides and training material may be yours to offer.
What does a buyer pay for support data?
No public price exists per ticket or per conversation. The one unit price is Turing's $50 per artifact for tickets, specs, docs and support threads when bundled with a codebase (24 Mar 2026). Support records usually sell inside a whole company's archive, priced per company.

In this topic

What sells

  1. Start hereWhich data do AI labs want, and who holds it?Each kind of data the market prices, the businesses that hold it, what labs are asking for now and what they turn away.
  2. GuideWhat Are RL Environments? Who Builds, Buys and Prices ThemAn RL environment is a sandbox where an AI agent attempts tasks with tools and a verifier scores each try. Who builds them, who buys them and what they cost.
  3. GuideHow to Sell Videos to AI Companies: Prices, Buyers and RightsAI companies license video by the hour or minute. What buyers pay, which footage sells, and the rights you need in every frame before you sell a library.
  4. PageWhere data comes fromWho holds each kind of data, how much exists and how it reaches buyers.
  5. RegisterCompanies closingCompanies closing, sold or gone dark, and the records they leave.

See what your tickets and records show, measured on your own computer. Recordings you describe.

Value your company's recordsCheck an offer you have
Selling support tickets and call recordings to AI labs | AIData.Broker